VOLUME 14, ISSUE 3, MARCH 2025

Abstract: This paper introduces a system that harnesses the recommendations of generative artificial intelligence (AI) and more specifically, large language models (LLMs) to develop machine learning (ML) models for the automated detection of non-alcoholic fatty liver disease (NAFLD) on the basis of liver B-mode ultrasound images. The image dataset is minimal so the option of utilizing convolutional neural networks (CNNs) and deep learning (DL) approaches built around artificial neural networks and comparable systems is not pursued. Rather, experiments are carried out with simpler machine learning algorithms and classifiers such as random forest classifier, logistic regression and decision tree classifier. Results indicate reasonable performance in light of the fact that the utilization of CNNs and comparable DL approaches could lead to overfitting of the data. The generative AI is prompted with tailored prompts engineered to elicit recommendations that account for the characteristics of the dataset.

Keywords: Generative Artificial Intelligence (AI), Large Language Model (LLM), Convolutional Neural Network (CNN), Deep Learning (DL), Machine Learning (ML), Healthcare System, Disease Diagnosis and Prediction, Non-alcoholic Fatty Liver Disease (NAFLD).

Abstract: Tuberculosis chest radiography image datasets are used to train convolutional neural networks designed for automated diagnosis of tuberculosis. First, a convolutional neural network of suitable complexity is designed, trained, tested and validated on the tuberculosis chest radiography image sequences. The resulting artificial intelligence models could then be refined and packaged into modules for the automated detection of tuberculosis in chest radiography images and could form part of a comprehensive artificial intelligence-driven framework for the detection, prediction, diagnosis and management of a wide variety of health conditions that could play a crucial clinical decision support role.

Keywords: Artificial Intelligence (AI), Convolutional Neural Network (CNN), TensorFlow, Healthcare System, Automated Disease Diagnosis and Prediction, Tuberculosis.

Abstract: Career success is not just about working hard; it’s about working smart. If I could go back and give my younger self career advice, it would be to focus on strategic growth rather than just technical expertise. Early in my career as a network engineer, I believed that dedication and technical skills alone would lead to success. However, I soon realized that professional advancement requires more than just long hours—it demands relationship-building, self-advocacy, and a willingness to take risks. This reflection highlights the crucial lessons I wish I had learned earlier: the importance of networking, the necessity of speaking up for one’s career growth, and the value of embracing change. By adopting these principles, young professionals can accelerate their careers and navigate the corporate world more effectively.

Keywords: Career advice, networking, self-advocacy, professional growth, risk-taking, corporate success, career development, working smart

Keywords: Artificial Intelligence, Cloud Computing, AI-as-a-Service, Resource Allocation, Scalability

Abstract: Square waves play a crucial role in signal processing, digital electronics, control systems, and communication protocols. However, resampling these signals to meet different hardware specifications can introduce spectral distortions and aliasing effects. This study uses the Continuous Wavelet Transform (CWT) to examine the impact of resampling on square wave characteristics, offering superior time-frequency resolution compared to traditional Fourier-based methods. A square wave is generated using harmonic summation and initially sampled at 100 Hz before being resampled to 200 Hz for hardware compatibility. The resampled signal is then analysed using CWT scalograms, Power Spectral Density (PSD) analysis, and waveform comparisons to assess spectral distortions. Results show that CWT provides a detailed understanding of the transient and frequency variations caused by resampling, ensuring optimal signal fidelity. This research highlights the importance of advanced time-frequency analysis techniques in maintaining signal integrity across different sampling rates, with applications in real-time signal processing and embedded system design.

Keywords: MATLAB, Fourier Series, Scalogram, Odd Harmonics, Continuous Wavelet Transform (CWT), FFT

Abstract: In the rapidly changing world of web application development, one problem stays forever the same: scalability. That’s particularly the case for large-scale enterprises that want to boost their technological agility and responsiveness. Large-scale is where the problems of monolithic architectures (both frontend and backend) hit you hardest. This paper looks at one possible way to work around those problems: micro front ends. The idea behind micro front ends is to apply the same principles of decomposition (scalability, maintainability, and team autonomy) that work so well when you use them on backend services. If the decomposition of backend services allows for independent deployment, development, and scaling of those services, then the same should be possible with frontend components. This paper synthesizes the current body of literature and real-world case studies from companies like IKEA, DAZN, and HelloFresh that have successfully implemented micro front end architectures. This paper provides a detailed analysis of these companies' performance, comparing key metrics like load times, times from initial commit to deployed code, and resource utilization, before and after the companies adopted micro front ends. The verdict on micro front ends seems to be deployment frequency has improved, error rates have been reduced, and developer satisfaction has increased. In addition, this paper talk about the architectural details and the strategic thinking required to move to micro front ends, covering the not-so-simple task of bringing together multiple frontend teams and making sure users have a coherent experience across a palette of differently behaving interfaces. Specific examples show how several happily-noted-as-solved problems have been handled by various organizations, and the two main things these organizations seem to have in common is some kind of rock-solid governance model that they all sing to and more than a few advanced deployment techniques (like server-side rendering and progressive web apps) that they all seem to be employing. This paper goes through an extensive review of both academic and practical literature to evaluate micro front ends. It sets the scene with a clear overview of what micro front ends are, discussing the architecture's components and how they work together, and working through some simple examples to make clear the kind of problems micro front ends might solve.

Keywords: Micro Front Ends, Scalability, Web Application Development, Modular Architecture, Software Engineering, Frontend Decomposition, Agile Development, Enterprise Applications

Abstract: Mobile health (mHealth) and artificial intelligence (AI) technologies are now being integrated to transform the healthcare through service delivery improvement, patient self-management and operational activity. The Convert Assist helps the sales agent to recommend personalized pathology tests, and the Copilot app empowers COPD patients to manage exacerbation. Despite these challenges of adoption barriers, IT integration, and user resistance, such tools are achieving success. Implementation of digital health solutions is dependent on institutional support, has a structured training, and workflow alignment to maximize the usage of the digital health solutions.

Keywords: Co-Pilot, Intelligence, Organization, Microsoft

Abstract: This Project provides a comprehensive overview of contactless charging devices, exploring its principles, forms, Types, applications, advantages, and challenges. it focuses on designing and building a wireless charging device that can be powered by a battery source, also known as a contactless charging device. The transmitter unit (input) and the receiver unit (output) make up this gadget. The device receives a 220V AC mains signal as input, which is rectified or converted to a DC signal. The gadget thus has two power sources (battery and mains). The backup batteries are then charged with this DC signal, which is also fed into a 48 kHz oscillating circuit. This circuit uses a square wave signal to create an alternating signal, which is then passed through an induction coil to create an electromagnetic field that transfers power. This is electromagnetic induction’s basic idea. The coil at the receiver, which is placed in the path of this field is induced with the alternating current (AC) which is then rectified again to DC before it’s passed through a 5V bulk converter circuit which regulates the output of the device to 5V which can be able to power different portable devices efficiently. Additionally, the device is noiseless and the output is taken through the USB port provided.

Keywords: Contact-less Charging, Transmitter, Receiver, Batteries, Power, DC Signal

Abstract: This paper is a case study of a project done at a manufacturing company wherein the project goal was to do Integration between Oracle Ebusiness Suite and Seibel CRM. This paper talks about the approach taken in this integration

Abstract: The Mechanical Training Platform of the Graduate Association is an innovative web platform to improve the interaction between graduates, students and academic institutions. This system is registered by graduates, interacts with other graduates and students, participates in the event at the Academy, and continues to recognize the academy news. Students can study the opportunities that graduates exhibit their work and review the text model based on whisker to summarize the explanation. Administrators can manage data for graduates and students who have approved registration and events and update the contents of academy such as events, news, and galleries. This platform integrates two improved machine learning models. A text model for text viewing to combine a chatbot for processing a user request based on LSTM with a task description. This portal provides services to several users, including administrators, graduates, students and guests. Graduates and students can talk about and communicate with chat and reservations, and potential users can get information about the academy by exploring the gallery, news and interactions of chatbots. The administrator processes the Backend process, such as graduates management, student registration, event requests and business presentations. The main design of the system creates an effective and optimized experience for all users related to automatic functions such as chatbots, text reviews and email notifications. In addition, this system uses a stable mechanism for authentication and view to ensure security to protect user data and interactions. The project strengthens the relationship between graduates and students, encourages further interactions between graduates, students, and ALMA, as well as providing a digital platform without problems.

Keywords: Alumni Network, Bidirectional and Auto-Regressive Transformers , Admin, Student ,User, Long Short-Term Memory, Chat box .

Abstract: This project focuses on developing a Cyber Security Detection System that utilizes various machine learning models to classify network traffic as either normal or malicious. The system preprocesses network traffic data, performs feature analysis, and trains models to detect different types of attacks. Key features include dataset handling, where network traffic data is read and pre processed, followed by feature engineering that examines categorical variables such as protocol type, login success, and attack distribution. The project implements several machine learning models, including Gaussian Naive Bayes, Decision Tree, Random Forest, Support Vector Machine (SVM), Logistic Regression, Gradient Boosting Classifier, and Artificial Neural Networks (ANN). Performance analysis of the models reveals high accuracy, with the best model achieving a training accuracy of 99.88% and a testing accuracy of 99.88%. The classification report shows excellent precision, recall, and F1-scores for various attack types, including Denial of Service (DoS) and normal traffic, both achieving 100%. Although detection rates for U2R attacks are lower due to fewer samples, the system demonstrates significant overall effectiveness in identifying other attack types such as Probe, R2L, and DoS attacks. Additionally, the system includes user management features, such as user registration with OTP verification, admin approval for login, and admin notifications for detected attacks. The system also offers user profile management, real-time attack detection input, and a feedback system to improve overall performance. Admins can analyze feedback to further enhance the system.

Keywords: Cyber Security, Machine Learning, Network Traffic, Intrusion Detection, Feature Engineering

Abstract: The rapid expansion of the Internet of Things (IoT) has led to the development of large-scale, energy-constrained networks where efficient and secure data transmission is crucial. Traditional routing protocols, such as the Routing Protocol for Low-Power and Lossy Networks (RPL), often suffer from high energy consumption, increased routing overhead, and vulnerability to security threats. To address these challenges, this paper proposes the Energy-Aware Routing Algorithm (EARA), a hybrid routing approach that integrates adaptive energy-efficient path selection, trust-based cooperative security, and optimized data forwarding mechanisms. The proposed method dynamically selects energy-efficient routes while mitigating security threats through cooperative trust evaluation. Simulations conducted in the Cooja simulator with 100 to 500 nodes in both static and dynamic environments demonstrate the effectiveness of EARA. The results show that EARA improves Packet Delivery Ratio (PDR) by 15-25%, reduces End-to-End Delay by 10-20%, lowers energy consumption by 20-30%, minimizes routing overhead by 15-25%, and increases network throughput by 10-18% compared to Standard RPL, Trust-Based RPL, and Secure RPL. These findings highlight EARA as a promising solution for sustainable IoT applications, including smart cities, industrial automation, healthcare monitoring, and environmental sensing. By balancing energy efficiency, security, and scalability, EARA enhances the longevity and reliability of IoT networks, making it a viable approach for next-generation IoT deployments.

Keywords: Green IoT, Energy-Aware Routing, RPL, Trust-Based Routing, Secure IoT, Sustainable IoT Networks, Low-Power and Lossy Networks (LLNs), Cooperative Routing, IoT Security, Adaptive Routing, Network Lifetime Optimization, Smart Cities, Industrial IoT, Contiki OS, Cooja Simulator

Abstract: Digitization speed in our current era shows mental health problems growing as a vital dilemma which particularly affects workers in demanding professional roles. The AI-Driven Mental Health Diagnosis platform performs mental health condition predictions through advanced machine learning algorithms which process user inputs as well as analyse behavioural patterns together with historical data. The system creates comprehensive mental well-being assessments through its ability to evaluate formatted data and free-form information about symptoms with added lifestyle conditions and workplace stress elements. The platform achieves this through exploratory data analysis methods which both reveal dominant patterns and danger elements behind mental health deteriorations. The predictive system uses multiple machine learning approaches which combine logistic regression with decision trees and random forests and neural networks for mental health condition diagnosis and prediction tasks. The AI-powered chatbot receives support from natural language processing (NLP) through Google Dialog flow to offer immediate relaxation techniques and music suggestions and yoga exercises to users. The platform allows users to track their mental health progress through an interactive control centre that provides AI-generated personalized reports and downloadable assessments. The system activates automated alert systems together with individual recommendations to simultaneously detect mental health warning indicators while implementing prompt assistance for better well-being. The platform combines AI analytics with chatbot equivalent and interactive tracking capabilities to establish itself as a groundbreaking instrument which boosts mental health education while providing active time-based help for those needing assistance.

Keywords: AI-driven mental health diagnosis, machine learning algorithms, mental well-being, behavioural patterns, exploratory data analysis (EDA), predictive modelling, natural language processing (NLP), real-time support, personalized recommendations, interactive dashboard, mental health tracking, early intervention, stress management, workplace well-being.

Abstract: Natural disasters such as floods, earthquakes, and tsunamis pose significant risks to human life and infrastructure. Early warning systems play a crucial role in minimizing the damage caused by such events. This project presents the Machine Learning Disaster Prediction System, an AI-powered platform designed to predict and analyze natural disasters. Using ma- chine learning algorithms, the system predicts the likelihood of disasters based on historical data, weather patterns, and seismic activity. The system is further enhanced by integrating real-time weather data from external APIs, which improves the accuracy of predictions. The platform features a user assistance system powered by natural language processing (NLP) to identify distress signals and connect users with emergency services. Additionally, an API-based chatbot extracts the latest disaster-related news and alerts, providing users with up-to-date information on current and predicted disasters. The system allows for secure user registration and feedback through OTP-based verification and admin approval processes, ensuring a safe and reliable environment for users. The project combines Python, Django, and various APIs to create a comprehensive disaster management tool that offers early warnings, facilitates user assistance, and contributes to better disaster preparedness and response.

Keywords: Machine Learning, Disaster Prediction, NLP, Weather API, Chatbot, Disaster Management, Early Warning, Seismic Activity, Flood Prediction, Earthquake Prediction, Tsunami Prediction

Abstract: This project introduces a An interactive job and internship platform developed using the MERN stack, consisting of modules tailored for Admins, Jobseekers/Students, and Recruiters/Companies. The system aims to revolutionize the recruitment landscape by addressing inefficiencies in traditional hiring processes and providing a seamless, user-centric platform. By integrating key features such as personalized job recommendations, online assessments, and real-time application tracking, the portal ensures enhanced transparency and efficiency for all stakeholders. The Admin module offers robust management capabilities, allowing administrators to oversee user registrations, job postings, and recruiter activities. This ensures data integrity, system security, and smooth platform operations. Jobseekers, through the User module, can create detailed profiles, search for jobs aligned with their skills, and receive recommendations tailored to their qualifications and preferences. Additionally, they can participate in online exams designed by recruiters, monitor application statuses, and gain real-time updates about their progress in the recruitment process. The Recruiter module equips companies with tools to post job openings, create and manage online assessments, and evaluate candidate profiles effectively. Recruiters can make informed hiring decisions based on exam results and candidate qualifications, streamlining the selection process. The development framework used MERN stack which integrates MongoDB alongside Express.js with React.js and Node.js —this system leverages modern web technologies to deliver high performance, scalability, and user-friendly interfaces. It caters to the diverse needs of jobseekers and recruiters, offering features such as skill-based recommendations and live exams that make recruitment faster and more efficient. The project not only addresses the need for an integrated hiring solution but also highlights the importance of adopting innovative technologies in modern recruitment practices.

Keywords: Admin, Jobseeker/Student, Recruiter/Company, MERN Stack, Online Assessments

Abstract: This research focuses on detecting forged images using a Convolutional Neural Network (CNN) for classification and a Dual-Stream UNet (D-UNet) for localizing manipulated regions. The system leverages Error Level Analysis (ELA) and Spatial Rich Model (SRM) filters to enhance forgery detection accuracy. The proposed approach provides a probability score for authenticity and highlights tampered areas, ensuring a robust and interpretable forgery detection framework With the increasing accessibility of digital image editing tools, image forgery has become a significant concern in various fields, including journalism, forensics, and security. This paper presents an advanced approach to detecting image forgery using deep learning techniques, particularly Convolutional Neural Networks (CNNs). Our method incorporates both traditional forgery detection techniques such as Error Level Analysis (ELA) and Frequency Domain Analysis, along with a dual-stream U-Net model. The first stream processes raw RGB images, while the second stream analyzes filtered images using Spatial Rich Model (SRM) features to capture subtle inconsistencies introduced during forgery. The combined feature representations are then used for classification, distinguishing between authentic and tampered images. Experimental results on benchmark datasets, including CASIA and Co Mo Fo D, demonstrate that our approach outperforms existing methods in terms of accuracy, precision, and recall. The proposed method not only enhances forgery detection capabilities but also contributes to the ongoing efforts in ensuring digital image integrity.

Keywords: Image Forgery Detection, Convolutional Neural Networks, U-Net, Error Level Analysis, Spatial Rich Model, Digital Forensics.

Abstract: The Online Voting System using Blockchain with Ethereum and Machine Learning is a decentralized and secure digital voting platform aimed at ensuring transparency and integrity in elections. By utilizing blockchain technology, specifically Ethereum with Ganache, this system guarantees immutable storage of votes, eliminating any possibility of manipulation. To enhance voter authentication, the system incorporates a face recognition module, which verifies a voter’s identity before allowing them to cast their vote. The voter registration process is managed by an administrator, who can add multiple voters through bulk data uploads, including images. Election and candidate management are also handled by the administrator, ensuring structured election processes. Once a voter casts their vote, it is permanently recorded on the Ethereum blockchain, preventing any unauthorized alterations. The system enforces a strict one-vote-per-voter rule, ensuring a fair electoral process. The election results are securely displayed after voting concludes, providing an unbiased outcome. Additionally, machine learning algorithms such as Decision Tree, Random Forest, and Logistic Regression are integrated to predict future election trends. These models analyze historical election data based on multiple factors, including candidate demographics, financial assets, liabilities, and voter behavior, providing insightful forecasts. The proposed system employs Python with Django for backend development, while the frontend is built using HTML, CSS, JavaScript, and Bootstrap. By combining blockchain technology for secure voting, face recognition for fraud prevention, and machine learning for predictive analytics, this system enhances trust in digital elections, promoting a fair and transparent democratic process.

Keywords: Face recognition, Blockchain, Django, Web development, Machine Learning

Abstract: Convolutional neural networks, or CNNs, are now the backbone of medical image processing and have revolutionized the interpretation and application of different medical data in clinical image, video,- decision-making in different classifications With a focus on significant advancements, cutting-edge trends, and enduring difficulties in the area, this survey study describes the investigation of CNN-based architectures for medical image processing. The study began with basic models of CNN, such as LeNet, AlexNet, VGG, and ResNet, before heading to the advanced architectures used in DenseNet, U-Net, and Vision Transformers (ViTs). From these architectures, the discussion reflects their applications to medical image tasks such as disease classification, organ and lesion segmentation, and anomaly detection that cut across imaging modalities like Pathology, Colonoscopy MRI, CT scans. The survey article provides a broad overview of Convolutional Neural Networks (CNNs), focusing on their applications in medical imaging. It demonstrates how various forms of CNN architectures are used for the interpretation of different types of medical imaging data such as x-ray, CT, MRI and ultrasound images. The paper covers the developments in CNN methods and their capability in analyzing complex medical data sets and performing tasks such as disease identification, organ delineation and abnormality recognition. In this regard, the survey gives an explanation of the use of CNNs in medical images, and those features provide possibilities for predicting changes in the course of the disease and improve the results of treatment.

Keywords: Deep -learning, Medical Image, CNN Architectures, Image Classification.

Abstract: The integration of Artificial Intelligence (AI), data analytics, and cloud computing in healthcare has revolutionized Medicaid services, predictive analytics, interoperability, and workforce training. This paper explores scalable AI solutions, machine learning applications, and cloud-enabled healthcare advancements. It emphasizes AI-driven predictive analytics for Electronic Health Record (EHR) management and blockchain-enabled data interoperability in Medicaid systems. The study highlights challenges in AI ethics, operational barriers, and security concerns. By leveraging AI-powered decision-making and IoT-enabled smart healthcare frameworks, Medicaid optimization enhances accessibility, cost efficiency, and population health management. Future directions in AI-driven healthcare, including robotic automation, generative AI, and real-time predictive analytics, present opportunities to further streamline Medicaid operations and improve patient outcomes.

Keywords: Artificial Intelligence (AI), Data Analytics, Cloud Computing, Medicaid Optimization, Predictive Analytics, Blockchain Interoperability, Electronic Health Records (EHR), Federated Learning, IoT in Healthcare, AI Ethics, Healthcare Decision-Making, Machine Learning, Security Challenges, Population Health Management, AI-driven Automation.

Abstract: This research presents two integrated systems designed to extract and summarize information from videos and text. The first system, titled AI Based Video Insights Generator, leverages deep learning techniques, particularly Long Short-Term Memory (LSTM) networks, for detecting themes from video and textual content. This approach incorporates speech-to-text transcription, timestamp extraction from videos, and an interactive question-answering capability. Additionally, the system supports multilingual theme detection, enabling translations via APIs.

Abstract: This project focuses on developing a Smart Inventory and Sales Analytics platform for an online marketplace. The system utilizes advanced technologies such as machine learning (ML), deep learning (DL), and optical character recognition (OCR) to streamline various operations like user registration, order management, product recommendations, and feedback analysis. The platform aims to enhance the experience for both users and administrators, offering secure login options, efficient order tracking, and personalized product suggestions. The admin section of the platform provides a comprehensive dashboard for monitoring key metrics such as total users, products, and user feedback. Advanced NLP models (BERT and ALBERT) are used to analyze feedback and gain valuable insights into marketplace trends. Additionally, the admin has the ability to manage user permissions and track orders seamlessly. The platform also integrates OCR technology for Aadhaar card data extraction, simplifying the identity verification process. For users, the platform provides a smooth registration and login process using fingerprint authentication, ensuring secure access. Users can browse products, manage their cart, and leave feedback, which is analyzed by the admin for sentiment insights. The platform also includes a chatbot, powered by AI, which provides personalized product recommendations and reviews based on user preferences. Key technologies include deep learning for fingerprint recognition and OCR, machine learning for product recommendations, and natural language processing for feedback analysis. The combination of these technologies ensures that both the users’ and administrators’ needs are met efficiently while maintaining security and automation.

Keywords: Smart Inventory, Sales Analytics, Machine Learning, Deep Learning, Optical Character Recognition, Feedback Analysis, Product Recommendations, Fingerprint Authentication, Aadhaar Data Extraction, Chatbot, NLP.

Abstract: This paper aims to present a new model to ease group travel by integrating the traveler and travel agency. The model works by uniting people with matching travel goals so that they can form groups effortlessly. Therefore, travel agencies can now provide specific groups with customized rental cars and travel packages. The above system tackles issues faced in conventional travel booking and increases user satisfaction while also enabling agencies to profit from it. This review details the problem definition, exploration, goals, architectural design, methods used, conclusions drawn, pros, and cons, and the possible future of the platform.

Keywords: Group travel , Traveler , Travel Agency , Model , Travel Goals , Customize rental cars , travel packages, Conventional travel booking , User satisfaction , Profit , System.

Abstract: Hand sign recognition is an innovative application of artificial intelligence that enables machines to interpret and understand human gestures. This project aims to develop an AI-powered Hand Sign Recognition System using deep learning techniques, particularly Convolutional Neural Networks (CNNs). The system will be trained on a dataset of hand gestures, allowing it to accurately classify and recognize different signs in real time. The project follows a structured workflow, including data collection, preprocessing, model training, and real-time recognition using a webcam. OpenCV is used for image processing, while TensorFlow/Keras handles model training and inference. Transfer learning techniques with pre-trained models such as MobileNetV2 or ResNet50 improve accuracy and efficiency. The system has applications in sign language interpretation, gesture-based human-computer interaction, and accessibility solutions for differently-abled individuals. Additionally, it can be extended to control devices using hand gestures, enhancing user experience in gaming, virtual reality, and robotics. By integrating AI with computer vision, this project demonstrates a practical and impactful approach to bridging the gap between human communication and machine understanding.

Keywords: Hand Sign Recognition,Artificial Intelligence (AI) , Deep Learning , Convolutional Neural Networks (CNN) Gesture Recognition , Sign Language Interpretation, Computer Vision , OpenCV , TensorFlow/Keras , Real-Time Processing Human-Computer Interaction (HCI) , Machine Learning , Transfer Learning , Image Classification.

Abstract: With the rising impact of extreme heat due to climate change, individuals exposed to high temperatures for prolonged periods such as outdoor workers, athletes, travelers, and the elderly are at significant risk of sunstroke and other heat-related illnesses. Sunstroke can cause severe dehydration, organ failure, and even death if not detected early. To address this, we propose the system called “RAY SHIELD-COMPREHENSIVE SUNSTROKE DETECTION IN HEAT EXPOSURE”. It is an advanced wearable device designed to monitor environmental and physiological parameters in real-time to prevent sunstroke. The system integrates multiple sensors to track UV exposure levels (240-370 nm), body temperature, and humidity. These data points are processed using an onboard microcontroller to assess the risk of heat-related illnesses. Upon detecting unsafe thresholds, the device activates an alert mechanism, which includes a buzzer alarm system, SMS alerts, and call notifications to designated emergency contacts. The SMS alerts contain critical information about the user’s condition, ensuring timely intervention. The primary goal of this project is to develop a functional and reliable prototype for individuals exposed to extreme heat, such as outdoor workers, athletes, and military personnel. This innovation enhances safety by preventing dehydration and sunstroke through proactive monitoring and alerting mechanisms.

Keywords: Sunstroke Detection; UV Sensor; Temperature sensor

Abstract: In today's fast-paced digital world, data centers are at the heart of modern businesses, processing huge amounts of information around the clock. As organizations transition from conventional on-site environments to cloud-based solutions, the focus is on enhancing system efficiency, security, and scalability. This paper explores the evolution of data centers and key challenges like virtualization, standardization, and the growing shift toward hybrid cloud models. One of the most difficult aspects of data center administration is keeping computers cool while remaining efficient. We describe a simple method for calculating the appropriate cooling capacity while taking into account the power of IT equipment, climatic conditions, and future growth requirements. In addition, we evaluate traditional and cloud data centers to assist organizations in making better IT decisions. We have applied a proposed methodology to identify total power consumption by using the total wattage of all IT equipment, including servers, storage devices, and networking equipment. By leveraging new technologies and making the most of their resources, companies can create a data management strategy that’s reliable, cost-effective, and ready for the future.

Keywords: Virtualization, Hybrid cloud, Cooling Capacity, Data center, Networking.

Abstract: GesturaX introduces a novel approach to touch-free computing by combining two interactive technologies— GestureSense and AirMouse—to redefine how users engage with digital systems. The GestureSense component employs advanced image processing techniques and machine learning algorithms, powered by the MediaPipe framework, to detect and interpret hand gestures in real time. This allows users to perform everyday tasks like moving cursors, clicking, scrolling, adjusting volume/brightness, and controlling media playback through intuitive hand motions. The AirMouse module complements this by blending hardware innovation with motion sensing. Built around an ESP32-WROOM microcontroller and Bluetooth technology, it uses an MPU6050 inertial sensor for precise cursor control, alongside physical buttons and a rotary dial for clicks and scrolling. A compact 3.7V battery paired with a tp5100 charging unit and integrated safety systems ensures reliable, cord-free operation. By merging visual gesture recognition with motion-based hardware controls, GesturaX offers a flexible solution for scenarios where traditional keyboards/mice fall short—from accessibility tools to gaming interfaces and smart home systems. This dual-modality design prioritizes both precision and ease of use, paving the way for more natural human- device interactions.

Keywords: Human-Computer Interaction, Gesture Recognition, Computer Vision, MediaPipe, ESP32, MPU6050, Bluetooth Connectivity, Touchless Interface, Sensor Fusion, Real-Time Control

Abstract: The concept of Total Team Quality represents a transformative approach to quality management within Scrum teams, moving away from traditional siloed testing and quality assurance methods toward a more integrated, collaborative model where quality becomes everyone's responsibility. This comprehensive analysis demonstrates that when properly implemented, Total Team Quality practices lead to significant improvements in both product quality and team velocity, with case studies showing up to 40% reduction in User Acceptance Testing (UAT) defects and duration alongside 20% increases in velocity. The approach centres on shift-left testing principles, shortened feedback cycles, reduced work-in-progress limits, and the cultivation of high-performing teams where quality is embedded into every phase of the development process rather than being an afterthought or separate activity.

Keywords: User Acceptance Testing (UAT), Scrum teams, SonarQube, work in progress (WIP), Test-Driven Development (TDD), Behaviour-Driven Development (BDD)

Abstract: The "Tours and Travel System" is an intelligent and automated travel management tool designed to streamline the process of organizing trips, managing tour bookings, customer interactions, and travel planning. Traditional travel agencies rely heavily on manual processing, which requires them to browse multiple travel websites, track price changes, and update clients regarding itineraries. This method is highly inefficient, prone to errors, and time-consuming. Our project introduces a robust system that automates these tasks by integrating real-time web scraping, dynamic pricing updates, and structured itinerary generation. The system fetches relevant travel details using Selenium WebDriver, processes information using Spring Boot, and converts it into structured Excel and JSON formats using Apache POI for improved accessibility and analysis. This automation reduces manual intervention, enhances accuracy, minimizes workload, and optimizes the efficiency of travel agencies and customers alike.

Keywords: Tour Management, Web Scraping, Dynamic Travel Pricing, Automated Booking, Spring Boot, Java Servlets, Apache POI, Travel Planning Optimization.

Abstract: As remote collaboration and online learning continue to evolve, traditional video conferencing tools often lack the interactivity and engagement needed for truly immersive experiences. VirtuVista, a web-based virtual reality (VR) platform, addresses these limitations by providing interactive, browser-based virtual environments that enhance real-time communication and collaboration. Unlike traditional VR systems that require specialized hardware, VirtuVista leverages Three.js, WebXR API, AngularJS, and Socket.io to deliver immersive experiences accessible on standard devices like desktops, laptops, and mobile phones. Users can navigate virtual spaces, engage in spatial audio conversations, and interact in a way that closely resembles physical presence. This research paper explores the development, implementation, and impact of VirtuVista in transforming digital communication. By merging web-based VR with real-time interaction, the platform offers a scalable and cost-effective alternative to conventional online meeting solutions, paving the way for the future of virtual collaboration.

Keywords: Virtual Reality (VR), Real-Time Communication, 3D Graphics, Collaborative spaces, AngularJS, WebXR API, Socket.io, Web Application, Three.js.

Abstract: The Smart Door Locking System is an innovative security solution to enhance safety and user convenience. By utilizing Internet of Things (IoT) technology, and mobile app controls, this system provides an advanced method for accessing doors. Users can unlock doors via smartphones, RFID tag removing the need for traditional keys. Features include time-based access control, real-time notifications for unauthorized entry attempts, and the ability to remotely lock or unlock doors. The system can also be integrated with other smart home devices, creating a fully automated and secure home environment. This system effectively addresses the limitations of traditional locking mechanisms while offering a modern, secure, and user-friendly alternative for residential, office, and commercial spaces.

Keywords: (Smart Door Locking System, ESP32, Arduino IDE, RFID Scanner, Servo Motor, Buzzer, LCD Display, Jumper Wires, Power Bank)

Abstract: The goal of this project is to create an integrated platform that will enable students from different colleges and universities to effectively manage and present their academic work. By giving students, a single location to upload, monitor, and edit their work, the platform will help with issues like project collaboration, documentation, and presentation. Peer review processes to improve learning through feedback, team collaboration tools, and structured project submission with version control are important features. In order to accommodate a wide range of academic disciplines, the platform will support several project domains, such as engineering, computer science, management, and more. For advice and assessment, a mentorship program will pair students with academics and business professionals. The platform will also provide a project repository where finished products can be consulted for further research and development. A step forward

Keywords: Project Collaboration, Academic Projects, Mentorship, Progress Tracking, Digital Repository

Abstract: Memristors, as the fourth fundamental circuit element, have garnered significant attention for their potential applications in non-volatile memory, neuromorphic computing, and arithmetic logic circuits. This paper presents the design and implementation of an original memristor model in LTSpice XVII, focusing on its electrical characteristics, behavior under different stimuli, and potential integration into modern electronic circuits. The study involves modeling the memristor using its fundamental equations, simulating its resistance switching characteristics, and analyzing its hysteresis behavior under AC and DC excitations. The simulation results provide insights into the nonlinear dynamics and memory-dependent behavior of the memristor, highlighting its feasibility for use in next-generation computing systems. This research contributes to the understanding and practical realization of memristor-based circuits, paving the way for innovative applications in low-power and high-density memory and logic designs. Index terms: Memristor, LTSpcie XVII ,Non-Volatile Memory,CMOS.

Abstract: The rapid growth of e-commerce has transformed the way consumers shop and businesses operate. This paper presents the design, development, and implementation of a Flipkart clone using the MERN (MongoDB, Express.js, React.js, and Node.js) stack. The project replicates key features of Flipkart, including user authentication, product catalog management, shopping cart functionality, order processing, and payment integration. The paper provides a detailed insight into system architecture, development challenges, and future enhancements, contributing to the broader knowledge base of full-stack e-commerce applications.

Keywords: MERN Stack, E-Commerce, Flipkart Clone, Web Development, Full-Stack Development, Online Shopping

Abstract: Pricing strategies are essential for optimizing revenue, profitability, and customer happiness in the fiercely competitive retail sector of today. The goal of this project is to create a machine learning-based price optimization model that will allow merchants to identify the best prices for their products by examining a number of influencing factors, such as market circumstances, competition pricing, demand trends, and historical sales data. The suggested solution makes use of predictive analytics to comprehend how pricing and demand elasticity are related, determining the price points that optimize profits without offending clients. To forecast sales success at various price points and suggest the most lucrative pricing strategies, sophisticated regression algorithms like Gradient Boosting (XGBoost) are used. To improve the accuracy of the model, feature engineering will take into account consumer segmentation, inventory levels, promotions, and seasonality. The model will be trained and validated using publicly available or retailer-provided data, and its performance will be evaluated using metrics such as Mean Absolute Error (MAE) and Revenue Growth Rate (RGR). One of the top priorities will be creating a dynamic and adaptable system that can respond to changes in the market in real time. The expected outcome is a data-driven pricing strategy that helps businesses increase profit margins, reduce inventory costs, and improve customer retention. This initiative may have practical benefits for retail chains, e-commerce sites, and other consumer-focused firms seeking to enhance their pricing tactics.

Keywords: Demand prediction, price optimization, data driven machine learning, retailing.

Abstract: Deep learning, in particular Convolutional Neural Networks (CNNs), have begun to serve as a great asset for improving many aspects in oncology including cancer detection, diagnosis, and treatment. This survey paper presents an overview of the works that employed CNN-based techniques towards the early detection of different cancers i.e. breast, lung, prostate and skin cancer. We investigate the application of CNN on medical image processing, primarily for radiographic imaging, including CTs, MRIs, and histopathological sections. Paper considers the actual studies devoted to the development of new CNN architectures, image preprocessing techniques, and transfer learning approaches aimed at increasing the cancer detection systems accuracy and efficiency. Nevertheless, a number of issues still need to be resolved, such as the high expense of acquiring high-quality data, the inability of deep learning models to be interpreted, and the requirement for big annotated datasets. Additionally, the survey article shows how CNNs could be used to increase the accuracy of cancer diagnosis when combined with other machine learning and imaging methods like multimodal imaging and genomics. Finally, the survey discusses the direction of subsequent research in the use of CNNs in oncology, including applying clinical workflows, diagnostics, and precision medicine in all its aspects.

Keywords: Deep - learning, Oncology, CNN Architectures, Classification, Cancer Detection, pre trained CNN Network

Abstract: In today's fast-paced world, road accidents remain a major concern, leading to severe injuries and fatalities due to delayed emergency response. The Crash Alerting and Detecting System is an IoT-based solution that automatically detects vehicular accidents through airbag deployment sensors and real-time GPS tracking, promptly notifying emergency services and designated contacts. The system integrates microcontrollers, GSM modules, and cloud-based storage to ensure seamless data transmission. By leveraging IoT technology, this project aims to enhance road safety, reduce emergency response time, and provide real-time accident reports, ultimately saving lives.

Keywords: Crash Detection, IoT, Real-time GPS Tracking, Automated Alert System

Abstract: Face recognition attendance systems have gained popularity in recent years as they offer an efficient and secure method of monitoring employee attendance. This system can help in reducing errors and increasing efficiency as compared to traditional manual methods of taking attendance. The system makes use of computer vision technology to detect and recognize the faces of employees and record their attendance automatically Our system consists of two main parts: face detection and recognition. The first step is to detect faces in an image using a pre-trained deep learning model. The model used for this task is the Single Shot Detector (SSD) model, which is trained on the COCO dataset. The model detects faces in an image and draws bounding boxes around them. The second step is face recognition, which involves comparing the detected faces with a pre-existing database of employee faces. For this task, the system uses the FaceNet model, which is trained on a large dataset of faces and can generate a high-dimensional feature vector for each face. The feature vectors are then compared using the cosine similarity measure to determine if a given face matches a face in the database. Our system also includes a user interface that allows administrators to view attendance records and add or remove student details from the database. The interface is built using the PyQt5 library and provides an easy-to- use graphical user interface. Our system has several advantages over traditional attendance systems. It eliminates the need for manual entry, reducing the chances of errors and fraud. It also saves time by automating the attendance process and reduces the workload of administrative staff. Furthermore, it provides enhanced security by preventing unauthorized access to the attendance records.

Keywords: Face Recognition, Attendance System, AI Algorithms, Automation, Scalable Solution

Abstract: The QR-Based Attendance System is a smart and hassle-free way to track attendance without using outdated manual methods. Instead of signing a register or calling out names, this system allows students or employees to scan a unique QR code to mark their attendance. The scanning can be done using a mobile phone or a web-based app, making the process fast, accurate, and effortless.Once scanned, the system instantly records the data in a secure database, ensuring that attendance records are well-organized and easily accessible. It also prevents proxy attendance (when someone else marks attendance for another person) by verifying unique codes and logging the exact time and location of the scan. With cloud storage integration, administrators and teachers can access attendance records anytime, anywhere, making monitoring more flexible and efficient.The system can also generate detailed reports and analytics, helping institutions and organizations track attendance trends over time. Additionally, it can send automated notifications to students, employees, or managers in case of frequent absences. The system is designed to be user-friendly, requiring minimal technical knowledge to operate.This project is especially useful for schools, colleges, offices, and events where attendance tracking is essential. By replacing traditional methods with QR code technology, it makes the attendance process quicker, more reliable, and transparent. It also reduces paperwork, saves time, and improves overall efficiency.With its ability to eliminate errors, prevent fraud, and offer real-time updates, this QRBased Attendance System is a modern, tech-savvy solution for smarter attendance management.

Keywords: Face Recognition, Attendance Management, AI Algorithms, Automation, Scalable Solution

Abstract: This project proposes an IOT-based smart public transportation system for buses and bus stations. The system aims to enhance the efficiency, safety and passenger experience of public transportation. The proposed system integrates various IOT sensors and technologies, such as GPS, RFID and WI-FI, to provide real –time monitoring and automation of bus operations.The rapid urbanization and increasing population density in cities have led to a pressing need for more efficient and user-friendly public transportation systems. This paper proposes an IOT-based automatic smart public transport bus and bus station system designed to enhance the reliability, safety and convenience of urban transit.

Keywords: • IOT and Technology 1. IOT (Internet of Things) 2. Smart Transportation 3. Intelligent Transport System (ITS) 4. Real-Time Data Analytics 5. Wireless Communication (Wi-Fi, RFID, etc.) 6. Sensor Networks 7. GPS Tracking 8. Automation

Abstract: Diabetic retinopathy is a severe eye disease that is fast spreading all over the world. It arises when blood sugar levels rise, leading to problems with the kidneys, eyes, and heart. Diabetic Retinopathy (DR) is an eye disease that is caused by the breakdown of blood vessels in the retina, which occurs as diabetes progresses. It is thought to be the main cause of visual impairment because it progresses asymptomatically at the early phases. This review article discusses the methods adopted in diabetic retinopathy detection, segmentation, and classification using deep and machine learning algorithms, and discusses their importance and limitations, along with potential future directions to overcome these limitations.

Keywords: Diabetic Retinopathy, Deep Learning, CNN, Accuracy.

Abstract: As we store more and more data in the cloud, like photos and schoolwork, it's super important to keep that information safe and private. Big data analysis helps us find patterns and learn cool things from this data, but we need to do it without peeking at anyone's personal stuff. This project looks at ways to build special computer programs, called algorithms, that let us analyze big groups of data without seeing the details of any one person's information. We'll explore techniques like making the data a little fuzzy (don't worry, it still works for finding patterns!), or using secret codes to keep things hidden. By using these privacy-preserving algorithms, we can use the power of big data while keeping everyone's information safe and sound in the cloud.

Keywords: Privacy, Big Data, Cloud Computing, Data Security, Encryption, Privacy Protection Techniques

Abstract: This project “Harmful Content Detection on Social Media Platform “is Python based project. It is designed using PYTHON/FLASK as front end and PHP as backend. The web application for the detection of offensive word is used to find the offensive Word and it performs a comparison between the words stored in the database and the words present in the text. The system then shows the user if any offensive words are detected. It shows the offensive and non-offensive words in graphical representation like chart, bar graph to find the presence of offensive word in the text. The proposed system is tested on a dataset of offensive words, and the results show that it can effectively detect offensive words in offline mode. Harmful Content Detection On Social Media Platforms implements our coded, machine learning algorithms, in finding a negative comment from the messages it receives by a user. The algorithm first gives the message a value and then based on our pre trained data, it decides if the comment is harsh enough to be transformed or not. It is assigned a value and if the value results in a positive sentence, the system will proceed to send the transformed positive sentence to the end user. Otherwise, the sentence will be placed through the models again. The users communicate through a developed web front face and they are connected to a central server. The users are termed as clients. If any messages are modified the receiving user will be notified along with the modified message. A major source of cyberbullying is social media. These platforms can have the opposite desired effect of uniting peers, and instead can be weaponized to harass and bully others. Most existing solutions have shown techniques/approaches to detect cyberbullying, but they are not freely available for end-users to use. They haven’t considered the evolution of language which makes a big impact on cyberbullying text. It doesn’t affect only for health, there are more different aspects which will lead life to a threat. Cyberbullying is a worldwide modern phenomenon which humans cannot avoid hundred percent but can be prevented.

Keywords: Harmful Content Detection, Social Media Moderation, Offensive Word Detection, Cyberbullying Prevention, Machine Learning Algorithms, Sentiment Analysis, Natural Language Processing (NLP), Flask Web Application, PHP Backend, Text Classification, Data Filtering, Content Moderation, Automated Censorship.

Abstract: This study extensively examines implementing Management Information Systems (MIS) across various sectors, focusing on the complex digital transformation processes and their effects on organizational performance. By thoroughly analyzing three leading organizations—a multinational manufacturing firm, a healthcare system, and a retail company—this research highlights the essential success factors, obstacles, and results associated with adopting advanced MIS. The results indicate that effective digital transformation necessitates a nuanced combination of technological advancements, organizational change management, and strategic coherence. This research enhances both the theoretical framework and practical application of MIS in today's business landscape.

Keywords: Management Information Systems, Digital Transformation, Enterprise Resource Planning, Healthcare Informatics, Business Intelligence, Industry 4.0, Artificial Intelligence, Case Study

Abstract: Landslides present significant risks to infrastructure, economies, and human safety, requiring advanced detection and predictive mapping strategies. This study explores the integration of deep learning and remote sensing techniques to enhance landslide identification. Utilizing Sentinel-2 multispectral imagery and ALOS PALSAR-derived slope and Digital Elevation Model (DEM) data, the research examines critical environmental factors such as vegetation cover, rainfall, and terrain features. Additionally, various geospatial analysis techniques are evaluated to determine their effectiveness in improving detection accuracy. The findings contribute to the advancement of early warning systems, disaster risk management, and sustainable land-use planning, fostering more reliable and scalable landslide prediction models.

Keywords: - Image Processing, Machine Learning, Deep Learning, Computer Vision, Remote Sensing.

Abstract: In every nation, democratic voting is an important and serious event. Currently, ballot papers or electronic voting machines are used for voting. These procedures have several disadvantages, including a lack of transparency, low voter turnout, vote tampering, mistrust of the electoral authority, voter ID card forgeries, result delays, and security concerns. Security is always the top priority when considering putting a digital voting system in place. There can be no question regarding the system's capacity to protect data and fend off possible attacks when such important choices are on the line. Blockchain technology is one possible solution to the security problems. There are countless uses for blockchain technology. Blockchain is a distributed ledger technology that facilitates peer-to-peer, decentralized network transactions involving digital assets. One interesting development in this area is distributed ledger technology. A block is a grouping of every transaction. Immutability, decentralization, security, transparency, and anonymity are some of the key characteristics of blockchain technology. A promising option for creating more transparent, safe, and secure electronic voting systems is blockchain technology with smart contracts. In this paper, we have used the Solidity language and blockchain technology to implement and test a sample e-voting application as a smart contract for the Ethereum network using wallets. To prevent vote duplication, a limited quantity of tokens (gas) are provided in the wallet and depleted when the user casts their ballot. In addition to outlining the benefits and drawbacks of utilizing blockchain technology, this paper presents a workable system by highlighting a voting web app and its restrictions.

Keywords: E-voting, Smart-contracts, Blockchain, Ethereum.

Abstract: The ICU Patient Monitoring System is an innovative healthcare solution designed to continuously monitor critical patient parameters such as heart rate, oxygen saturation (SpO2), and body temperature. The system integrates Raspberry Pi, medical sensors, a Flask-based web application, an SQLite database, and an automated alert system (SMS, email, voice calls) to enhance patient care and emergency response. This paper details the progress, challenges, and future enhancements of the project. The ultimate goal is to provide a cost-effective, scalable, and real-time monitoring solution for hospitals and healthcare institutions, reducing manual intervention and ensuring rapid response to critical conditions.

Keywords: ICU Patient Monitoring, Raspberry Pi, Medical Sensors, Flask Web Application, SQLite Database, Automated Alert System, Heart Rate Monitoring, SpO2 Measurement, ECG Sensor, Temperature Sensor.

Abstract: The integration of Artificial Intelligence (AI) in agriculture is revolutionizing farming practices by enhancing productivity, optimizing resources, and enabling precision farming. This project explores the use of AI technologies to improve agricultural processes, focusing on key tools, models, and steps involved. This project includes precision farming, crop health monitoring, yield prediction, market insights. The project utilizes machine learning algorithms such as decision trees, support vector machines (SVM), and deep learning models, data collection. Data inputs include environmental variables, soil conditions, crop health, weather patterns. Through the application of these AI models, the system can predict crop yields, detect diseases, optimize irrigation schedules, and recommend fertilizers and pesticides. The output includes actionable insights for farmers, providing them with precise recommendations to enhance crop management, reduce costs, and increase sustainability. The project demonstrates how AI can drive innovation in agriculture, ultimately improving food security and farming efficiency.

Keywords: Artificial Intelligence (AI), Precision Farming, Crop Health Monitoring, Yield Prediction, Market Insights, Environmental Variables, Soil Conditions, Weather Patterns, Disease Detection, Sustainability, Farming Efficiency.

Abstract: The rising number of missing child cases globally highlights the urgent need for a more efficient and intelligent identification and recovery system. Conventional methods, including manual tracking and public awareness initiatives, often fall short due to time limitations and insufficient data management. This research proposes a comprehensive Missing Child Identification System that leverages deep learning, facial recognition, and big data analytics to enhance identification accuracy and operational efficiency. By employing convolutional neural networks (CNNs) and transfer learning, the system compares images of missing children with those in existing databases and surveillance footage. It also incorporates biometric data, such as facial embeddings and age progression algorithms, to adapt to changes in appearance over time. Additionally, the system features an AI-powered alert mechanism that promptly notifies law enforcement and relevant authorities when a match is identified. Real-time analysis and pattern recognition capabilities significantly reduce search times and improve recovery rates. The system’s scalable architecture allows seamless integration with existing surveillance networks and law enforcement databases, making it a viable solution for large-scale deployments. Furthermore, the implementation of privacy-preserving techniques ensures data security and compliance with legal standards. Experimental evaluations validate the system's effectiveness, demonstrating robust performance with high precision and recall rates in diverse scenarios. This study presents a scalable and intelligent solution designed to expedite the recovery of missing children, addressing the limitations of traditional investigative approaches while offering a proactive and efficient response mechanism.

Keywords: Missing Child Identification, Facial Recognition, Deep Learning, Convolutional Neural Networks, Transfer Learning, AI-Powered Alert System, Child Recovery System.

Abstract: Emotion recognition from speech signals has gained significant attention in human-computer interaction, offering applications in entertainment, mental health monitoring, and personalized user experiences. This paper presents a web-based Speech Emotion Recognition and Music Recommendation System that utilizes deep learning for emotion classification and integrates music streaming services for personalized recommendations. The system records speech input, extracts Mel-Frequency Cepstral Coefficients (MFCC) as features, and classifies emotions using a pre-trained Convolutional Neural Network (CNN) model. Based on the detected emotion, the system retrieves genre-specific music recommendations from Spotify. Implemented using Flask, TensorFlow, and Librosa, the proposed approach achieves efficient real-time emotion classification and enhances user engagement through tailored music selection. Experimental results demonstrate the model’s accuracy and the effectiveness of the recommendation system.

Keywords: Speech Emotion Recognition (SER), Deep Learning, Convolutional Neural Networks (CNN), Mel-Frequency Cepstral Coefficients (MFCC), Natural Language Processing (NLP), Audio Signal Processing, Flask Web Application, Music Recommendation System, Spotify API, Human-Computer Interaction (HCI).

Abstract: The proposed blockchain-based organ donation and transplantation system enhances efficiency, security, and transparency in managing organ donations. Traditional organ donation frameworks depend on centralized databases, which are prone to unauthorized access, data breaches, and inefficiencies in organ allocation. To overcome these challenges, this system leverages a private Ethereum blockchain, ensuring secure, immutable, and decentralized management of donor and recipient records. Smart contracts, developed using Solidity, automate key processes such as donor registration, recipient matching, and organ allocation. By eliminating manual intervention, the system ensures fairness and accuracy. Medical institutions can securely register donors and recipients, with every transaction permanently recorded on a tamper-proof ledger, promoting transparency. Once a suitable match is identified, real-time notifications alert all relevant stakeholders, enabling recipients to track their organ request status through a hospital-issued unique identification number. The platform employs advanced cryptographic methods and role-based access controls to safeguard sensitive medical data, restricting access to authorized personnel only. Additionally, integrated data visualization tools offer insights into organ donation patterns, aiding policymakers in making informed decisions.By incorporating blockchain technology and smart contracts, this system introduces a decentralized, trust-driven approach to organ donation, ensuring data integrity, optimizing efficiency, and reinforcing security. It represents a transformative advancement in medical technology, setting a new benchmark for ethical and transparent organ transplantation processes.

Keywords: Organ Allocation Integrity, Blockchain based system, Smart Contracts, Data Security, Decentralized System, Ethereum blockchain

Abstract: Inherited retinal diseases in children can lead to blindness and diagnosing them is difficult due to the many possible causes. Current diagnostic methods are complex and sometimes invasive, making them unsuitable for young children. This research introduces a new system to help diagnose these diseases using a technique called Chromatic Pupillometry, which measures how the pupil reacts to different colours of light. The new system combines a special pupillometer device with a computer program that uses machine learning. The program analyses the pupillometry data and helps doctors determine if a child has an inherited retinal disease. Specifically, they tested the system on Retinitis Pigmentosa, a type of inherited retinal disease. The results were promising, showing good accuracy, sensitivity (correctly identifying those with the disease), and specificity (correctly identifying those without the disease). This is the first time machine learning has been used with pupillometry to diagnose a genetic disease in children.

Keywords: Machine Learning, Clinical decision support system, Python, Pupillometry, Retinopathy, Support Vector Machine, ELM (Ensemble Extreme Learning Machine), pigmentosa.

Abstract: Agro Craft is an innovative e-commerce platform designed to connect farmers, artisans, and suppliers of agricultural tools and pesticides. The website provides a seamless marketplace for buying and selling agricultural products, integrating a user-friendly interface with secure payment gateways and efficient logistics. A unique feature of Agro Craft is the tutorial module, which educates users on website navigation, product listing, order placement, and account management. Additionally, the platform integrates an AI-powered chatbot that provides real-time weather updates and crop recommendations based on location, helping farmers make informed decisions. This project aims to bridge the gap between rural producers and urban consumers, enhancing market accessibility and promoting sustainable agricultural practices.

Keywords: E-commerce, Agriculture, Online Marketplace, Digital Farming, Rural Empowerment, AI Chatbot

Abstract: E-mail is a widely used method of communication, serving both professional and social purposes. However, for blind users, traditional email applications present significant challenges, particularly since they must rely on memory to navigate through the keyboard or keypad. The proposed system aims to address this issue by offering a voice-driven email platform, enabling blind users to send, receive, and view emails without needing to memorize a keyboard layout. This voice-based system provides access to various email folders, such as the inbox, sent items, and trash, all through spoken commands. The integration of a speech recognition model helps in recognizing both the language and voice of the user. By transcribing spoken words into written text accurately, the system allows users to interact with their email accounts hands-free. The process of converting speech to text is known as "speech-to-text" or "automatic speech recognition" (ASR). Additionally, Google Text-to-Speech (GTTS) technology is employed to convert the input text into audible speech. Two important protocols, SMTP and IMAP, are used in email communication: SMTP handles the sending of emails, while IMAP is responsible for retrieving and managing messages. The outcome of this work is the development of the “Blind-Friendly Email System” prototype.

Keywords: google text to speech (GTTS), Speech recognition, speech to text (STT), Conversational Email Assistant, Voice-Controlled Inbox Navigation

Abstract: Deepfake technology has become a serious concern due to its potential misuse in misinformation, fraud, and privacy violations. Traditional detection methods struggle to keep up with increasingly sophisticated fake videos. This project leverages deep learning and computer vision techniques to detect DeepFake content in images and videos using the XceptionNet model. The system processes images and videos by extracting frames, preprocessing them, and passing them through a trained Xception model to classify them as real or fake. The video classification is based on majority voting of analyzed frames. The application is built using Streamlit for an interactive user interface, enabling users to upload and analyze media in real-time.Future improvements include optimizing model inference, enhancing dataset diversity, and integrating real-time DeepFake detection for live streaming applications.

Keywords: DeepFake Detection, XceptionNet, Deep Learning, Computer Vision, Image Processing, Video Processing, Frame Extraction, Streamlit, Realtime Detection, Majority Voting, Fake Video Classification, Media Analysis, Model Inference Optimization, Dataset Diversity, Live Streaming Applications.

Abstract: Diesel-powered engines are major contributors to CO₂ emissions, driving environmental pollution and climate change. Predicting and mitigating these emissions is essential for improving fuel efficiency and minimizing environmental impact. Over a 10-month period (October 2022 to December 2023), this project aims to deliver actionable insights. This study applies machine learning models to analyse and predict CO₂ emissions using historical data. The models, including Linear Regression, Random Forest, and XGBoost, are trained on engine parameters, fuel traits, and operational data to generate accurate predictions. Key input variables, such as engine load, fuel consumption, and temperature, are processed to provide real-time emission estimates, categorized as low, moderate, or high. This approach enhances diesel engine efficiency and enables industries, researchers, and policymakers to make informed, data-driven decisions for reducing carbon footprints. Through AI-driven methods, the project advances sustainability by offering precise, actionable guidance for emission control and regulatory compliance. This initiative fosters decarbonization, balancing environmental responsibility with operational efficiency.

Keywords: CO₂ emissions, diesel engines, machine learning, emission prediction, fuel consumption, engine load, temperature, Linear Regression, Random Forest, XGBoost, sustainability, emission control, regulatory compliance.

Abstract: This research presents RFCNN, a hybrid machine learning and deep learning framework for traffic accident severity prediction using decision-level fusion. The proposed approach combines Random Forest (RF) for feature selection and Convolutional Neural Networks (CNN) for deep feature extraction, followed by ensemble-based classification. The model leverages full and selected feature sets to improve predictive accuracy while addressing challenges like high-dimensional data and class imbalance. Experimental results on real-world accident datasets demonstrate that RFCNN outperforms traditional machine learning models (e.g., AdaBoost, Gradient Boosting, and Voting Classifiers) in terms of accuracy, precision, recall, and F1-score. The system includes a user-friendly GUI for data preprocessing, model training, and performance visualization. The study highlights the effectiveness of feature selection and model fusion in enhancing accident severity prediction, contributing to improved road safety analytics.

Keywords: Traffic Accident Severity Prediction, Machine Learning (ML), Deep Learning (DL), Random Forest (RF), Convolutional Neural Network (CNN), Feature Selection, Ensemble Learning, Decision-Level Fusion, Road Safety Analytics, Predictive Modeling

Abstract: RideTogether is a community-driven ride-sharing platform designed to tackle urban challenges such as traffic congestion, high commuting costs, and environmental concerns. By integrating carpooling, bike-sharing, and alternative transport options, it offers an affordable and eco-friendly commuting solution. Unlike traditional ride-hailing services, RideTogether prioritizes user verification, AI-powered ride matching, and secure digital transactions to enhance safety and efficiency. It supports multiple transport modes, including cars, bikes, scooters, and public transit, while features like a built-in wallet, automated fare splitting, and a carbon footprint tracker promote seamless payments and sustainable travel choices. Safety measures such as ID verification, live tracking, and an SOS emergency button ensure a secure ride-sharing experience. Additionally, the platform fosters community engagement through workplace carpools, college rideshares, and local ride groups, supported by an admin dashboard for ride analytics, user management, and policy enforcement. Future enhancements, including EV ride-sharing, AI-driven fare optimization, and loyalty rewards, aim to further improve user engagement. By reducing commuting costs, easing traffic congestion, and fostering trust-based ride-sharing, RideTogether aspires to revolutionize urban mobility with a scalable and efficient transportation solution.

Keywords: Ride-sharing, Carpooling, Urban Mobility, Sustainable Transportation, Digital Wallet, Fare Splitting, User Verification, Community-driven Transport, Alternative Transport Modes, EV Ride-sharing, Safety Features, Live Tracking, Carbon Footprint Reduction.

Abstract: Health Pixel is an innovative web-based platform that utilizes artificial intelligence to perform comprehensive medical image analysis. By employing advanced convolutional neural networks (CNNs), the platform offers users quick and non-invasive health insights by analyzing images from four key body areas: the tongue, eyes, nails, and skin. This research highlights the platform's unique approach to enhancing preventive healthcare by leveraging intelligent image recognition techniques and providing personalized health recommendations based on visual biomarkers. The proposed multi-modal diagnostic system effectively bridges critical gaps in early disease detection and continuous health monitoring by delivering accessible, technology-driven medical insights.

Keywords: Artificial Intelligence, Medical Image Analysis, Convolutional Neural Networks, Preventive Healthcare, Multi-Modal Diagnostics.

Abstract: In recent years, there has been a growing demand for intelligent transportation systems (ITS) to enhance road safety, traffic management, and law enforcement. This paper proposes an efficient Vehicle Monitoring System. The system comprises four primary components: vehicle speed estimation, number plate recognition, Deblurring number plate and moving vehicle detection.For vehicle speed estimation, by tracking vehicle movement and applying optical flow techniques, the system accurately calculates the speed of each passing vehicle in real-time.In parallel, the system integrates a Number Plate Detection module. Upon detecting vehicles within the camera's field of view, the system extracts their number plates using advanced object detection algorithms.The proposed method utilizes advanced image deblurring techniques to restore the clarity of vehicle number plates in blurred images. Initially, the system detects and extracts candidate regions containing number plates using state-of-the-art object detection algorithms. Also the proposed approach utilizes a convolutional neural network (CNN) architecture to detect and localize moving vehicles in video streams. By leveraging the temporal information inherent in consecutive frames, the system accurately distinguishes between static background elements and dynamic objects, such as vehicles in motion.Overall, the proposed system offers a robust and scalable solution for real-time vehicle speed and number plate detection, contributing to enhanced road safety, traffic management, and law enforcement in urban environments. Index Terms: YOLO (You Only Look Once), CNN (Convolutional Neural Network), OCR(Optical Character Recognition).

Abstract: The Stock Price prediction is a critical area of financial analytics that aims to forecast stock prices based on historical data patterns. This project, Stock Price Prediction Model combines Deep learning, web development, and cloud deployment to provide an interactive platform for stock price forecasting The system utilizes Long Short Term Memory (LSTM) neural networks for time series prediction, ensuring high accuracy in forecasting stock trends.This system not only recommends stocks but also provides detailed research information about companies, helping users make informed investment decisions By integrating Deep Learning (ML), Full Stack Web Development, and Financial Data Analysis, our platform offers a seamless and intelligent stock selection experience.

Keywords: Stock Price, Deep Learning, LSTM, Financial forecasting, Stock Recommendation

Abstract: SmartMedConnect is an innovative, web-based platform designed to streamline the process of booking medical appointments, connecting patients and healthcare providers in a seamless, efficient, and secure manner. Built using the MERN stack (MongoDB, Express, React, Node.js), SmartMedConnect offers a modern solution to the growing demand for digital healthcare services. The platform enables patients to easily browse through a wide range of doctor profiles, check their availability, and schedule appointments at their convenience. Patients can also securely make payments for consultations, leave feedback, and access their appointment history via a personalized dashboard. On the other hand, healthcare providers (doctors) can register, manage their professional details, set their availability, and list consultation fees. Doctors can efficiently manage appointments, communicate with patients, and maintain a comprehensive profile. SmartMedConnect ensures the highest level of security, utilizing bcryptjs for encrypting passwords, ensuring that sensitive patient and provider data remains protected. The platform’s intuitive design and responsive interface cater to a wide range of users, providing access across devices. With key features such as secure payment processing, real-time appointment scheduling, user feedback, and scalability, SmartMedConnect aims to enhance the accessibility, efficiency, and quality of healthcare services. By leveraging cutting-edge technology, SmartMedConnect bridges the gap between patients and doctors, offering a secure, user-friendly, and scalable solution for modern healthcare management.

Keywords: Medical Appointment Platform, Patient-Doctor Communication, Online Appointment Scheduling,Secure Payment Gateway, User Feedback System, Healthcare Data Security, Web-Based Healthcare Solution

Abstract: Budget Buddy is an AI-powered finance tracking application that simplifies personal financial management while prioritizing user security and privacy. Developed with Next.js, Tailwind CSS, and Supabase (Prisma), it offers a seamless experience for tracking expenses and managing budgets. Secure authentication is handled by Clerk, allowing users to maintain multiple accounts and set a default account for automatic transaction entries. Transactions are categorized as income or expense, with attributes such as description, recurrence, and receipt data extraction powered by Google Generative AI API, which enables automatic transaction detail filling from scanned receipts. The platform includes data visualization tools like bar graphs and pie charts, helping users gain insights into their financial activity. Users can set monthly budgets and receive email notifications when expenditures exceed 90% of the budget, along with a financial report on the 1st of each month, enriched by Google AI with tailored suggestions to stay on track. Budget Buddy also ensures platform security through Arcjet’s rate-limiting (10 transactions per day) and bot protection middleware, which blocks malicious bots while permitting trusted ones like search engines and Inngest. The app supports recurring transactions, automatically recording them at specified intervals, and allows users to edit, delete, filter, and search transactions based on preferences. With advanced AI and security features, Budget Buddy enables users to track finances effectively, gain actionable insights, and achieve financial goals securely.

Keywords: AI-Powered Finance Tracking, Personal Financial Management, Secure Authentication, Expense and Budget Tracking, Google Generative AI, Data Visualization in Finance

Abstract: The study "Regression Analysis on Financial Statements at Konigtronics Pvt Ltd" investigates the role of regression analysis in improving financial statement comparability and effectiveness. The study focuses on identifying important profitability factors, estimating future financial performance, and defining the most often analysed financial accounts. Using primary data from surveys and secondary data from firm records, the study does multiple regression analyses to identify key profitability predictors. The key findings show that quantifying correlations between financial variables is the most successful way (40%) for identifying profitability drivers, followed by analysing cash flow patterns (25%), and comparing industry benchmarks (20%). Cash flow statements are the most often analysed financial statements (25%). The study emphasises the need of using standardised regression models and conducting regular financial analyses to improve financial reporting. The study suggests that regression analysis greatly improves financial statement quality and reliability, allowing for better decision-making and strategic planning. Companies that apply standardised models and invest in financial team training can produce transparent, comparable, and successful financial reporting. This study emphasises the need of correct financial data and ongoing professional growth in utilising regression approaches to improve financial outcomes.

Keywords: Regression Analysis, Statistical tools, Forecasting, Decision making, profitability and Financial statement Analysis.

Abstract: The Flexy Store People Counting and Assistance System is a cutting-edge solution designed to enhance customer management, optimize operations, and improve resource allocation in retail environments. By leveraging advanced technologies such as IoT sensors, machine learning algorithms, and computer vision, the system accurately monitors the number of people entering, exiting, and moving within a store. Strategically placed cameras capture video feeds, which are analyzed using computer vision techniques to ensure precise people detection and movement tracking, while prioritizing privacy and anonymity. One of the system's key features is its ability to provide real-time occupancy updates. Store managers can monitor foot traffic patterns, identify peak hours, and analyze customer flow, enabling data-driven decisions regarding staffing, inventory placement, and store layout optimization. This valuable insight allows businesses to allocate resources effectively, minimizing overcrowding and enhancing operational efficiency. In addition to people counting, the system incorporates a dynamic customer assistance feature. By identifying areas with higher customer density or potential bottlenecks, the system alerts staff to provide immediate assistance where it is needed most. This not only ensures a smoother shopping experience but also increases customer satisfaction and loyalty. The Flexy Store system is highly scalable and adaptable, making it suitable for a wide range of retail establishments, from small shops to large retail chains. Its modular design allows seamless integration with existing infrastructure, while its robust analytics dashboard offers actionable insights that help businesses respond quickly to changing market demands. The Flexy Store system is highly scalable and adaptable, making it suitable for a wide range of retail establishments, from small shops to large retail chains. Its modular design allows seamless integration with existing infrastructure, while its robust analytics dashboard offers actionable insights that help businesses respond quickly to changing market demands. In today’s dynamic and competitive retail landscape, the Flexy Store People Counting and Assistance System stands as a vital tool for businesses seeking to maximize efficiency, improve customer experience, and achieve a competitive edge.

Keywords: Flexy Store, people counting , face reading algorithm, IoT

Abstract: The healthcare industry is experiencing a transformation because of artificial intelligence, which delivers both powerful diagnosis and prognosis and treatment design capabilities. Opacity in many AI models creates concerns about clinical decision-making transparency while also threatening trust in medical decision systems as well as ethical standards. Such issues gain particular importance in critical fields, including oncology, together with mental health treatment and personalized medical practices. The emergence of explainable AI (XAI) represents a fundamental solution to these problems by giving healthcare professionals understandable insights that show how AI systems operate. This work examines why healthcare needs XAI solutions through an explanation of various explainable methods while addressing the human-focused ethical and legal barriers to implementation. Explainable technology serves as a basic requirement to build trust because it exists as both a technological need and a social requirement and a legal essential and clinical necessity. The successful adoption of XAI into clinical settings requires proper regulatory oversight while using interdisciplinary teamwork and continuous staff training because it ensures accountable, equitable applications of AI in healthcare.

Abstract: Our initiative creates a dubbing system powered by AI to convert English audio into various Indian regional languages, thus improving accessibility. By utilizing natural language processing (NLP), speech-to-text (STT), and text-to-speech (TTS) technologies, the system captures spoken language, translates it into text, and produces audio that sounds natural when dubbed [1]. This automated solution streamlines the dubbing process, making it quicker, more affordable, and scalable, which is advantageous for sectors like education, entertainment, and business. It minimizes the need for manual dubbing while enhancing speech fluency and synchronization [2]. We tackle challenges such as timing adjustments and linguistic precision to ensure translations sound natural [3]. Future improvements could feature real-time dubbing capabilities and advanced voice cloning techniques to further enhance quality and maintain speaker consistency [4].

Keywords: AI- Powered Dubbing, Speech-to-Text, Text-to-Speech, Natural Language Processing (NLP), Multilingual accessibility

Abstract: Cloud computing has become a vital part of modern digital infrastructure, offering scalable, on-demand computing resources. However, security remains a primary concern, as cyber threats continue to evolve in complexity. Machine learning (ML) has emerged as a powerful tool in enhancing cloud security by detecting, preventing, and mitigating various cyber threats. This paper provides a detailed review of ML-based security mechanisms in cloud computing, covering key algorithms, applications, benefits, challenges, and future research directions.

Keywords: Artificial intelligence, Machine learning, Deep Learning, IOT, Cyber Security.

Abstract: Deep fake technology poses significant threats to the authenticity of digital media, leading to misinformation, reputational damage, and security risks. The ability to manipulate videos and images with AI has resulted in concerns over trustworthiness in media, cyber threats, and fraudulent activities. Traditional detection methods, including manual inspection and rule-based algorithms, have proven inadequate in identifying these rapidly evolving deep fake techniques. This project introduces a deep learning-based solution utilizing Convolutional Neural Networks (CNNs) for detailed image analysis and Recurrent Neural Networks (RNNs) for detecting temporal inconsistencies in videos. The system integrates attention mechanisms to focus on subtle artifacts and adversarial training to enhance detection robustness. Additionally, it continuously learns from new deep fake patterns, ensuring adaptability against emerging manipulation techniques. Designed for scalability and real-time performance, our system is optimized to run efficiently on standard hardware while achieving high accuracy and low false-positive rates. By providing a reliable tool for deep fake detection, this project contributes to media integrity and cybersecurity

Keywords: Deep Fake, Deep Learning, Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Adversarial Training, Image Forgery Detection, Real- Time Detection, Artificial Intelligence (AI).

Abstract: This research paper explores the development of an application leveraging fine-tuned Large Language Models (LLMs) for advanced interaction with research papers and GitHub repositories. Traditional methods often involve manual effort or siloed tools, lacking integrated insights across textual and code-based resources. Our system bridges this gap, enabling developers and researchers to analyze papers, understand implementations, and create adaptations seamlessly. By fine-tuning a pre-trained LLM on research papers and source code, and integrating Retrieval-Augmented Generation (RAG) techniques, the system delivers contextual and dynamic interactions. Unlike previous approaches focusing on isolated analysis or summarization, our unified platform combines academic and practical insights, helping users navigate complex topics, validate claims, and prototype ideas efficiently. This work addresses key challenges in integrating text and code for knowledge discovery, setting the stage for enhanced research and development workflows.

Keywords: Large Language Models, AI-Powered Code Understanding, Codebase Analysis, Intelligent Document Processing, Automated Literature Review, Contextual Code Retrieval.

Abstract: Emotion recognition from human voice has emerged as a crucial technology in various fields, including healthcare, human-computer interaction, and artificial intelligence-based applications. The ability to detect emotions based on speech signals enhances system adaptability and improves user experience. This study presents a progressive implementation of an emotion detection system that integrates Natural Language Processing (NLP) and speech feature extraction techniques. The system utilizes machine learning and deep learning models to classify emotions, including happiness, sadness, anger, and fear, based on vocal expressions. The approach involves extracting speech parameters such as pitch, tone, energy, and amplitude, which are analyzed using ML-based classifiers. Additionally, NLP techniques, including text sentiment analysis and word embedding’s, enhance classification accuracy by providing contextual insights. The system is implemented on Raspberry Pi hardware, making it portable and scalable for real-world applications. Initial findings indicate that deep learning models outperform traditional ML approaches, offering improved accuracy. Future advancements will focus on reducing background noise, optimizing feature selection, and incorporating real-time emotion tracking. Keyword-Speech Emotion Recognition, NLP, Machine Learning, Deep Learning, Speech Processing, Human-Computer Interaction.

Abstract: Effective waste management is crucial for environmental sustainability and public health. This research presents the development of a Smart Waste Segregation System using image processing and deep learning to automate the classification and sorting of waste. The system classifies waste into five categories: paper, glass, metal, plastic, and organic waste. It integrates both hardware and software components to enhance the accuracy and efficiency of waste segregation. The hardware consists of an ESP32-CAM module to capture waste images and an ESP32 development board to control the mechanical sorting system. Captured images are processed on a laptop using the MobileNetV2 deep learning model for real-time classification. Upon identification, the waste is sorted into the appropriate bin using a conveyor belt and servo motors. To ensure optimal performance, the system was tested using five deep learning models: MobileNetV2, VGG16, ResNet50, InceptionV3, and Xception. Experimental analysis revealed that MobileNetV2 offers the best balance of accuracy and computational efficiency, making it ideal for real-time waste classification. Key features of the system include automated image-based waste identification, real-time sorting, and LED indicators to display the detected waste category. This automated approach reduces human intervention, improves sorting accuracy, and increases operational efficiency. The proposed system is scalable, cost-effective, and suitable for applications in smart cities and industrial waste management, offering a sustainable and efficient solution for modern waste handling challenges.

Keywords: Smart Waste Segregation, Image Processing, Deep Learning, Automated Waste Classification, MobileNetV2, VGG16, ResNet50, InceptionV3, Xception, ESP32-CAM, Waste Management, Real-Time Sorting, Environmental Sustainability, Mechanical Automation, Smart Cities.

Abstract: Advanced digital platforms that cater to a wide range of shopping requirements have emerged as a result of the rapid expansion of e-commerce, which has altered the way consumers interact with businesses. DSTS.com is a cutting-edge online shopping site made to make shopping easy, secure, and personalized for customers. This paper focuses on DSTS.com's innovative approach to e-commerce and examines its core functionalities, technological framework, and market positioning. To increase user engagement, the platform incorporates AI-driven recommendation systems that suggest relevant products based on past browsing and purchasing patterns. Additionally, it makes use of secure payment gateways to guarantee secure and dependable transactions. Additionally, a user-friendly and intuitive interface makes it simple to navigate, making shopping enjoyable for customers of all ages. Customers can get their hands on a wide range of high-quality goods at reasonable prices thanks to the extensive product catalog that DSTS.com provides, which includes clothing, electronics, household necessities, and other categories. Additionally, the website includes effective logistics management, allowing for prompt and dependable delivery services. In addition, live assistance and chatbots powered by AI increase customer satisfaction by addressing queries and concerns in real time. This study examines DSTS.com's technological architecture, business model, and market impact to determine how it changed online shopping. The platform aims to redefine e-commerce convenience and efficiency by making use of data analytics, machine learning, and design that is focused on the customer. The paper ends with a discussion of DSTS.com's potential for the future, with a focus on ongoing innovation and adapting to changing consumer trends.

Keywords: E-commerce, marketing strategy, consumer segmentation, collection of comments from consumers, business analysis, security encryption

Abstract: The objective of this project is to create an advanced robotic system to use as a nurse in hospitals that utilizes the aspects of the Internet of Things (IoT). The primary objective is to enhance patient care with limited direct human-to-human contact, especially in the case of epidemics and pandemics, where such contact is a significant threat to both medical staff and patients.This system combines a robot, a control interface board, sensors, and software to react to real healthcare needs. The robot is designed to assist healthcare workers by automating many processes like dispensing medication, monitoring patient condition, and offering assistance to healthcare workers.The robot mechanism is equipped with an online management command to control and monitor remotely. Seamless connectivity is enabled by a Wi-Fi-supported controller with effective data transfer between the robot and the medical staff. Integration of sensors enables the robot to capture key patient data such as temperature, heart rate, oxygen saturation, and other life-critical parameters, which are transmitted to the doctors and nurses in real time.With IoT application, data processing, and robotic automation, this system greatly enhances efficiency in hospitals. It alleviates the burden of the medical staff, increases timely treatment of patients, and minimizes opportunities for infection during health crises. The proposed robotic solution will be a smart and efficient aid in the healthcare industry, thereby enhancing the quality of medical care and patient safety

Abstract: The convergence of Artificial Intelligence (AI) and Machine Learning (ML) with cloud computing has catalyzed transformative advancements across various sectors by delivering scalable, efficient, and cost-effective solutions. This journal examines the ramifications of AI and ML within cloud environments, elucidating the benefits, challenges, applications, and emerging trends in this dynamic landscape.

Abstract: This study introduces a YouTube-inspired video streaming platform with an AI assistant that enables users to upload, watch, and control video content. The platform includes an AI chatbot driven by OpenAI's ChatGPT API in addition to allowing user authentication, channel creation, video uploads, and subscriptions. By responding to inquiries, suggesting videos, and enhancing content discoverability, the AI assistant raises user engagement. PHP, MySQL, HTML, CSS, JavaScript, and AI technology are all used in the development of this project to improve user experience. The design, development, and effects of this system are examined in this paper along with a comparison to more conventional video streaming services.

Keywords: USB to TTL UART Serial Converter, ADS1115, MS51FB9AE, etc.

Abstract: In modern computing systems, fast and efficient arithmetic operations are essential for enhancing overall performance. The Carry Look-Ahead Adder (CLA) is widely used due to its reduced propagation delay compared to conventional ripple-carry adders. However, further optimization is required to improve speed, power efficiency, and circuit complexity. This paper presents a novel design of an optimized CLA using hybrid logic, integrating CMOS technology with a custom-designed memristor model implemented in LTSpice XVII. The proposed approach leverages the low-power characteristics of memristors while maintaining the robustness and switching reliability of CMOS technology. By utilizing Voltage-Controlled Resistors (VCRs) as memristive elements, the design achieves significant reductions in propagation delay and power consumption.Extensive simulations and performance evaluations demonstrate the superiority of the proposed CLA in terms of speed and energy efficiency compared to conventional CLAs. The results indicate that hybrid CMOS-memristor logic can be a promising approach for designing next-generation arithmetic circuits. This study provides valuable insights into the practical implementation of hybrid logic circuits and establishes a foundation for future research in memristor-based arithmetic units.

Keywords: Memrisor, CLA, LTSpice XVII, Hybrid Logic

Abstract: Through a digital platform that improves accessibility, transparency, and efficiency in the agricultural supply chain, the Farmers Network Website Megaproject aims to bridge the gap between farmers, consumers, and agricultural stakeholders. This paper discusses the platform's development, implementation, and impact, focusing on how it connects farmers and potential buyers, provides real-time agricultural insights, and streamlines logistics throughout the supply chain. The study provides empirical results on the platform's adoption and economic benefits and evaluates the platform's effectiveness using various data-driven metrics. By providing a digital platform that makes it easier for farmers, consumers, and agricultural stakeholders to interact directly, the Farmers Network Website Megaproject hopes to revolutionize agricultural trade and communication. The primary objective is to make the market more accessible, to streamline the supply chain, and to guarantee transaction transparency. An in-depth look at the platform's development, features, and approaches to implementing various technological solutions are provided in this paper. A comprehensive empirical evaluation of its impact is also included in the study, highlighting important improvements in price predictability, buyer engagement, and logistics efficiency. The Farmers Network Website is a scalable and effective model for digital agricultural marketplaces because it incorporates cutting-edge technologies like machine learning, blockchain, and cloud computing.

Keywords: Farmers' Networks, Agricultural Productivity, Network Participation, Social Capital, Digital Platforms, ICT Tools in Agriculture, Technology Adoption, Cooperative Marketing

Abstract: Breach detection helps in identifying unauthorized access or suspicious activities in a network system. It detects threats at an early stage before they cause serious damage, protecting sensitive information like personal data, financial records, and confidential files. Breach detection systems provide real-time alerts, allowing security teams to take quick action and prevent further harm. However, most existing breach detection systems face challenges like false alarms, slow detection speed, and inability to detect new attacks. Some systems also struggle to detect intrusions in encrypted data and consume high system resources. These limitations affect the accuracy and performance of the detection system. To overcome these issues, the breach detection system is implemented using Temporal Fusion Transformers (TFT), which analyses time-based patterns in network traffic to detect intrusions accurately. The current study incorporates the Simargyl2022 dataset to enhance the quality of our results and analyses, which contains both normal network traffic and malicious attack data, making it suitable for evaluating detection performance. The system achieved 95.40 accuracy, with a recall of 95.40, precision of 91.01, and an F1-score of 93.15, showing its high efficiency in detecting breaches. The outcomes of this study have significant implications for network security, providing valuable insights for practitioners and researchers working towards building robust and intelligent breach detection systems. Key Words: Breach Detection, Temporal Fusion Transformer (TFT), Cybersecurity, Anomaly Detection, Time-Series Forecasting, Temporal Dependencies, Multi-Headed Attention, Gating Layers, Scalable Systems, Advanced Deep Learning.

Abstract: The integration of digital technology and Artificial Intelligence (AI) into healthcare systems has immense potential to enhance doctor availability, streamline appointment scheduling, and improve patient outcomes. Our application is designed to facilitate efficient healthcare access by providing comprehensive information about doctors, including details such as name, qualifications, specialties, location, and nearby medical facilities. Patients can easily access emergency medical service contacts, locate blood banks, and schedule appointments with specific doctors based on their needs. The application also provides a centralized view of appointment statuses, showing completed, upcoming, and canceled appointments for better patient management. Additionally, a search functionality enables patients to find specialists tailored to their specific health concerns. In the next phase, we are working on AI integration to incorporate a Virtual Doctor Assistant. This feature will allow for disease prediction based on patient-reported symptoms, offering preliminary diagnosis and personalized recommendations. These include primary medication suggestions, dietary advice, and preventive measures, ensuring immediate assistance in the absence of doctors. This AI- driven addition aims to empower patients with proactive healthcare guidance, [1] reduce patient wait times, and facilitate informed medical decision-making, marking a significant step toward smart healthcare delivery.

Abstract: Accurate human body measurement is essential in the clothing and fashion industry to improve sizing accuracy and reduce product returns. Traditional measurement methods require expensive 3D scanners or manual tape measurements, which are time-consuming and less accessible. This paper presents a comprehensive analysis of AI-powered body measurement systems using a single RGB camera. By leveraging computer vision and machine learning techniques, this study examines body detection, anthropometric estimation, and elliptical modeling for circumference calculations. Additionally, we discuss challenges in measurement accuracy, ethical considerations, and future research directions. The findings contribute to the ongoing efforts to standardize body measurement techniques in the fashion industry, ensuring inclusivity, sustainability, and technological advancement in garment sizing. Furthermore, we explore the social, economic, and technological implications of AI-based measurement techniques in different industries. Index Terms: Human body measurement, Machine learning, Computer vision, Fashion technology, Anthropometry, Sizing standardization, Privacy concerns, Inclusivity, AI ethics.

Abstract: Farmers are the backbone of the Indian economy but farmers are involved in age-old practices, poor use of resources and financial risk. AgriGyan is one AI based web app developed with the intention of combining the latest technology with the age old farming practices and bridge this gap. It is a platform that provides features such as field mapping, crop planning, expenses and income, and real time advisory services. This paper presents the system design of AgriGyan, the approaches taken, and effectiveness analysis of AgriGyan, along with the role of AgriGyan in enhancing agriculture productivity and environmental sustainability. But despite their crucial role, many farmers are still relying on centuries-old practices that are outdated, unsustainable and poorly suited to 21st-century challenges. These old methodologies, combined with lack of access to advanced tools and technologies, lead to bad resource usage, low productivity, and a huge financial risk.

Keywords: Smart Agriculture, Data Visualization, Recommendations, Agriculture Technology, Sustainable Farming, Data-Driven Farming. Etc.

Abstract: Cross-Domain Recommendation Systems (CDRS) enhance traditional recommendation models by transferring knowledge across different domains, thereby improving the personalization of suggested content. The integration of Explainable AI (XAI) ensures transparency in recommendation systems, addressing concerns regarding trust and interpretability. Additionally, sentiment analysis is crucial in refining recommendations by capturing the emotions embedded in user reviews and feedback. This paper explores the concept of a sentiment-aware and explainable AI-based cross-domain recommendation system by discussing key methodologies, challenges, and potential advancements in the field. The research highlights how a hybrid recommendation model can optimize user satisfaction by balancing accuracy, interpretability, and personalized content delivery.

Keywords: Cross-Domain Recommendation, Explainable AI, Sentiment Analysis, Hybrid Recommendation Model, User Trust.

Abstract: Sustainable waste management solutions, particularly for EV batteries and electrical components, are in high demand due to the growing popularity of electric vehicles (EVs). Our Go-Lang-based platform provides a comprehensive digital solution to efficiently manage and recycle EV waste as a response to this challenge. The platform's core functions, technological framework, and impact on the environment are examined in this paper, highlighting its role in sustainable waste disposal. The system includes real-time map-based waste collection, allowing EV users to effortlessly schedule pickups for recyclable materials in order to increase user engagement. Optimized route planning for collection agents also reduces carbon emissions and ensures faster and more effective waste retrieval. Transparency and accountability during the recycling process are enhanced by data-driven analytics and secure transactions. Users of all backgrounds can easily navigate and interact with the platform thanks to its intuitive and user-friendly interface. It ensures that discarded materials undergo proper processing, minimizing environmental harm, by connecting EV owners with authorized showrooms and recycling centers. Additionally, future waste collection strategies can be improved with the help of AI-powered predictive analytics. This study demonstrates the platform's potential to transform EV waste recycling by examining its technological architecture, business model, and impact on sustainable waste management. Insights into possible future developments are provided at the paper's conclusion, with an emphasis on continuous innovation and adapting to changing environmental regulations.

Keywords: Electric Vehicle Recycling, Waste Management, Sustainable Disposal, Go-Lang Application, Route Optimization, AI in Waste Collection, Environmental Protection.

Abstract: In today's competitive employment landscape, crafting a professional and impactful resume is crucial for job seekers aiming to secure interviews and career opportunities. However, many individuals face challenges in creating resumes that effectively communicate their skills, qualifications, and experiences. This paper introduces an AI-powered online resume builder developed using PHP, MySQL, HTML, CSS, JavaScript, and the ChatGPT API. The system is designed to streamline the resume-building process by offering real-time suggestions, intelligent content recommendations, and formatting assistance tailored to specific job roles and industry standards. By integrating the capabilities of ChatGPT, the system enables users to receive personalized guidance in crafting compelling resumes with optimized content. The AI assistant assists in generating professional summaries, refining skill descriptions, and incorporating relevant keywords that align with the job market demands. The platform supports multiple resume templates and ensures usability for users from diverse backgrounds, including non-technical individuals. The paper details the system architecture, frontend and backend technologies, database design, and AI integration. User testing and feedback highlight significant improvements in both the quality and efficiency of resume creation compared to traditional methods. The findings support the effectiveness of AI-assisted resume generation in enhancing user satisfaction and the potential for broader applications in career development tools. Future enhancements aim to incorporate AI-driven resume scoring, multilingual support, and integration with recruitment platforms to expand the system’s functionality and accessibility.

Abstract: This system fills the communication gap between the deaf and hearing population by interpreting oral language into Indian Sign Language (ISL) video streams, making it more accessible. It begins with Whisper ASR, a transformer-based automatic speech recognition algorithm that transcribes speech to accurate text. The transcribed text is then processed with Natural Language Processing (NLP) methods, such as tokenization, part-of-speech tagging, stop-word elimination, and lemmatization, to prepare the structure for ISL translation. For additional enhancement of compatibility with ISL grammar, a Sequence-to-Sequence (Seq2Seq) model with Recurrent Neural Networks (RNNs) is employed for restructuring sentences to produce fluent and natural translations. The optimized text is then translated into pre-recorded ISL video clips, and MoviePy performs the seamless stitching and synchronization of sign segments. A web interface based on Flask offers users a simple platform to upload audio files, input text, and create ISL videos in real-time. The system is optimized for efficiency and ease of use, and it can be useful in education, healthcare, government services, and customer support. The future developments will emphasis on developing the ISL vocabulary dataset, real-time processing optimization, mobile compatibility, and cloud platforms deployment for improved scalability. These enhancements will make the system more efficient, accurate, and accessible, further improving communication for the deaf and hard-of-hearing.

Keywords: Speech-to-Sign Language, Indian Sign Language (ISL), Seq2Seq Model, Natural Language Processing (NLP), Audio-to-Video Conversion, Gesture Recognition, Whisper ASR, Deep Learning.

Abstract: This project integrates cutting-edge computer vision techniques used to enhance the medical diagnostics by simultaneously addressing two critical aspects of healthcare—blood sample analysis for leukemia and brain hemorrhage classification in MRI images. Leveraging the YOLO (You Only Look Once) algorithm, our system employs deep learning to efficiently detect and classify various stages of leukemia in blood samples. YOLO's real- time object detection capabilities enable swift identification of abnormal cells, facilitating early diagnosis and intervention. The model is trained on a comprehensive dataset, ensuring robust performance across diverse cases. In parallel, Convolutional Neural Networks (CNNs) are employed for the intricate task of brain hemorrhage classification in MRI scans. The CNN model learns complex hierarchical features from brain images, enabling it to accurately differentiate between different types and stages of hemorrhages. This dual-faceted approach aims to provide a comprehensive diagnostic tool, facilitating healthcare professionals in timely and accurate decision-making.

Abstract: Power line inspection is a critical task that requires regular monitoring and maintenance to ensure the reliability and safety of electrical distribution infrastructure. With the advancements in robotics, artificial intelligence (AI), and unmanned aerial vehicles (UAVs), integrating robotic manipulators with drones and automating their maneuvers has emerged as a promising solution for power line inspection. This paper proposes a quadcopter design and implementation with a gripper mechanism to dock automatically on a power line using AI-enabled camera feedback. The machine learning model implemented onboard will detect the power line, align the drone to it, and activate the gripper for automated perching. The drone also includes a light weight three degree of freedom (DoF) robotic manipulator with an additional camera incorporated into it for AI-assisted power line inspection. The insulator fault detection can be carried out with a deep learning model. Power line inspection begins with the take-off of the drone from the ground and its perch on the power line. After disarming the drone, the manipulator comes into action. The arm is lifted through a controlled manipulator action to focus the camera on the insulators. The video of the insulators will be shared with a server through wireless means. A custom-trained deep-learning model in the server will identify the faulty insulators.

Keywords: Power line inspection, UAVs, manipulator, degree of freedom, payload capability, deep learning.

Abstract: An innovative platform created to transform the identification and treatment of esophageal conditions by medical practitioners is the AI-Driven Esophageal Disease Diagnosis and Predictive treatment forecasting Platform. With the incorporation of deep learning models and advanced machine learning algorithms, this platform makes precise prognosis of illness progression therapy predictions, and diagnostics. Automated disease classification, recurrence prediction, and treatment outcome forecasting based on models like Convolutional Neural Networks (CNN), Random Forest (RF), and Long Short-Term Memory (LSTM) are some of its major features. Healthcare professionals can observe and understand diagnostic insights in real time due to the integration of the platform with an intuitive dashboard developed using Streamlit. Excel simplifies data management by keeping data in an easily accessible and user friendly format. By helping medical professionals develop personalized and efficient treatment plans, this artificial intelligence based technology improves patient outcomes and optimizes healthcare resource utilization.

Keywords: Deep Learning, Machine Learning, Streamlit, Healthcare AI, Esophageal disease, Diagnosis and Prediction

Abstract: This paper presents AI-PrepMate: AI-Assisted Mock Interview and Feedback System, an AI-powered mock interview application designed to enhance technical interview preparation through real-time feedback and performance evaluation. Built using React, TypeScript, Firebase, and Google Gemini AI, the system enables users to log in, create mock interviews, record responses, and receive AI-generated assessments. By leveraging natural language processing (NLP) and machine learning, the platform evaluates user responses against predefined criteria, providing structured feedback to improve interview readiness. Additionally, Firebase ensures seamless data storage and authentication via Clerk, making the system scalable and secure. The application is deployed on Firebase Hosting, ensuring real-world usability. This research explores the implementation process, technical challenges, and future enhancements to optimize AI-driven interview simulations for aspiring professionals.

Keywords: Mock interview system, Natural Language Processing (NLP), Google Gemini AI, AI-driven assessment, Automated interview feedback, Firebase cloud database, Scalable web application for interview practice AI-assisted learning and evaluation.

Abstract: Softball pitching is a complex and dynamic movement involving the integration of multiple joints, muscles, and biomechanical principles. Unlike the overhand motion in baseball, softball pitching utilizes a windmill or underhand motion that places unique stresses on the musculoskeletal system. This review aims to synthesize current biomechanical research related to softball pitching, focusing on kinematic and kinetic variables, energy transfer through the kinetic chain, muscle activation patterns, and common mechanical inefficiencies that can lead to injury. The literature indicates that the windmill pitch consists of six phases: wind-up, stride, arm rotation, release, deceleration, and follow-through. Proper sequencing of these phases is critical for performance and injury prevention. Peak angular velocities at the shoulder often exceed 5000°/s during release, emphasizing the importance of trunk-shoulder coordination. The kinetic chain concept plays a central role, as energy generated from the lower body and core must efficiently transfer to the upper limbs. Ground reaction force (GRF) data suggest the importance of stride and drive leg strength, while EMG analyses highlight high muscular demand in the deltoids, latissimus dorsi, gluteals, and rectus femoris. A common finding across studies is that mechanical deficiencies, such as early trunk rotation or inadequate stride length, increase the risk of overuse injuries, particularly in the shoulder and lumbar spine. This review also explores the impact of fatigue, developmental differences (e.g., youth vs. collegiate pitchers), and training interventions aimed at improving mechanics. Overall, this review highlights the critical role of biomechanical analysis in understanding softball pitching. By applying these insights, coaches and clinicians can better tailor strength training, corrective exercises, and pitching technique to enhance performance and reduce injury risks. Continued interdisciplinary research combining biomechanics, motor control, and sports medicine is essential for advancing the science of softball pitching.

Keywords: Biomechanics, Softball pitching, Windmill motion, Kinematic chain, Ground reaction force, Injury prevention, Muscle activation

Abstract: In this paper we studied With the exponential rise in digital audio streaming, protecting intellectual property from unauthorized use and piracy has become a critical challenge. This thesis presents a robust and imperceptible audio watermarking system designed to resist piracy through an improved hybrid technique that combines Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT). The proposed method leverages the multi-resolution analysis capability of DWT and the energy compaction property of DCT to embed a secure, imperceptible watermark into the audio signal. The watermark is embedded in the mid-frequency coefficients of selected DWT sub-bands after DCT transformation, ensuring a balance between robustness and audio quality. A pseudo-random sequence, generated using a secret key, determines the watermark’s location, enhancing security against common signal processing and malicious attacks. The system is tested under various attack scenarios, including compression, filtering, noise addition, and re-sampling, to evaluate its robustness.

Keywords: DCT, DWT, SVD, ICA, Resampling.