National Conference on Recent Trends in Engineering and Technology

Abstract: Microarray data has been widely applied to cancer classification, where the purpose is to classify and predict the category of a sample by its gene expression profile. The cancer classification is required to identify the significant genes that are a good subset of original features. For evaluating the goodness of a subset of features, the feature selection methods fall into two broad categories: the filter approach and the wrapper approach. Since wrapper methods are computationally very intensive, filter approach is chosen for selecting the most informative genes. DNA microarray is a gene chip which consists of expression levels for a huge number of genes a relatively small number of samples. However, only a small number of genes contribute in accurate classification of cancer. Therefore, the challenging task is to identify a small subset of informative genes which has maximum amount of information about the class. The feature selection method is used to find the informative genes which helps to minimize the classification errors. The hybrid correlation methods are used to find out the correlated and negative correlated features. The classifier Support Vector Machine along with Decision Tree Algorithm is proposed to classify the features. The result is compared with the performance of neural network classifier which gives better accuracy in positive correlated features than hybrid correlated and negative correlated features. Keywords: DNA Microarray, Classification, Correlation, Neural Network, Backpropagation Algorithm.

Abstract: The tremendous growth of World Wide Web made the world to joined as together. It paved the way to minimize the distance of each other. Even though it provides greater advantages to internet users, the same way it poses great insecurity to information. Web security is an important issue of concern nowadays. Web attacks are increasing day by day. Clickjacking attack is a web framing attack that has recently received wide media coverage. In a clickjacking attack, the attacker creates a malicious page such that it tricks into clicking on an element of a page that is only barely visible. It redresses the users click for malicious purposes. By stealing user�s click, they force the users to perform unintended actions, thus posing a significant threat to the Internet users. Although clickjacking attack has been the matter of discussion and disturbing reports, it is currently unclear what extent it clickjacking is used by attackers. But still, it is a problem incurred by typical internet users. This paper focuses on various clickjacking defenses on different web browsers. Keywords: Information Security, Web Security, Clickjacking, Web Browser, Web framing attack.

Abstract: Preparing a data set for analysis is the most time consuming task in a data mining project, requiring many complex SQL queries, joining tables, and aggregating columns. Existing SQL aggregations have limitations to prepare data sets because they return one column per aggregated group, and a significant manual effort is required to build data sets. In this paper proposing simple, efficient methods make SQL code return multiple columns in horizontal aggregation tables. It will return set of values instead of one value for one aggregation query. These functions of class are called as horizontal aggregations. It build data sets with a horizontal de normalized layout, which is the standard layout required by most data mining algorithms. In order to transform the data into suitable form three fundamental horizontal aggregation methods are used: SPJ (select, project, and join) method, CASE method and PIVOT method. The analysis become more efficient if the dataset obtained is in the horizontal form. Keywords: SQL Operators, Aggregate functions, Data Set Preparation.

Abstract: Facial expression recognition system using evolutionary particle swarm optimization (PSO)- based feature optimization employs modi?ed local binary patterns, which conduct horizontal and vertical neighborhood pixel comparison, to generate a discriminative initial facial representation. Then, a PSO variant embedded with the concept of a micro genetic algorithm (mGA), called mGA embedded PSO, is proposed to perform feature optimization. It incorporates a nonreplaceable memory, a small-population secondary swarm, a new velocity updating strategy, a subdimension- based in-depth local facial feature search, and a cooperation of local exploitation and global exploration search mechanism to mitigate the premature convergence problem of conventional PSO. Multiple classi?ers are used for recognizing seven facial expressions. Based on a comprehensive study using within- and cross-domain images from the extended Cohn Kanade and MMI benchmark databases, respectively, the empirical results indicate that proposed system outperforms other state-of-the-art PSO variants, conventional PSO, classical GA, and other related facial expression recognition models reported in the literature by a signi?cant margin. Keywords: Ensemble classi?er, facial expression recognition, feature selection, particle swarm optimization (PSO).

Abstract: Analyzing the users mainly refers to the ones who are able. But there is no complete acceptance behavior for such an approach. The disabled people also must have complete chance to use all the technologies. So these users need to be studied. All the accessibility tools can be utilized here and verify what is the performance of disabled users using these accessibilities such as keyboard accessibilities, screen accessibilities etc. methods should be identified which are for ease of these users. How do disabled users utilize these keys and accessibilities? Are they able to use it same as like normal users using keyboard keys and mouse movements? To study these characteristics of disabled users a support vector machine can be used. State space would create the data that users (disabled users) used to a certain task. Edge pace captures certain transitions. Thus support vector machine can be applied to these representations to have successful study. Keywords: Accessibility, Support vector machine (SVM), State space, Edge space.

Abstract: Multiple distortion assessment is a big challenge in image quality assessment (IQA). Here developing a no reference IQA model for multiply-distorted images. The NSS features, which are sensitive to each distortion, make a difference in quality. Multiple distortions are impaired by varying degree of distortions like blur, uneven illumination, noise and jpeg. The basis of model is to create different IQA metrics that are sensitive to blur, jpeg, noise and uneven illumination. The output of these metrics is combined to predict quality score. Subjective score is expensive and time consuming; due to that reason objective assessment should be taken for assessment of images. Applications in liveness detection and can be used for direct image recapture or correction. Keywords: No-Reference Quality Assessment, Multiple Distortion, Image Quality Metrics, NSS features.

Abstract: Cracks are an important indicator of the safety status of infrastructures. Detection of cracks on civil structure is a vital task for maintaining the structural health and reliability of buildings. The proposed system has the ability to i) identify crack, ii) report the type, iii) estimate the length and width of the crack, iv) also estimate the depth of crack in pillars by 3D approach . This method presents a novel automated crack detection algorithm using particle filter. Cracks can be classified in accordance with their geometrical form as vertical, horizontal, diagonal and complex. The proposed approach eliminates the complex cracks geometry because it is very rare in structures. However, with this proposed method, we can measure 94% of all cracks. This system eliminates the need for manually tuning threshold parameters. In this paper, the system based on machine vision concepts has been developed with the goal to automate the process of crack geometry measurement. A single camera installed in a truck or even in a robot is used to take sequence of images is processed and the crack dimensions are estimated. Keywords: Image analysis, machine vision, particle filter, crack depth, volume rendering.

Abstract: 2D barcodes are considered as an interface to connect potential customers and advertisement contents. Due to their crowdie appearance it gives no human readable information before the barcode is successfully decoded. The information content present in a 2D barcode can be delivered via a simple camera phone with a suitable decoding software in it.With the advanced communication principle a picture could be integrated into a 2D barcode called picode is developed, with this idea, a video clip,an image and an audio is integrated into a series of 2D barcode called vicode is also developed. To realise both picode and vicode a new modulation and demodulation schemes are developed and a new decoding scheme known as low density parity check code is used to provide better error rate performance than a traditional 2D barcode. The use of OFDM will increase the speed and accuracy while transmitting the multimedia messages. Picode and Vicode has been implemented in the MATLAB on a PC and it is successfully demonstrated for the real-world applications. Keywords: 2D barcode, 3D barcode, embedded picture, video and audio, Perceptual quality.

Abstract: Sparse representation has been successfully applied to visual tracking for finding the suitable candidate by using the target templates. But most of the sparse representation based trackers only consider the holistic representation of the target object and do not make the full use of the sparse coefficients to discriminate between target and background. Hence may fail with more possibility, when there is similar object or occlusion in the scene. This paper studies the visual tracking problem in video sequences and presents a sparse tracker using coarse and fine dictionaries. This representation exploits both partial and structural information of the target based on averaging and alignment-pooling method. The similarity obtained by pooling across the local patches helps not only locate the target more accurately but also handle partial occlusion. Object detection, identification and tracking are the three main objectives of this paper. For object/person detection a superpixel based face detection algorithm is used here that is followed by moment-based matching and isosceles triangle matching. Object tracking can be done by using extended Kalman filtering method. In addition, this method employs a template update strategy which combines incremental subspace learning and local sparse representation. This strategy adapts the template to the appearance change of the target with less possibility of drifting and reduces the influence of the occluded target template as well. The proposed algorithm is superior in accuracy and it has better robustness against to partial and full occlusion. Keywords: Object tracking, sparse coding, averaging, alignment-pooling, occlusion detection.

Abstract: The power quality (PQ) in distribution systems is principally affected by the pollution introduced by the customers. It is necessary to protect the sensitive loads from disturbances such as sags, swells, source voltage imbalances, etc. The actual solution for this case is to employ a dynamic voltage restorer (DVR) device. Dynamic Voltage Restorer (DVR) is a series connected compensators. A DVR has to supply energy to the load during the voltage sag and swell. The use of multilevel inverters in DVR improves the harmonic performance of the System. The synchronous reference frame based control is used for the DVR control. In this paper the design and simulations of three level inverter based DVR is discussed. A three level inverter based DVR and a two level inverter DVR is simulated in MATLAB software and the results are analysed. From the result it is observed that load voltage is maintained at reference value by injecting a series voltage and the load voltage THD is improved by employing the three level inverter. Keywords: PQ-Power Quality, DVR-Dynamic Voltage Restorer, THD-Total Harmonic Distortion, PWM-pulse width modulated.

Abstract: We are today in the era in which communication and information sharing has become integral part. Thus more and more technological advancements are coming up. A peer-to-peer network is among them. While sharing any information to another person we first have this thought if our data will be safe with the person or not. Thus, came the concept of securing the data and how significant trust can be. Many trust models have been proposed timely. One among these models is the Self ORganizing Trust (SORT)Model. Even though many attempts are still going on to improvise the calculations, SORT has been able to solve more issues than its contemporaries. We suggest this paper to improvise the existing SORT model accompanied with a hardware concept called the Trusted Platform Module (TPM), which was introduced by Trusted Computing Group. Since most of the vulnerabilities lie in the software part, it is believed that TPM can withstand such attacks. Keywords: SORT, TPM, trustworthiness, trusted computing.

Abstract: The idea of gesture recognition plays a vital role in human machine interaction in which the perception of human gestures by a computer is utilized. The main usage of gesture recognition research is in detecting a particular human gesture and convey the information related to each gesture to the user. After the identification of specific gesture of interest from the collection of gestures, clear-cut command for execution of action can be given to remote system. It is intended to make the computer understand human body language, thereby removing the gap between machine and human. Without depending on conventional input methods we can make use of gesture recognition to enhance human�computer communication. Using this approach, it is aimed to improve the effectiveness and efficiency of gesture recognition by providing an alternative method with some advancement in gesture identification and rectification using the calculations related to slop. Keywords: Human-machine interaction, Gestures, Image processing, Gesture recognition.

Abstract: Sentiment Analysis is a natural language processing task that mines opinion information from various text forms such as reviews, news, and blogs and classifies themby their polarity as positive, negative or neutral. With the rise of information being communicated via regional languages like Malayalam, comes a promising chance of mining this information. The works are very less in dialectal languages like Malayalam even though so many are there for universal languages like English. Mining sentiments in Malayalam comes with a lot of issues and challenges. As compared to English, Malayalam is a free order and morphologically rich language, which adds complexity while handling the user-generated content. This paper gives anoverview of the sentiment analysis works that has been donein the Malayalam language and challenges faced in these works. Keywords: Sentiment Analysis, Natural Language Processing, Malayalam Sentiment Analysis.

Abstract: Data mining is the extraction of useful knowledge from a large amount of data from different heterogeneous sources. Nowadays data is growing rapidly, and mining from these massive sets of data become the most complex task. The bigdata mining is the ability to extract information from large complex data due to its volume, velocity, verity, veracity and value. Uncovering of the huge amount of heterogeneous bigdata will maximize the knowledge in the target domain. So bigdata mining become one of the exciting opportunities today. The traditional data mining tools are not capable of handling large distributed data. Effective big data mining requires scalable and efficient solutions that are also useful to all kind of users. So the combination of efficient and user-friendly data mining tools will provide a more effective and scalable bigdata data mining platform for users with all levels of expertise. Keywords: Big Data, Data Mining, Frameworks, Distributed Systems.

Abstract: The intrusion detection is done in the data mining by means of using the clustering technique. Due to the risks of average clustering ways for intrusion detection, I am performing a graph-based intrusion detection algorithm with the aid of utilizing outlier detection ways. Compared to other intrusion detection algorithm of clustering, this algorithm is mindless to preliminary cluster quantity. In the meantime, it is strong within the outlier�s detection and capable to notice any shape of cluster alternatively that the circle one best. This paper makes use of graph-based cluster algorithm (GB) to get an initial partition of knowledge set to valid clusters by using an precision parameter. On the other hand, since of this intrusion detection mannequin is cantered on mixed training dataset, so it must have high label accuracy to assurance its efficiency. Hence, in labelling phrase, the algorithm imposes outlier detection algorithm to label the influence of GB algorithm once more. This measure is equipped to reinforce the labelling accuracy. Keywords: clustering, intrusion detection, anomaly based, graph algorithm, DBSCAN

Abstract: Malware or malicious software refers to software programs designed to damage or do other unwanted actions on a computer system which includes disrupt computer operations, gather sensitive information, gain access to private computer systems, or display unwanted advertising. Smartphones and mobile tablets are rapidly becoming indispensable in daily life. Given their large distribution, and also their capabilities, in the last two years mobile devices have become the main target for attackers . Android, the open source OS introduced by Google, has currently the largest market share, which is greater than 80%.Due to the openness and popularity, Android is the main target of attacks against mobile devices (98.5%), with more than 1 million of malicious apps currently available in the wild . Equipped withthe knowledge of malware�s capabilities, the detection of malware is an area of major concern not only to the research community but also to the general public. This paper focuses on the survey on different categories of malwares, its features and detection techniques. Keywords: Malwares, Static Analysis, Dynamic Analysis, Security, Android.

Abstract: In the field of Information Filtering we have many term-based or pattern �based methods for generating user�s needed form information from a set of documents .A basic general thinking is that documents in a set of particular collection is related to only a single topic .But in real life user�s interest is different and documents in a set or collection includes multiple topics. Most commonly used topic modeling method is Latent Dirichlet Allocation (LDA) which generates a structural model to represent multiple topics in a set of documents. Patterns generally are more descriptive and efficiently used in real time applications. So to select most descriptive and efficient patterns from the discovered set of patterns here a Maximum matched Pattern-based Topic Model is introduced. It helps us to get the relevant document according to user needs by filtering out unwanted documents. Keywords: Topic Model, Information Filtering, Pattern mining, relevance ranking, user interest model.

Abstract: Wireless sensor networks consist of autonomous sensor nodes attached to one or more base stations. Security is critical for many sensor network applications. Traditional key management techniques, such as public key cryptography or key distribution center (e.g., Kerberos), are often not effective for wireless sensor networks for the serious limitations in terms of computational power, energy supply, network bandwidth and defection of center authority. In order to balance the security and efficiency using group key agreement protocol and also support dynamic operations like join, leave, merge, etc. by using ECC based Diffie Hellman key exchange. This protocol employs ternary tree like structure instead of binary tree in the process of group key generation. Keywords: ECC, group key agreement, ternary tree, ECC based Diffie-Hellman.