Development of deepfakes detection model using deep learning framework

Abstract: In response to the growing threat of deepfake technology, which can deceive and manipulate individuals, leading to identity theft, financial fraud, and political manipulation. This paper proposes a deepfakes detection model using a deep learning framework and image processing techniques. The proposed utilizes the ResNeXt architecture as a powerful feature extractor to capture intricate patterns and discriminative features from input images or video frames. Additionally, the LSTM architecture is employed to handle temporal dependencies, enabling the model to analyze the temporal coherence and consistency of video sequences. By leveraging ResNeXt and LSTM, the model achieves enhanced accuracy and robustness in detecting deepfake content.

Keywords: LSTM- long short term memory, cnn- convolution neural network.