Deep Fake Detection Of Videos

Authors: Thejashwini M, Thejaswini K P, Vismitha K, Vivek A M, Shylaja B

Abstract: Deep fake videos—synthetically manipulated visual content created by deep learning techniques—pose an alarming threat issues in sectors involving from media and politics to healthcare. This study aims to develop a robust deep fake detec- tion framework using deep learning methods and algorithms, focusing on generalization across datasets and video types. Our approach is informed by an extensive literature review of recent advancements, including methods using xception, Mobile Net, Mask R-CNN, and affective cue-based models. The review highlights challenges such as dataset diversity, generalization issues, and real-time detection limitations. Our proposed deep learning system will address these challenges by integrating temporal and spatial video analysis, leveraging convolutional and recurrent neural networks to detect subtle manipulative traces.

DOI: http://doi.org/