Deepfake Detection with Deeplearning Using Resnet CNN Algorithm
Rpunithavathi, Mr. Mukesh Sai M, Mr.Hiruthik R, Mr.Sripadmesh S, Mr.Kishore RV · 2023
Due to information security breaches, video forgery is constantly rising in the digital world, creating a need for video and photo material surveillance for forgery identification.The proliferation of false films increases societal chaos and security dangers.The increase in malware has made it easier for users (anyone) to post, download, or exchange objects online that include audio, photos, or video, which is the cause of video forgeries.Recent years have seen a significant increase in media manipulation due to the advancement of technology and simplicity of producing fake information.Applications for video forgery detection include multimedia science, forensic examination, digital investigations, and video authenticity verification.The goal of video forensics technologies is to extract characteristics that can be used to tell false content frames apart from genuine videos.Deepfake media is produced and disseminated widely throughout social media platforms, and its identification is considered as posing a significant threat to media integrity.Falsification detection in video has been supplied with a proposed method for Deepfake detection.Convolutional neural network (CNN) method ResNet is employed as a method to identify Deepfake movies.The model tries to improve the reliability of the detector as well as the performance of identifying forgeries movies made using a certain technique.In order to identify the counterfeit in the movie, the suggested method simply makes use of the deep features that were recovered from of the ResNet Classification algorithm and then applies the standard mathematical method to these features.In an effort to address Deepfake video identification, the detector will offer a preliminary solution and be updated frequently with data from the actual world.