Deepfake Detection using Inception-ResNet-V2 Network

Deepfakes entails obscene videos in which a face can be changed with someone else's utilizing neural networks. Deepfakes are a public problem, thus developing methods to detect them is critical. With deepfake and human-level control, deep learning has already utilized to address complicated issues. Deepfake is a new branch of AI technology, which in that one's face is superimposed on another's face, which is popular on social media. Deep learning techniques were also employed in the development of software that affects national security, democracy and privacy. One of the most current applications that use deep learning is deepfakes. Algorithms for deepfake may generate fake photos and movies which are indistinguishable from real ones. As a result, the development of technology capable of automatically detecting and assessing the reliability of digital video is critical. This research describes algorithms for detecting deepfakes. By reviewing the influence of deepfakes and deepfake recognition systems, this work enables the creation of new and so many effective methodologies to cope with increasingly complex deepfakes. InceptionResNetV2 architecture in Convolutional Neural Networks (CNN) is utilized in this comparative study to distinguish real and deepfake images.