Deepfake video detection using InceptionResnetV2
Sarra Guefrechi, Marwa Ben Jabra, Habib Hamam · 2022
Recently, deepfake face-swapping technology has become popular, making it easy to create ultra-realistic fake videos. Detecting the authenticity of videos is becoming increasingly important due to the potential negative impact videos can have on the world. This research is a method for developing a deep learning model, which can make a distinction between real and fake videos. Research papers on the subject of transfer learning in the domain of computer vision to take advantage of the earlier created neural network functions for image classification and to create new models based on them. Deep learning continues to evolve in both generating and detecting deepfakes. Models developed to detect deepfakes are designed using older datasets, may become outdated over time, and require new detection techniques all the time. Research results are promising with over 90% accuracy and areas for development and further development