Cost Sensitive Optimization of Deepfake Detector
This work is significant for video streaming platforms and other users who need to screen large volumes of videos for deepfakes, by proposing a cost-sensitive optimization approach for deepfake detectors.
This paper addresses the problem of deepfake detection from the perspective of a screening task, where a large number of videos are screened daily, with only a small fraction being deepfakes. The authors propose that detection performance and model parameter estimation should be measured and optimized in a cost-sensitive manner.
Since the invention of cinema, the manipulated videos have existed. But generating manipulated videos that can fool the viewer has been a time-consuming endeavor. With the dramatic improvements in the deep generative modeling, generating believable looking fake videos has become a reality. In the present work, we concentrate on the so-called deepfake videos, where the source face is swapped with the targets. We argue that deepfake detection task should be viewed as a screening task, where the user, such as the video streaming platform, will screen a large number of videos daily. It is clear then that only a small fraction of the uploaded videos are deepfakes, so the detection performance needs to be measured in a cost-sensitive way. Preferably, the model parameters also need to be estimated in the same way. This is precisely what we propose here.