An Overview of Face Manipulation Detection

Abstract

Due to the power of editing tools, new types of fake faces are being created and synthesized, which has attracted great attention on social media. It is reasonable to acknowledge that one human cannot distinguish whether the face is manipulated from the real faces. Therefore, the detection of face manipulation becomes a critical issue in digital media forensics. This paper provides an overview of recent deep learning detection models for face manipulation. Some public dataset used for face manipulation detection is introduced. On this basis, the challenges for the research and the potential future directions are analyzed and discussed.