Exposing Deep Fakes Using Inconsistent Head Poses

TL;DR

This paper introduces a novel method to detect Deep Fakes by analyzing inconsistencies in estimated 3D head poses, revealing artifacts from face splicing that distinguish fake images from real ones.

Contribution

The paper presents a new detection approach leveraging head pose inconsistencies caused by face splicing in Deep Fakes, validated through experiments with an SVM classifier.

Findings

Deep Fakes exhibit detectable head pose inconsistencies.

The proposed method effectively distinguishes real from fake images.

Experimental results demonstrate high classification accuracy.

Abstract

In this paper, we propose a new method to expose AI-generated fake face images or videos (commonly known as the Deep Fakes). Our method is based on the observations that Deep Fakes are created by splicing synthesized face region into the original image, and in doing so, introducing errors that can be revealed when 3D head poses are estimated from the face images. We perform experiments to demonstrate this phenomenon and further develop a classification method based on this cue. Using features based on this cue, an SVM classifier is evaluated using a set of real face images and Deep Fakes.