Lips Don't Lie: A Generalisable and Robust Approach to Face Forgery Detection

TL;DR

This paper introduces LipForensics, a face forgery detection method that leverages high-level mouth movement analysis via lipreading to achieve superior generalisation to unseen manipulations and robustness against distortions.

Contribution

It presents a novel approach combining lipreading with a two-stage neural network training process for improved face forgery detection.

Findings

Outperforms state-of-the-art in generalisation to unseen manipulations

Demonstrates robustness to various distortions and post-processing

Highlights importance of high-level semantic features in detection

Abstract

Although current deep learning-based face forgery detectors achieve impressive performance in constrained scenarios, they are vulnerable to samples created by unseen manipulation methods. Some recent works show improvements in generalisation but rely on cues that are easily corrupted by common post-processing operations such as compression. In this paper, we propose LipForensics, a detection approach capable of both generalising to novel manipulations and withstanding various distortions. LipForensics targets high-level semantic irregularities in mouth movements, which are common in many generated videos. It consists in first pretraining a spatio-temporal network to perform visual speech recognition (lipreading), thus learning rich internal representations related to natural mouth motion. A temporal network is subsequently finetuned on fixed mouth embeddings of real and forged data in order to detect fake videos based on mouth movements without overfitting to low-level, manipulation-specific artefacts. Extensive experiments show that this simple approach significantly surpasses the state-of-the-art in terms of generalisation to unseen manipulations and robustness to perturbations, as well as shed light on the factors responsible for its performance. Code is available on GitHub.