Metric Learning for Anti-Compression Facial Forgery Detection

TL;DR

This paper introduces a metric learning-based framework that effectively detects facial forgeries regardless of compression artifacts by learning a compression-insensitive embedding space, improving robustness in multimedia forensics.

Contribution

The paper proposes a novel anti-compression facial forgery detection method combining adversarial feature extraction, metric loss, and attention transfer to enhance detection accuracy across compressed and uncompressed data.

Findings

High effectiveness in handling compressed facial forgeries

Improved localization accuracy of tampered regions

Robustness against various compression formats

Abstract

Detecting facial forgery images and videos is an increasingly important topic in multimedia forensics. As forgery images and videos are usually compressed into different formats such as JPEG and H264 when circulating on the Internet, existing forgery-detection methods trained on uncompressed data often suffer from significant performance degradation in identifying them. To solve this problem, we propose a novel anti-compression facial forgery detection framework, which learns a compression-insensitive embedding feature space utilizing both original and compressed forgeries. Specifically, our approach consists of three ideas: (i) extracting compression-insensitive features from both uncompressed and compressed forgeries using an adversarial learning strategy; (ii) learning a robust partition by constructing a metric loss that can reduce the distance of the paired original and compressed images in the embedding space; (iii) improving the accuracy of tampered localization with an attention-transfer module. Experimental results demonstrate that, the proposed method is highly effective in handling both compressed and uncompressed facial forgery images.