SAVe: Self-Supervised Audio-visual Deepfake Detection Exploiting Visual Artifacts and Audio-visual Misalignment

TL;DR

SAVe introduces a self-supervised framework for audio-visual deepfake detection that learns from authentic videos, using pseudo-manipulations and lip-speech synchronization to identify subtle artifacts and inconsistencies.

Contribution

The paper presents a novel self-supervised approach that does not rely on synthetic training data, using on-the-fly pseudo-manipulations and cross-modal alignment for robust deepfake detection.

Findings

Achieves competitive in-domain detection performance.

Demonstrates strong cross-dataset generalization.

Effective in identifying subtle visual artifacts and audio-visual misalignments.

Abstract

Multimodal deepfakes can exhibit subtle visual artifacts and cross-modal inconsistencies, which remain challenging to detect, especially when detectors are trained primarily on curated synthetic forgeries. Such synthetic dependence can introduce dataset and generator bias, limiting scalability and robustness to unseen manipulations. We propose SAVe, a self-supervised audio-visual deepfake detection framework that learns entirely on authentic videos. SAVe generates on-the-fly, identity-preserving, region-aware self-blended pseudo-manipulations to emulate tampering artifacts, enabling the model to learn complementary visual cues across multiple facial granularities. To capture cross-modal evidence, SAVe also models lip-speech synchronization via an audio-visual alignment component that detects temporal misalignment patterns characteristic of audio-visual forgeries. Experiments on FakeAVCeleb and AV-LipSync-TIMIT demonstrate competitive in-domain performance and strong cross-dataset generalization, highlighting self-supervised learning as a scalable paradigm for multimodal deepfake detection.