Not made for each other- Audio-Visual Dissonance-based Deepfake Detection and Localization

TL;DR

This paper introduces a novel deepfake detection method based on audio-visual dissonance, leveraging the inconsistency between modalities to identify manipulated videos and localize forgery segments effectively.

Contribution

It proposes the Modality Dissonance Score (MDS) for detecting deepfakes by measuring audio-visual dissonance, and demonstrates superior performance and forgery localization capabilities.

Findings

Outperforms state-of-the-art by up to 7% on benchmark datasets

Effectively localizes manipulated video segments

Demonstrates robustness across multiple datasets

Abstract

We propose detection of deepfake videos based on the dissimilarity between the audio and visual modalities, termed as the Modality Dissonance Score (MDS). We hypothesize that manipulation of either modality will lead to dis-harmony between the two modalities, eg, loss of lip-sync, unnatural facial and lip movements, etc. MDS is computed as an aggregate of dissimilarity scores between audio and visual segments in a video. Discriminative features are learnt for the audio and visual channels in a chunk-wise manner, employing the cross-entropy loss for individual modalities, and a contrastive loss that models inter-modality similarity. Extensive experiments on the DFDC and DeepFake-TIMIT Datasets show that our approach outperforms the state-of-the-art by up to 7%. We also demonstrate temporal forgery localization, and show how our technique identifies the manipulated video segments.