Glitch in the Matrix: A Large Scale Benchmark for Content Driven Audio-Visual Forgery Detection and Localization

TL;DR

This paper introduces a new large-scale benchmark dataset, LAV-DF, for content-driven audio-visual deepfake detection and localization, along with a novel multimodal detection method, BA-TFD+, which outperforms existing approaches.

Contribution

The paper presents the LAV-DF dataset for content-driven audio-visual deepfake detection and localization, and proposes an improved multimodal detection architecture, BA-TFD+, with enhanced accuracy.

Findings

BA-TFD+ outperforms baseline in deepfake detection

LAV-DF dataset captures content-driven audio-visual manipulations

Proposed methods achieve superior localization accuracy

Abstract

Most deepfake detection methods focus on detecting spatial and/or spatio-temporal changes in facial attributes and are centered around the binary classification task of detecting whether a video is real or fake. This is because available benchmark datasets contain mostly visual-only modifications present in the entirety of the video. However, a sophisticated deepfake may include small segments of audio or audio-visual manipulations that can completely change the meaning of the video content. To addresses this gap, we propose and benchmark a new dataset, Localized Audio Visual DeepFake (LAV-DF), consisting of strategic content-driven audio, visual and audio-visual manipulations. The proposed baseline method, Boundary Aware Temporal Forgery Detection (BA-TFD), is a 3D Convolutional Neural Network-based architecture which effectively captures multimodal manipulations. We further improve (i.e. BA-TFD+) the baseline method by replacing the backbone with a Multiscale Vision Transformer and guide the training process with contrastive, frame classification, boundary matching and multimodal boundary matching loss functions. The quantitative analysis demonstrates the superiority of BA-TFD+ on temporal forgery localization and deepfake detection tasks using several benchmark datasets including our newly proposed dataset. The dataset, models and code are available at https://github.com/ControlNet/LAV-DF.