Mining Forgery Traces from Reconstruction Error: A Weakly Supervised Framework for Multimodal Deepfake Temporal Localization

TL;DR

This paper introduces RT-DeepLoc, a weakly supervised framework that uses reconstruction errors from a Masked Autoencoder to accurately localize deepfake manipulations over time without requiring detailed annotations.

Contribution

The authors propose a novel reconstruction-based approach with a new contrastive loss for effective weakly supervised deepfake localization, outperforming existing methods.

Findings

Achieves state-of-the-art results on large-scale datasets.

Effectively localizes forgeries without dense annotations.

Robustly generalizes to unseen forgery methods.

Abstract

Modern deepfakes have evolved into localized and intermittent manipulations that require fine-grained temporal localization to mitigate severe digital security risks. The prohibitive cost of frame-level annotation makes weakly supervised methods a practical necessity, which rely only on video-level labels. To this end, we propose Reconstruction-based Temporal Deepfake Localization (RT-DeepLoc), a weakly supervised temporal forgery localization framework that identifies forgeries via reconstruction errors. Our framework uses a Masked Autoencoder (MAE) trained exclusively on authentic data to learn its intrinsic spatiotemporal patterns; this allows the model to produce significant reconstruction discrepancies for forged segments, effectively providing the missing fine-grained cues for accurate localization without demanding dense human annotations. To robustly leverage these indicators, we introduce a novel Asymmetric Intra-video Contrastive Loss (AICL). By focusing on the compactness of authentic features guided by these reconstruction cues, AICL establishes a stable decision boundary that enhances local discrimination while preserving generalization to unseen forgeries by advanced generative models. Extensive experiments on large-scale datasets, including LAV-DF, demonstrate that RT-DeepLoc achieves state-of-the-art performance in weakly-supervised temporal forgery localization.