English

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

Computer Vision and Pattern Recognition 2026-05-19 v2

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.

Keywords

Cite

@article{arxiv.2601.21458,
  title  = {Mining Forgery Traces from Reconstruction Error: A Weakly Supervised Framework for Multimodal Deepfake Temporal Localization},
  author = {Midou Guo and Qilin Yin and Wei Lu and Rui Yang},
  journal= {arXiv preprint arXiv:2601.21458},
  year   = {2026}
}
R2 v1 2026-07-01T09:25:20.845Z