GIFGuard: Proactive Forensics against Deepfakes in Facial GIFs via Spatiotemporal Watermarking

2026-04-29

AI Summary

This work addresses the limitations of existing active forensic methods, which are confined to static images and struggle to counter deepfake threats in facial GIFs prevalent on social networks. To bridge this gap, the study proposes the first spatiotemporal watermarking framework tailored for GIFs, enabling proactive defense through dedicated embedding and extraction stages. Key contributions include the creation of GIFfaces—the first large-scale GIF-based forensic dataset—and the development of a spatiotemporally adaptive watermark encoding and recovery mechanism. The framework integrates a 3D convolutional backbone, adaptive channel recalibration, a spatiotemporal hourglass architecture, and 3D attention mechanisms. Experimental results demonstrate that the method achieves high visual fidelity while exhibiting significantly greater robustness against deepfakes compared to current state-of-the-art techniques.

Abstract

The rapid evolution of deepfake technology poses an unprecedented threat to the authenticity of Graphics Interchange Format (GIF) imagery, which serves as a representative of short-loop temporal media in social networks. However, existing proactive forensics works are designed for static images, which limits their applicability to animated GIFs. To bridge this gap, we propose GIFGuard, the first spatiotemporal watermarking framework tailored for deepfake proactive forensics in GIFs. In the embedding stage, we propose the Spatiotemporal Adaptive Residual Encoder (STARE) to ensure robustness against high-level semantic tampering. It employs a 3D convolutional backbone with adaptive channel recalibration to capture globally coherent temporal dependencies. In the extraction stage, we design the Deep Integrity Restoration Decoder (DIRD). It utilizes a spatiotemporal hourglass architecture equipped with 3D attention to restore latent features, allowing for the accurate extraction of watermark signals even under severe facial manipulation. Furthermore, we construct GIFfaces, the first large-scale benchmark dataset curated for GIF proactive forensics to facilitate research in this domain. Extensive results show that GIFGuard achieves high-fidelity visual quality and remarkable robustness performance against deepfakes. Related code and dataset will be released.

Problem

• deepfakes
• GIF
• proactive forensics
• spatiotemporal watermarking
• facial manipulation

Innovation

• spatiotemporal watermarking
• deepfake forensics
• 3D convolutional network
• adaptive channel recalibration
• GIF dataset