RGFVR: Reference-Guided Face Video Restoration with Flow Matching

2026-06-15

AI Summary

This work addresses the limitations of existing face video restoration methods, which either suffer from identity loss under no-reference settings or exhibit poor generalization when tailored to specific identities. To overcome this trade-off, the authors propose an identity-agnostic, reference-guided restoration framework that integrates bimodal perceptual-descriptive identity conditions into a pre-trained flow-matching-based text-to-video generation model for the first time. A two-stage training strategy is introduced to enhance identity-aware guidance without requiring identity-specific fine-tuning. The proposed approach achieves significant improvements in visual quality, temporal consistency, and identity fidelity across complex degradation scenarios—including downsampling, blur, noise, and compression artifacts—outperforming current state-of-the-art methods while maintaining broad applicability.

Abstract

Face video restoration from degraded observations is challenging, as it requires simultaneously recovering visual fidelity, temporal consistency, and subject identity. Existing approaches are often either reference-free, which can lead to identity loss when person-specific facial details are lost, or subject-specific, which limits generalization to unseen identities. We propose a subject-agnostic, reference-guided framework for identity-preserving face video restoration. Our method introduces bimodal perceptual-descriptive identity conditioning into a pretrained flow-based text-to-video generator and employs a two-stage training strategy to strengthen identity guidance during restoration. Experiments show that our approach improves restoration fidelity, temporal consistency, and identity preservation, achieving superior performance under challenging video degradations, including downsampling, blur, noise, and compression artifacts. The code is available under: https://github.com/batuhanntosun/RG-FVR.

Problem

• face video restoration
• identity preservation
• temporal consistency
• visual fidelity
• reference-guided

Innovation

• reference-guided
• face video restoration
• flow matching
• identity preservation
• bimodal conditioning