GeoMAR: Unleashing Geometrically Aligned Features for Masked Autoregressive Blind Face Restoration

2026-08-04

AI Summary

This work addresses the limitations of existing codebook-based blind face restoration methods, which suffer from ambiguous conditional features and fragile prediction mechanisms under severe degradation. To overcome these issues, the authors propose a multi-step masked autoregressive generation framework that leverages dual-input geometric feature extraction and an alignment prior injector to produce spatially precise geometric guidance. Additionally, a KV-Q swapping strategy is introduced to enhance feature interaction. By reformulating the single-step mapping into a coarse-to-fine multi-step generation process, the method significantly improves restoration robustness and structural consistency. Extensive experiments demonstrate that the proposed model achieves state-of-the-art perceptual quality and visual coherence across one synthetic and three real-world benchmarks.

Abstract

Codebook-based blind face restoration (BFR) often suffers from ambiguous conditioning features and a fragile prediction mechanism under severe degradation. To address these challenges, we propose GeoMAR, a framework designed to unleash geometrically aligned features with masked autoregressive (MAR) refinement for robust face restoration. For feature conditioning, we introduce a dual-input extraction pipeline to extract component-based geometric descriptions with explicit, spatially faithful anchors. These textual priors are integrated with low-quality (LQ) features via an Aligned Geometric Priors Injector, which employs a KV-Q exchange strategy to generate geometrically aligned features. For prediction mechanism, we reformulate the one-step mapping into a multi-step MAR process. This coarse-to-fine generation progressively refines complex facial regions based on increasingly reliable context. Experiments on one synthetic and three real-world benchmarks demonstrate that GeoMAR achieves highly competitive perceptual quality and coherent visual structures compared with existing methods. The code is available at https://github.com/BRL-SYSU/GeoMAR.git.

Problem

• blind face restoration
• ambiguous conditioning features
• fragile prediction mechanism
• severe degradation

Innovation

• Geometrically Aligned Features
• Masked Autoregressive
• Blind Face Restoration
• Codebook-based
• KV-Q Exchange