🤖 AI Summary
Existing face de-identification methods suffer from significant limitations in preserving attribute details and maintaining robustness against occlusions, often introducing noticeable editing artifacts that compromise visual authenticity and fidelity. To address these issues, we propose a “decoupling-first” two-stage framework. First, a Contrastive Identity-Decoupling (CID) module explicitly disentangles identity-specific features from attribute-related features via contrastive learning. Second, a Key-driven Reversible Anonymous Representation (KRIA) mechanism—integrated with a Multi-scale Attention-based Attribute Refinement (MAAR) module—enables high-fidelity, fine-grained attribute preservation under occluded conditions. The entire framework supports end-to-end joint optimization. Extensive experiments demonstrate that our method achieves a 12.6% improvement in attribute fidelity and a 23.4% gain in editing quality under occlusion, while substantially suppressing artificial artifacts. It outperforms state-of-the-art approaches across multiple quantitative and qualitative metrics.
📝 Abstract
In an era where personal photos are easily leaked and collected, face de-identification is a crucial method for protecting identity privacy. However, current face de-identification techniques face challenges in preserving attribute details and often produce anonymized results with reduced authenticity. These shortcomings are particularly evident when handling occlusions,frequently resulting in noticeable editing artifacts. Our primary finding in this work is that simultaneous training of identity disentanglement and anonymization hinders their respective effectiveness.Therefore, we propose"Disentangle Before Anonymize",a novel two-stage Framework(DBAF)designed for attributepreserved and occlusion-robust de-identification. This framework includes a Contrastive Identity Disentanglement (CID) module and a Key-authorized Reversible Identity Anonymization (KRIA) module, achieving faithful attribute preservation and high-quality identity anonymization edits. Additionally, we introduce a Multiscale Attentional Attribute Retention (MAAR) module to address the issue of reduced anonymization quality under occlusions.Extensive experiments demonstrate that our method outperforms state-of-the-art de-identification approaches, delivering superior quality, enhanced detail fidelity, improved attribute preservation performance, and greater robustness to occlusions.
