MGFace: Mask-Gated Face Matching via Conditional Similarity Routing

2026-07-14

AI Summary

This work addresses the dual challenges of degraded face recognition accuracy under partial occlusions such as masks and the high computational cost of large-scale retrieval. To this end, the authors propose a conditional routing mechanism that dynamically selects a matching strategy based on the predicted presence of a mask: when no mask is detected, global embedding matching is employed; when a mask is present, a mask-aware patch-level re-ranking based on the upper facial region is activated. Integrated with FaceNet or ArcFace backbones, the method achieves over 80% and 90% accuracy, respectively, on the LFW-Mask dataset, outperforming the EMD re-ranking approach while reducing computation time by approximately 20×. This strategy significantly lowers computational overhead without compromising robustness.

Abstract

Face identification has achieved remarkable performance under normal conditions. Yet, its accuracy often degrades significantly when query faces are partially occluded, especially by facial masks. Existing re-ranking approaches improve robustness by exploiting patch-level similarities. Still, they often rely on costly, fine-grained matching mechanisms, which limit their efficiency in large-scale retrieval scenarios. In this paper, we propose MGFace, a mask-gated face identification pipeline that predicts the mask status of a query face and conditionally routes the similarity computation accordingly. Specifically, MGFace distinguishes between masked and unmasked queries, applies global embedding matching to unmasked queries, and activates mask-aware patch-level re-ranking only for masked queries. This design focuses on reliable upper-face regions while avoiding unnecessary fine-grained computation. Experiments on the extended LFW-Mask dataset show that MGFace achieves over 80% identification accuracy with the FaceNet backbone and over 90% with the ArcFace backbone. Compared with a previous EMD-based re-ranking method, MGFace achieves better identification performance while reducing query time by approximately 20x. These results demonstrate the effectiveness of MGFace in improving masked-face identification accuracy with low computational overhead. The source code is available at https://github.com/chequanghuy/MGFace.

Problem

• masked face recognition
• occlusion
• face identification
• re-ranking
• computational efficiency

Innovation

• mask-gated routing
• conditional similarity
• patch-level re-ranking
• masked face recognition
• efficient face identification