MagFace: A Universal Representation for Face Recognition and Quality Assessment
This addresses face recognition in the wild by enhancing robustness to low-quality samples, though it is incremental as it builds on prior quality monitoring methods.
The paper tackles the problem of face recognition performance degradation due to varying face quality by proposing MagFace, a loss function that learns a feature embedding where magnitude indicates face quality, proven to increase with recognition likelihood, and improves state-of-the-art results in experiments.
The performance of face recognition system degrades when the variability of the acquired faces increases. Prior work alleviates this issue by either monitoring the face quality in pre-processing or predicting the data uncertainty along with the face feature. This paper proposes MagFace, a category of losses that learn a universal feature embedding whose magnitude can measure the quality of the given face. Under the new loss, it can be proven that the magnitude of the feature embedding monotonically increases if the subject is more likely to be recognized. In addition, MagFace introduces an adaptive mechanism to learn a wellstructured within-class feature distributions by pulling easy samples to class centers while pushing hard samples away. This prevents models from overfitting on noisy low-quality samples and improves face recognition in the wild. Extensive experiments conducted on face recognition, quality assessments as well as clustering demonstrate its superiority over state-of-the-arts. The code is available at https://github.com/IrvingMeng/MagFace.