SubFace: Learning with Softmax Approximation for Face Recognition

TL;DR

SubFace introduces a subspace feature approximation method that dynamically selects features during training to enhance face recognition performance, significantly improving baseline CNN models on benchmark datasets.

Contribution

The paper proposes a novel subspace feature approximation approach for softmax-based loss functions, improving face recognition accuracy by dynamically selecting features during training.

Findings

Significant performance improvement over baseline CNNs.

Effective enhancement of discriminability in face recognition.

Validation on benchmark datasets confirms the method's effectiveness.

Abstract

The softmax-based loss functions and its variants (e.g., cosface, sphereface, and arcface) significantly improve the face recognition performance in wild unconstrained scenes. A common practice of these algorithms is to perform optimizations on the multiplication between the embedding features and the linear transformation matrix. However in most cases, the dimension of embedding features is given based on traditional design experience, and there is less-studied on improving performance using the feature itself when giving a fixed size. To address this challenge, this paper presents a softmax approximation method called SubFace, which employs the subspace feature to promote the performance of face recognition. Specifically, we dynamically select the non-overlapping subspace features in each batch during training, and then use the subspace features to approximate full-feature among softmax-based loss, so the discriminability of the deep model can be significantly enhanced for face recognition. Comprehensive experiments conducted on benchmark datasets demonstrate that our method can significantly improve the performance of vanilla CNN baseline, which strongly proves the effectiveness of subspace strategy with the margin-based loss.