L2-constrained Softmax Loss for Discriminative Face Verification

Abstract

In recent years, the performance of face verification systems has significantly improved using deep convolutional neural networks (DCNNs). A typical pipeline for face verification includes training a deep network for subject classification with softmax loss, using the penultimate layer output as the feature descriptor, and generating a cosine similarity score given a pair of face images. The softmax loss function does not optimize the features to have higher similarity score for positive pairs and lower similarity score for negative pairs, which leads to a performance gap. In this paper, we add an L2-constraint to the feature descriptors which restricts them to lie on a hypersphere of a fixed radius. This module can be easily implemented using existing deep learning frameworks. We show that integrating this simple step in the training pipeline significantly boosts the performance of face verification. Specifically, we achieve state-of-the-art results on the challenging IJB-A dataset, achieving True Accept Rate of 0.909 at False Accept Rate 0.0001 on the face verification protocol. Additionally, we achieve state-of-the-art performance on LFW dataset with an accuracy of 99.78%, and competing performance on YTF dataset with accuracy of 96.08%.

Cited by in corpus (22)

• A Survey of Deep Learning-based Object Detection
• Deep Face Recognition: A Survey
• On Low-Resolution Face Recognition in the Wild: Comparisons and New Techniques
• SFace: Sigmoid-Constrained Hypersphere Loss for Robust Face Recognition
• Recent Advances in Deep Learning Techniques for Face Recognition
• Circle Loss: A Unified Perspective of Pair Similarity Optimization
• Variational Graph Normalized Auto-Encoders
• An Experimental Evaluation of Covariates Effects on Unconstrained Face Verification
• UR Channel-Robust Synthetic Speech Detection System for ASVspoof 2021
• Deep Feature Space: A Geometrical Perspective
• Relational Deep Feature Learning for Heterogeneous Face Recognition
• Deep Representation Learning on Long-tailed Data: A Learnable Embedding Augmentation Perspective
• Spatial Pyramid Encoding with Convex Length Normalization for Text-Independent Speaker Verification
• Regular Polytope Networks
• The complementarity of a diverse range of deep learning features extracted from video content for video recommendation
• Adjusting Logit in Gaussian Form for Long-Tailed Visual Recognition
• Person Recognition in Personal Photo Collections
• Data Uncertainty Learning in Face Recognition
• Prototype Memory for Large-scale Face Representation Learning
• Adaptive Sparse Softmax: An Effective and Efficient Softmax Variant
• CAMRI Loss: Improving Recall of a Specific Class without Sacrificing Accuracy
• Spherical Feature Transform for Deep Metric Learning