SphereFace: Deep Hypersphere Embedding for Face Recognition

Abstract

This paper addresses deep face recognition (FR) problem under open-set protocol, where ideal face features are expected to have smaller maximal intra-class distance than minimal inter-class distance under a suitably chosen metric space. However, few existing algorithms can effectively achieve this criterion. To this end, we propose the angular softmax (A-Softmax) loss that enables convolutional neural networks (CNNs) to learn angularly discriminative features. Geometrically, A-Softmax loss can be viewed as imposing discriminative constraints on a hypersphere manifold, which intrinsically matches the prior that faces also lie on a manifold. Moreover, the size of angular margin can be quantitatively adjusted by a parameter $m$. We further derive specific $m$ to approximate the ideal feature criterion. Extensive analysis and experiments on Labeled Face in the Wild (LFW), Youtube Faces (YTF) and MegaFace Challenge show the superiority of A-Softmax loss in FR tasks. The code has also been made publicly available.

CVPR 2017 (v4: updated the Appendix)

References in corpus (3)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Going Deeper with Convolutions
• Large-Margin Softmax Loss for Convolutional Neural Networks

Cited by in corpus (95)

• AdvHat: Real-world adversarial attack on ArcFace Face ID system
• SphereReID: Deep Hypersphere Manifold Embedding for Person Re-Identification
• Unsupervised Feature Learning via Non-Parametric Instance-level Discrimination
• Enabling and Emerging Technologies for Social Distancing: A Comprehensive Survey and Open Problems
• Discrimination-aware Channel Pruning for Deep Neural Networks
• Rethinking Feature Discrimination and Polymerization for Large-scale Recognition
• Circle Loss: A Unified Perspective of Pair Similarity Optimization
• One-shot Face Recognition by Promoting Underrepresented Classes
• Efficient Facial Representations for Age, Gender and Identity Recognition in Organizing Photo Albums using Multi-output CNN
• Hyperspherical Prototype Networks
• von Mises-Fisher Mixture Model-based Deep learning: Application to Face Verification
• Biometric Template Protection for Neural-Network-based Face Recognition Systems: A Survey of Methods and Evaluation Techniques
• PersEmoN: A Deep Network for Joint Analysis of Apparent Personality, Emotion and Their Relationship
• Minimum Margin Loss for Deep Face Recognition
• Face Recognition in Low Quality Images: A Survey
• Crystal Loss and Quality Pooling for Unconstrained Face Verification and Recognition
• A Study on various state of the art of the Art Face Recognition System using Deep Learning Techniques
• Exploring the Encoding Layer and Loss Function in End-to-End Speaker and Language Recognition System
• Spatially and Temporally Efficient Non-local Attention Network for Video-based Person Re-Identification
• Iterative Learning with Open-set Noisy Labels
• Striking the Right Balance with Uncertainty
• Face Recognition via Centralized Coordinate Learning
• Meta-Learned Confidence for Few-shot Learning
• Robust Classification with Convolutional Prototype Learning
• Feature Transfer Learning for Deep Face Recognition with Under-Represented Data
• Deep Learning For Face Recognition: A Critical Analysis
• LOTR: Face Landmark Localization Using Localization Transformer
• The Devil of Face Recognition is in the Noise
• Secure Face Matching Using Fully Homomorphic Encryption
• Seeing Small Faces from Robust Anchor's Perspective
• Siamese Capsule Networks
• Inducing Predictive Uncertainty Estimation for Face Recognition
• Mis-classified Vector Guided Softmax Loss for Face Recognition
• Light Can Hack Your Face! Black-box Backdoor Attack on Face Recognition Systems
• Learning to Anonymize Faces for Privacy Preserving Action Detection
• On the Capacity of Face Representation
• Feature Incay for Representation Regularization
• Spectral Feature Transformation for Person Re-identification
• Finding Missing Children: Aging Deep Face Features
• Viewpoint-Aware Loss with Angular Regularization for Person Re-Identification
• Pairwise Relational Networks for Face Recognition
• Unsupervised Out-of-Distribution Detection by Maximum Classifier Discrepancy
• Video Face Recognition: Component-wise Feature Aggregation Network (C-FAN)
• Neural Architecture Search for Deep Face Recognition
• Orthogonal Deep Features Decomposition for Age-Invariant Face Recognition
• Domain Alignment with Triplets
• Masked Face Recognition Challenge: The InsightFace Track Report
• Ring loss: Convex Feature Normalization for Face Recognition
• Attention Control with Metric Learning Alignment for Image Set-based Recognition
• RobFR: Benchmarking Adversarial Robustness on Face Recognition
• Susceptibility to Image Resolution in Face Recognition and Trainings Strategies
• WarpGAN: Automatic Caricature Generation
• Deep Metric Learning with Hierarchical Triplet Loss
• AI Oriented Large-Scale Video Management for Smart City: Technologies, Standards and Beyond
• Ensemble Soft-Margin Softmax Loss for Image Classification
• Softmax Dissection: Towards Understanding Intra- and Inter-class Objective for Embedding Learning
• Ranked List Loss for Deep Metric Learning
• Technical Challenges for Training Fair Neural Networks
• GhostVLAD for set-based face recognition
• Analysis of Length Normalization in End-to-End Speaker Verification System
• DocFace: Matching ID Document Photos to Selfies
• Student Network Learning via Evolutionary Knowledge Distillation
• DashCam Pay: A System for In-vehicle Payments Using Face and Voice
• Learning Channel Inter-dependencies at Multiple Scales on Dense Networks for Face Recognition
• Instance-level Sketch-based Retrieval by Deep Triplet Classification Siamese Network
• Proxy Synthesis: Learning with Synthetic Classes for Deep Metric Learning
• SELM: Siamese Extreme Learning Machine with Application to Face Biometrics
• Pairwise Relational Networks using Local Appearance Features for Face Recognition
• Normalization Before Shaking Toward Learning Symmetrically Distributed Representation Without Margin in Speech Emotion Recognition
• More Information Supervised Probabilistic Deep Face Embedding Learning
• Deep Face Feature for Face Alignment
• An Attention Model for group-level emotion recognition
• Tackling Early Sparse Gradients in Softmax Activation Using Leaky Squared Euclidean Distance
• Generating 2D and 3D Master Faces for Dictionary Attacks with a Network-Assisted Latent Space Evolution
• Boosting Semi-Supervised Face Recognition with Noise Robustness
• Bias-Awareness for Zero-Shot Learning the Seen and Unseen
• Self-supervised pre-training with acoustic configurations for replay spoofing detection
• Robust Face Recognition with Deeply Normalized Depth Images
• Weakly supervised discriminative feature learning with state information for person identification
• Information Theory-Guided Heuristic Progressive Multi-View Coding
• A Deep, Information-theoretic Framework for Robust Biometric Recognition
• Unique Identification of Macaques for Population Monitoring and Control
• Reconstructing A Large Scale 3D Face Dataset for Deep 3D Face Identification
• Large-scale Multi-modal Person Identification in Real Unconstrained Environments
• A Machine Learning Framework for Data Ingestion in Document Images
• Dynamic Metric Learning: Towards a Scalable Metric Space to Accommodate Multiple Semantic Scales
• Self-attention aggregation network for video face representation and recognition
• Imbalance Robust Softmax for Deep Embeeding Learning
• Center Loss Regularization for Continual Learning
• Natural Attribute-based Shift Detection
• Bidirectional Multiscale Feature Aggregation for Speaker Verification
• Open Source Face Recognition Performance Evaluation Package
• On Improving the Generalization of Face Recognition in the Presence of Occlusions
• Investigation of Large-Margin Softmax in Neural Language Modeling
• Top-Related Meta-Learning Method for Few-Shot Object Detection