CosFace: Large Margin Cosine Loss for Deep Face Recognition

Abstract

Face recognition has made extraordinary progress owing to the advancement of deep convolutional neural networks (CNNs). The central task of face recognition, including face verification and identification, involves face feature discrimination. However, the traditional softmax loss of deep CNNs usually lacks the power of discrimination. To address this problem, recently several loss functions such as center loss, large margin softmax loss, and angular softmax loss have been proposed. All these improved losses share the same idea: maximizing inter-class variance and minimizing intra-class variance. In this paper, we propose a novel loss function, namely large margin cosine loss (LMCL), to realize this idea from a different perspective. More specifically, we reformulate the softmax loss as a cosine loss by $L_2$ normalizing both features and weight vectors to remove radial variations, based on which a cosine margin term is introduced to further maximize the decision margin in the angular space. As a result, minimum intra-class variance and maximum inter-class variance are achieved by virtue of normalization and cosine decision margin maximization. We refer to our model trained with LMCL as CosFace. Extensive experimental evaluations are conducted on the most popular public-domain face recognition datasets such as MegaFace Challenge, Youtube Faces (YTF) and Labeled Face in the Wild (LFW). We achieve the state-of-the-art performance on these benchmarks, which confirms the effectiveness of our proposed approach.

Accepted by CVPR 2018

References in corpus (14)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Deep Residual Learning for Image Recognition
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Learning Face Representation from Scratch
• YFCC100M: The New Data in Multimedia Research
• DeepID3: Face Recognition with Very Deep Neural Networks
• Aggregated Residual Transformations for Deep Neural Networks
• Targeting Ultimate Accuracy: Face Recognition via Deep Embedding
• Beyond triplet loss: a deep quadruplet network for person re-identification
• von Mises-Fisher Mixture Model-based Deep learning: Application to Face Verification
• Range Loss for Deep Face Recognition with Long-tail
• Deeply learned face representations are sparse, selective, and robust
• When Face Recognition Meets with Deep Learning: an Evaluation of Convolutional Neural Networks for Face Recognition
• Directional Statistics-based Deep Metric Learning for Image Classification and Retrieval

Cited by in corpus (55)

• AdvHat: Real-world adversarial attack on ArcFace Face ID system
• Circle Loss: A Unified Perspective of Pair Similarity Optimization
• Biometric Template Protection for Neural-Network-based Face Recognition Systems: A Survey of Methods and Evaluation Techniques
• Feature Fusion from Head to Tail for Long-Tailed Visual Recognition
• MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices
• Modality-Agnostic Attention Fusion for visual search with text feedback
• Support Vector Guided Softmax Loss for Face Recognition
• AdaCos: Adaptively Scaling Cosine Logits for Effectively Learning Deep Face Representations
• 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
• Jointly De-biasing Face Recognition and Demographic Attribute Estimation
• Spectral Feature Transformation for Person Re-identification
• Finding Missing Children: Aging Deep Face Features
• Contrastive Learning for Debiased Candidate Generation in Large-Scale Recommender Systems
• SegSort: Segmentation by Discriminative Sorting of Segments
• Video Face Recognition: Component-wise Feature Aggregation Network (C-FAN)
• Improving Generalization via Scalable Neighborhood Component Analysis
• Orthogonal Deep Features Decomposition for Age-Invariant Face Recognition
• Masked Face Recognition Challenge: The InsightFace Track Report
• Deep Metric Learning Meets Deep Clustering: An Novel Unsupervised Approach for Feature Embedding
• FaceLeaks: Inference Attacks against Transfer Learning Models via Black-box Queries
• RobFR: Benchmarking Adversarial Robustness on Face Recognition
• Joint Feature Distribution Alignment Learning for NIR-VIS and VIS-VIS Face Recognition
• Towards Generalized and Incremental Few-Shot Object Detection
• 1st Place Solution to Google Landmark Retrieval 2020
• Learning towards Minimum Hyperspherical Energy
• Softmax Dissection: Towards Understanding Intra- and Inter-class Objective for Embedding Learning
• MultiFace: A Generic Training Mechanism for Boosting Face Recognition Performance
• DocFace: Matching ID Document Photos to Selfies
• Efficient Realistic Data Generation Framework leveraging Deep Learning-based Human Digitization
• Student Network Learning via Evolutionary Knowledge Distillation
• A Fast and Accurate System for Face Detection, Identification, and Verification
• DashCam Pay: A System for In-vehicle Payments Using Face and Voice
• End-to-end Evaluation of Practical Video Analytics Systems for Face Detection and Recognition
• Regularizing Neural Networks via Minimizing Hyperspherical Energy
• Face Recognition in Unconstrained Conditions: A Systematic Review
• DotFAN: A Domain-transferred Face Augmentation Network for Pose and Illumination Invariant Face Recognition
• More Information Supervised Probabilistic Deep Face Embedding Learning
• Proxy Synthesis: Learning with Synthetic Classes for Deep Metric Learning
• Loop Closure Detection with RGB-D Feature Pyramid Siamese Networks
• DocFace+: ID Document to Selfie Matching
• Learning to Cluster Faces on an Affinity Graph
• Boosting Semi-Supervised Face Recognition with Noise Robustness
• On measuring the iconicity of a face
• Domain-Aware SE Network for Sketch-based Image Retrieval with Multiplicative Euclidean Margin Softmax
• Unified Batch All Triplet Loss for Visible-Infrared Person Re-identification
• Characters as Graphs: Recognizing Online Handwritten Chinese Characters via Spatial Graph Convolutional Network
• Robust Angular Local Descriptor Learning
• Learning discriminative features in sequence training without requiring framewise labelled data
• Investigation of Large-Margin Softmax in Neural Language Modeling
• Reconstructing A Large Scale 3D Face Dataset for Deep 3D Face Identification
• Learning Clusterable Visual Features for Zero-Shot Recognition
• LAE : Long-tailed Age Estimation
• Learning Metrics from Mean Teacher: A Supervised Learning Method for Improving the Generalization of Speaker Verification System