A Light CNN for Deep Face Representation with Noisy Labels

Abstract

The volume of convolutional neural network (CNN) models proposed for face recognition has been continuously growing larger to better fit large amount of training data. When training data are obtained from internet, the labels are likely to be ambiguous and inaccurate. This paper presents a Light CNN framework to learn a compact embedding on the large-scale face data with massive noisy labels. First, we introduce a variation of maxout activation, called Max-Feature-Map (MFM), into each convolutional layer of CNN. Different from maxout activation that uses many feature maps to linearly approximate an arbitrary convex activation function, MFM does so via a competitive relationship. MFM can not only separate noisy and informative signals but also play the role of feature selection between two feature maps. Second, three networks are carefully designed to obtain better performance meanwhile reducing the number of parameters and computational costs. Lastly, a semantic bootstrapping method is proposed to make the prediction of the networks more consistent with noisy labels. Experimental results show that the proposed framework can utilize large-scale noisy data to learn a Light model that is efficient in computational costs and storage spaces. The learned single network with a 256-D representation achieves state-of-the-art results on various face benchmarks without fine-tuning. The code is released on https://github.com/AlfredXiangWu/LightCNN.

arXiv admin note: text overlap with arXiv:1507.04844. The models are released on https://github.com/AlfredXiangWu/LightCNN, IEEE Transactions on Information Forensics and Security, 2018

References in corpus (11)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Wide Residual Networks
• Learning Face Representation from Scratch
• Training Deep Neural Networks on Noisy Labels with Bootstrapping
• Training Convolutional Networks with Noisy Labels
• Scalable, High-Quality Object Detection
• Targeting Ultimate Accuracy: Face Recognition via Deep Embedding
• MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition
• Recurrent Regression for Face Recognition
• VIPLFaceNet: An Open Source Deep Face Recognition SDK

Cited by in corpus (31)

• On the Reconstruction of Face Images from Deep Face Templates
• Beyond Face Rotation: Global and Local Perception GAN for Photorealistic and Identity Preserving Frontal View Synthesis
• von Mises-Fisher Mixture Model-based Deep learning: Application to Face Verification
• MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices
• Privacy-Protective-GAN for Face De-identification
• Maximum A Posteriori Estimation of Distances Between Deep Features in Still-to-Video Face Recognition
• Wasserstein CNN: Learning Invariant Features for NIR-VIS Face Recognition
• Feature Transfer Learning for Deep Face Recognition with Under-Represented Data
• Deep Learning For Face Recognition: A Critical Analysis
• Global and Local Consistent Age Generative Adversarial Networks
• Cardea: Context-Aware Visual Privacy Protection from Pervasive Cameras
• Coupled Deep Learning for Heterogeneous Face Recognition
• Adversarial Discriminative Heterogeneous Face Recognition
• Anti-Makeup: Learning A Bi-Level Adversarial Network for Makeup-Invariant Face Verification
• What Face and Body Shapes Can Tell About Height
• Large-scale Bisample Learning on ID Versus Spot Face Recognition
• Face Translation between Images and Videos using Identity-aware CycleGAN
• Learning Disentangling and Fusing Networks for Face Completion Under Structured Occlusions
• Robust RGB-D Face Recognition Using Attribute-Aware Loss
• Adversarial Occlusion-aware Face Detection
• Face Synthesis for Eyeglass-Robust Face Recognition
• Learning Channel Inter-dependencies at Multiple Scales on Dense Networks for Face Recognition
• A Supervised Learning Methodology for Real-Time Disguised Face Recognition in the Wild
• Cosine similarity-based adversarial process
• DeMeshNet: Blind Face Inpainting for Deep MeshFace Verification
• Image Generation from Sketch Constraint Using Contextual GAN
• Attention-Set based Metric Learning for Video Face Recognition
• Self-supervised pre-training with acoustic configurations for replay spoofing detection
• Fast Face Image Synthesis with Minimal Training
• Deep Generative Variational Autoencoding for Replay Spoof Detection in Automatic Speaker Verification
• Global Norm-Aware Pooling for Pose-Robust Face Recognition at Low False Positive Rate