One-shot Face Recognition by Promoting Underrepresented Classes

By Yandong Guo, Lei Zhang. 2017

108 citations recorded by the source corpus.

In this paper, we study the problem of training large-scale face identification model with imbalanced training data. This problem naturally exists in many real scenarios including large-scale celebrity recognition, movie actor annotation, etc. Our solution contains two components. First, we build a face feature extraction model, and improve its performance, especially for the persons with very limited training samples, by introducing a regularizer to the cross entropy loss for the multi-nomial logistic regression (MLR) learning. This regularizer encourages the directions of the face features from the same class to be close to the direction of their corresponding classification weight vector in the logistic regression. Second, we build a multi-class classifier using MLR on top of the learned face feature extraction model. Since the standard MLR has poor generalization capability for the one-shot classes even if these classes have been oversampled, we propose a novel supervision signal called underrepresented-classes promotion loss, which aligns the norms of the weight vectors of the one-shot classes (a.k.a. underrepresented-classes) to those of the normal classes. In addition to the original cross entropy loss, this new loss term effectively promotes the underrepresented classes in the learned model and leads to a remarkable improvement in face recognition performance. We test our solution on the MS-Celeb-1M low-shot learning benchmark task. Our solution recognizes 94.89% of the test images at the precision of 99\% for the one-shot classes. To the best of our knowledge, this is the best performance among all the published methods using this benchmark task with the same setup, including all the participants in the recent MS-Celeb-1M challenge at ICCV 2017.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

iTAML: An Incremental Task-Agnostic Meta-learning Approach (2020, background): “MS-Celeb-10K is a subset of MS-Celeb-1M dataset [8].”
Large Scale Incremental Learning (2019, background): “Celeb-10000: a random subset of 10,000 classes are selected from MS-Celeb-1M-base [5] face dataset which has 20,000 classes.”
Incremental Classifier Learning with Generative Adversarial Networks (2018, methodology): “Our method has excellent results outperforming the state of arts on three datasets CIFAR-100, FLower-102 and MS-Celeb-1M-Base by a large margin.”
Deep Face Recognition: A Survey (2018, methodology): “[58] trained the softmax classifier combined with underrepresentedclasses promotion (UP) loss term to enhance the performance.”
Deep Imbalanced Learning for Face Recognition and Attribute Prediction (2018, background): “centers [49] or weight norms [50] of the minority classes to the majority.”
Past, Present, and Future of Face Recognition: A Review (2020, methodology): “Several promising ideas have been explored to bring advances in CNNs, such as the use of proper activation [143] and various loss functions [15,124–126,130,134,136,138,141,142,145,148–152], the use of metric learning algorithms [38,127], normalization of features and weights [133,135,139], extractio”
BBN: Bilateral-Branch Network With Cumulative Learning for Long-Tailed Visual Recognition (2019, methodology): “Following [9], we visualize the 2-norm of these classifiers.”
Calibrating CNNs for Lifelong Learning (2020, methodology): “We perform experiments on the SVHN [28], CIFAR [29], ImageNet [30], and MS-Celeb-10K [31] datasets.”
Boosting Adversarial Training with Hypersphere Embedding (2020, methodology): “WN normalizes the weights in the softmax layer to encourage the model to equally treat samples from different classes, facilitating the cases with unbalanced training data (Guo & Zhang, 2017; Liu et al., 2017; Pang et al., 2018; 2020).”
Feature Transfer Learning for Face Recognition With Under-Represented Data (2019, methodology): “Compared to methods [6, 11] with similar setting, we achieve competitive performance on the base classes and much better accuracy on the novel classes by 15% .”