Angular Margin-Mining Softmax Loss for Face Recognition

Abstract

Face recognition methods have been significantly improved in recent years owing to the advances made in loss functions. Typically, loss functions are designed to enhance the separability power by concentrating on hard samples in mining-based approaches or by increasing the feature margin between different classes in margin-based approaches. However, margin-based methods lack the utilization of informative hard sample, and mining-based methods also fail to learn the latent correlations between classes. Moreover, there are no methods that simultaneously consider the effects of hard samples and feature margin through the same shape of feature angular margin. Therefore, this paper introduces the Angular Margin-Mining Softmax (AMM-Softmax) loss function, which adaptively emphasizes hard samples while also increasing the decision margins. The proposed AMM-Softmax loss function introduces a linear angular margin for hard samples, enabling the direct optimization of the geodesic distance margin and maximization of class separability. Furthermore, the proposed AMM-Softmax loss function is computationally efficient and can be easily converged by rapidly switching from the hard samples to easy samples. The results of the extensive experimental analyses conducted on popular benchmarks demonstrate the superiority of the proposed AMM-Softmax loss function over the existing state-of-the-art methods.

Enhancing the Classification Accuracy by Intra-Concentration and the Distance between the Class Boundaries Instead of the Class Centers

“…The center loss [7] simultaneously learns a center for deep features of each class and penalizes the distances between the deep features and their corresponding class centers. However, the center loss faces the drawbacks of high computational cost and sensitivity to the sample selecting strategy, which significantly affects the performance [18]. A-Softmax loss [19] tries to incorporate the angular margin into the softmax loss function, which has achieved promising performance.…”

“…Due to the values of weights and features being unbounded, samples in different classes still can overlap. AMM-Softmax [18] design a linear angular margin for hard samples, which allows the optimization of the geodesic distance margin and maximization of class separability. According to the article the consistency of the margin provides stability and acceleration in the training too [18].…”

“…AMM-Softmax [18] design a linear angular margin for hard samples, which allows the optimization of the geodesic distance margin and maximization of class separability. According to the article the consistency of the margin provides stability and acceleration in the training too [18]. The L-GM [24] loss is used to distinguish abnormal inputs by estimating the likelihood of an input based on its distance from the learned training feature distribution.…”