Abstract:Feature extractors significantly impact the performance of biometric systems. In the field of hand gesture authentication, existing studies focus on improving the model architectures and behavioral characteristic representation methods to enhance their feature extractors. However, loss functions, which can guide extractors to produce more discriminative identity features, are neglected. In this paper, we improve the margin-based Softmax loss functions, which are mainly designed for face authentication, in two aspects to form a new loss function for hand gesture authentication. First, we propose to replace the commonly used cosine function in the margin-based Softmax losses with a linear function to measure the similarity between identity features and proxies (the weight matrix of Softmax, which can be viewed as class centers). With the linear function, the main gradient magnitude decreases monotonically as the quality of the model improves during training, thus allowing the model to be quickly optimized in the early stage and precisely fine-tuned in the late stage. Second, we design an adaptive margin scheme to assign margin penalties to different samples according to their separability and the model quality in each iteration. Our adaptive margin scheme constrains the gradient magnitude. It can reduce radical (excessively large) gradient magnitudes and provide moderate (not too small) gradient magnitudes for model optimization, contributing to more stable training. The linear function and the adaptive margin scheme are complementary. Combining them, we obtain the proposed linear adaptive additive angular margin (L3AM) loss. To demonstrate the effectiveness of L3AM loss, we conduct extensive experiments on seven hand-related authentication datasets, compare it with 25 state-of-the-art (SOTA) loss functions, and apply it to eight SOTA hand gesture authentication models. The experimental results show that L3AM loss further improves the performance of the eight authentication models and outperforms the 25 losses. The code is available at https://github.com/SCUT-BIP-Lab/L3AM.

Angle-based Softmax Loss for Face Verification

InterFace:Adjustable Angular Margin Inter-class Loss for Deep Face Recognition

X2-Softmax: Margin adaptive loss function for face recognition

Margin Loss: Making Faces More Separable

Class-Variant Margin Normalized Softmax Loss for Deep Face Recognition

Qamface: quadratic additive angular margin loss for face recognition

Inter-class Angular Margin Loss for Face Recognition

Loss Function Search for Face Recognition

HARD SAMPLES BASED MARGIN LOSS FOR FACE VERIFICATION d

Softmax Dissection: Towards Understanding Intra- and Inter-class Objective for Embedding Learning

Angular Softmax Loss for End-to-end Speaker Verification.

A Loss Function Base on Softmax for Expression Recognition

Semantic Softmax Loss for Zero-Shot Learning

An Equalized Margin Loss for Face Recognition.

Chi-square Loss for Softmax: an Echo of Neural Network Structure

Advanced Joint Bayesian Method for Face Verification

L3AM: Linear Adaptive Additive Angular Margin Loss for Video-Based Hand Gesture Authentication

Companion Guided Soft Margin for Face Recognition

SubFace: learning with softmax approximation for face recognition

Learning Towards The Largest Margins.

Adaptive Adversarial Cross-Entropy Loss for Sharpness-Aware Minimization