A Performance Comparison of Loss Functions for Deep Face Recognition

TL;DR

This paper compares the performance of various loss functions used in CNN-based face recognition, evaluating their effectiveness across different architectures and datasets to identify the most suitable loss functions for improved accuracy.

Contribution

It provides a comprehensive comparison of multiple loss functions for face recognition, highlighting their relative strengths and weaknesses across different CNN architectures and datasets.

Findings

ArcFace achieves the highest accuracy among tested loss functions.

Additive-Margin Softmax improves discriminative power.

Performance varies with CNN architecture and dataset used.

Abstract

Face recognition is one of the most widely publicized feature in the devices today and hence represents an important problem that should be studied with the utmost priority. As per the recent trends, the Convolutional Neural Network (CNN) based approaches are highly successful in many tasks of Computer Vision including face recognition. The loss function is used on the top of CNN to judge the goodness of any network. In this paper, we present a performance comparison of different loss functions such as Cross-Entropy, Angular Softmax, Additive-Margin Softmax, ArcFace and Marginal Loss for face recognition. The experiments are conducted with two CNN architectures namely, ResNet and MobileNet. Two widely used face datasets namely, CASIA-Webface and MS-Celeb-1M are used for the training and benchmark Labeled Faces in the Wild (LFW) face dataset is used for the testing.