SphereFace2: Binary Classification is All You Need for Deep Face Recognition

TL;DR

SphereFace2 introduces a binary classification framework for deep face recognition that overcomes softmax limitations, improving representation quality and outperforming existing methods on benchmarks.

Contribution

The paper proposes SphereFace2, a novel binary classification approach that bypasses softmax normalization, aligning training and evaluation for better face recognition performance.

Findings

Consistently outperforms state-of-the-art methods on benchmarks

Effectively bridges the training-evaluation gap in face recognition

Provides principles for designing effective binary classification frameworks

Abstract

State-of-the-art deep face recognition methods are mostly trained with a softmax-based multi-class classification framework. Despite being popular and effective, these methods still have a few shortcomings that limit empirical performance. In this paper, we start by identifying the discrepancy between training and evaluation in the existing multi-class classification framework and then discuss the potential limitations caused by the "competitive" nature of softmax normalization. Motivated by these limitations, we propose a novel binary classification training framework, termed SphereFace2. In contrast to existing methods, SphereFace2 circumvents the softmax normalization, as well as the corresponding closed-set assumption. This effectively bridges the gap between training and evaluation, enabling the representations to be improved individually by each binary classification task. Besides designing a specific well-performing loss function, we summarize a few general principles for this "one-vs-all" binary classification framework so that it can outperform current competitive methods. Our experiments on popular benchmarks demonstrate that SphereFace2 can consistently outperform state-of-the-art deep face recognition methods. The code has been made publicly available.