MixFace: Improving Face Verification Focusing on Fine-grained Conditions

TL;DR

This paper introduces MixFace, a new loss function for face verification that enhances robustness under fine-grained conditions, validated through experiments on a specialized dataset and benchmarks.

Contribution

The paper proposes MixFace, a novel loss function combining classification and metric losses, specifically designed to improve face verification under fine-grained conditions.

Findings

MixFace outperforms existing methods in robustness and effectiveness.

Fine-grained conditions significantly impact face recognition performance.

Experimental validation on K-FACE and benchmarks confirms MixFace's superiority.

Abstract

The performance of face recognition has become saturated for public benchmark datasets such as LFW, CFP-FP, and AgeDB, owing to the rapid advances in CNNs. However, the effects of faces with various fine-grained conditions on FR models have not been investigated because of the absence of such datasets. This paper analyzes their effects in terms of different conditions and loss functions using K-FACE, a recently introduced FR dataset with fine-grained conditions. We propose a novel loss function, MixFace, that combines classification and metric losses. The superiority of MixFace in terms of effectiveness and robustness is demonstrated experimentally on various benchmark datasets.