Harnessing Geometric Constraints from Emotion Labels to improve Face Verification

TL;DR

This paper proposes leveraging facial emotion labels as auxiliary information to impose geometric constraints on face embeddings, enhancing verification performance without altering the neural network architecture.

Contribution

It introduces novel loss functions that incorporate emotion-based geometric constraints, improving face verification models using existing architectures.

Findings

Enhanced verification accuracy with emotion-based constraints

Effective use of manually annotated or automatic emotion labels

No changes needed to existing neural network backbones

Abstract

For the task of face verification, we explore the utility of harnessing auxiliary facial emotion labels to impose explicit geometric constraints on the embedding space when training deep embedding models. We introduce several novel loss functions that, in conjunction with a standard Triplet Loss [43], or ArcFace loss [10], provide geometric constraints on the embedding space; the labels for our loss functions can be provided using either manually annotated or automatically detected auxiliary emotion labels. Our method is implemented purely in terms of the loss function and does not require any changes to the neural network backbone of the embedding function.