SegFace: Face Segmentation of Long-Tail Classes

TL;DR

SegFace introduces a transformer-based face segmentation model that effectively handles long-tail classes, significantly improving segmentation accuracy for infrequent facial regions while maintaining high efficiency for real-time applications.

Contribution

This work is the first to apply transformer models with class-specific tokens to face parsing, addressing long-tail class segmentation issues and enabling efficient, high-performance face segmentation.

Findings

Achieves a mean F1 score of 88.96 on CelebAMask-HQ

Outperforms previous state-of-the-art models

Runs at 95.96 FPS on edge devices

Abstract

Face parsing refers to the semantic segmentation of human faces into key facial regions such as eyes, nose, hair, etc. It serves as a prerequisite for various advanced applications, including face editing, face swapping, and facial makeup, which often require segmentation masks for classes like eyeglasses, hats, earrings, and necklaces. These infrequently occurring classes are called long-tail classes, which are overshadowed by more frequently occurring classes known as head classes. Existing methods, primarily CNN-based, tend to be dominated by head classes during training, resulting in suboptimal representation for long-tail classes. Previous works have largely overlooked the problem of poor segmentation performance of long-tail classes. To address this issue, we propose SegFace, a simple and efficient approach that uses a lightweight transformer-based model which utilizes learnable class-specific tokens. The transformer decoder leverages class-specific tokens, allowing each token to focus on its corresponding class, thereby enabling independent modeling of each class. The proposed approach improves the performance of long-tail classes, thereby boosting overall performance. To the best of our knowledge, SegFace is the first work to employ transformer models for face parsing. Moreover, our approach can be adapted for low-compute edge devices, achieving 95.96 FPS. We conduct extensive experiments demonstrating that SegFace significantly outperforms previous state-of-the-art models, achieving a mean F1 score of 88.96 (+2.82) on the CelebAMask-HQ dataset and 93.03 (+0.65) on the LaPa dataset. Code: https://github.com/Kartik-3004/SegFace