Learning Unsupervised Gaze Representation via Eye Mask Driven Information Bottleneck

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jiang, Yangzhou, Lin, Yinxin, Wang, Yaoming, Li, Teng, Ke, Bilian, Ni, Bingbing
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911938141749248
author Jiang, Yangzhou
Lin, Yinxin
Wang, Yaoming
Li, Teng
Ke, Bilian
Ni, Bingbing
author_facet Jiang, Yangzhou
Lin, Yinxin
Wang, Yaoming
Li, Teng
Ke, Bilian
Ni, Bingbing
contents Appearance-based supervised methods with full-face image input have made tremendous advances in recent gaze estimation tasks. However, intensive human annotation requirement inhibits current methods from achieving industrial level accuracy and robustness. Although current unsupervised pre-training frameworks have achieved success in many image recognition tasks, due to the deep coupling between facial and eye features, such frameworks are still deficient in extracting useful gaze features from full-face. To alleviate above limitations, this work proposes a novel unsupervised/self-supervised gaze pre-training framework, which forces the full-face branch to learn a low dimensional gaze embedding without gaze annotations, through collaborative feature contrast and squeeze modules. In the heart of this framework is an alternating eye-attended/unattended masking training scheme, which squeezes gaze-related information from full-face branch into an eye-masked auto-encoder through an injection bottleneck design that successfully encourages the model to pays more attention to gaze direction rather than facial textures only, while still adopting the eye self-reconstruction objective. In the same time, a novel eye/gaze-related information contrastive loss has been designed to further boost the learned representation by forcing the model to focus on eye-centered regions. Extensive experimental results on several gaze benchmarks demonstrate that the proposed scheme achieves superior performances over unsupervised state-of-the-art.
format Preprint
id arxiv_https___arxiv_org_abs_2407_00315
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Learning Unsupervised Gaze Representation via Eye Mask Driven Information Bottleneck
Jiang, Yangzhou
Lin, Yinxin
Wang, Yaoming
Li, Teng
Ke, Bilian
Ni, Bingbing
Computer Vision and Pattern Recognition
Appearance-based supervised methods with full-face image input have made tremendous advances in recent gaze estimation tasks. However, intensive human annotation requirement inhibits current methods from achieving industrial level accuracy and robustness. Although current unsupervised pre-training frameworks have achieved success in many image recognition tasks, due to the deep coupling between facial and eye features, such frameworks are still deficient in extracting useful gaze features from full-face. To alleviate above limitations, this work proposes a novel unsupervised/self-supervised gaze pre-training framework, which forces the full-face branch to learn a low dimensional gaze embedding without gaze annotations, through collaborative feature contrast and squeeze modules. In the heart of this framework is an alternating eye-attended/unattended masking training scheme, which squeezes gaze-related information from full-face branch into an eye-masked auto-encoder through an injection bottleneck design that successfully encourages the model to pays more attention to gaze direction rather than facial textures only, while still adopting the eye self-reconstruction objective. In the same time, a novel eye/gaze-related information contrastive loss has been designed to further boost the learned representation by forcing the model to focus on eye-centered regions. Extensive experimental results on several gaze benchmarks demonstrate that the proposed scheme achieves superior performances over unsupervised state-of-the-art.
title Learning Unsupervised Gaze Representation via Eye Mask Driven Information Bottleneck
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.00315