DMAGaze: Gaze Estimation Based on Feature Disentanglement and Multi-Scale Attention

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Haohan, Liu, Hongjia, Lan, Shiyong, Wang, Wenwu, Qiao, Yixin, Li, Yao, Deng, Guonan
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910966082437120
author Chen, Haohan
Liu, Hongjia
Lan, Shiyong
Wang, Wenwu
Qiao, Yixin
Li, Yao
Deng, Guonan
author_facet Chen, Haohan
Liu, Hongjia
Lan, Shiyong
Wang, Wenwu
Qiao, Yixin
Li, Yao
Deng, Guonan
contents Gaze estimation, which predicts gaze direction, commonly faces the challenge of interference from complex gaze-irrelevant information in face images. In this work, we propose DMAGaze, a novel gaze estimation framework that exploits information from facial images in three aspects: gaze-relevant global features (disentangled from facial image), local eye features (extracted from cropped eye patch), and head pose estimation features, to improve overall performance. Firstly, we design a new continuous mask-based Disentangler to accurately disentangle gaze-relevant and gaze-irrelevant information in facial images by achieving the dual-branch disentanglement goal through separately reconstructing the eye and non-eye regions. Furthermore, we introduce a new cascaded attention module named Multi-Scale Global Local Attention Module (MS-GLAM). Through a customized cascaded attention structure, it effectively focuses on global and local information at multiple scales, further enhancing the information from the Disentangler. Finally, the global gaze-relevant features disentangled by the upper face branch, combined with head pose and local eye features, are passed through the detection head for high-precision gaze estimation. Our proposed DMAGaze has been extensively validated on two mainstream public datasets, achieving state-of-the-art performance.
format Preprint
id arxiv_https___arxiv_org_abs_2504_11160
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DMAGaze: Gaze Estimation Based on Feature Disentanglement and Multi-Scale Attention
Chen, Haohan
Liu, Hongjia
Lan, Shiyong
Wang, Wenwu
Qiao, Yixin
Li, Yao
Deng, Guonan
Computer Vision and Pattern Recognition
Artificial Intelligence
Gaze estimation, which predicts gaze direction, commonly faces the challenge of interference from complex gaze-irrelevant information in face images. In this work, we propose DMAGaze, a novel gaze estimation framework that exploits information from facial images in three aspects: gaze-relevant global features (disentangled from facial image), local eye features (extracted from cropped eye patch), and head pose estimation features, to improve overall performance. Firstly, we design a new continuous mask-based Disentangler to accurately disentangle gaze-relevant and gaze-irrelevant information in facial images by achieving the dual-branch disentanglement goal through separately reconstructing the eye and non-eye regions. Furthermore, we introduce a new cascaded attention module named Multi-Scale Global Local Attention Module (MS-GLAM). Through a customized cascaded attention structure, it effectively focuses on global and local information at multiple scales, further enhancing the information from the Disentangler. Finally, the global gaze-relevant features disentangled by the upper face branch, combined with head pose and local eye features, are passed through the detection head for high-precision gaze estimation. Our proposed DMAGaze has been extensively validated on two mainstream public datasets, achieving state-of-the-art performance.
title DMAGaze: Gaze Estimation Based on Feature Disentanglement and Multi-Scale Attention
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2504.11160