GazeOnce360: Fisheye-Based 360 ∘ Multi-Person Gaze Estimation with Global-Local Feature Fusion

Zhuojiang CaiZhenghui Sun, Feng Lu State Key Laboratory of VR Technology and Systems, School of CSE, Beihang University {caizhuojiang, sunzhenghui, lufeng}@buaa.edu.cn · 2026

Read GazeOnce360: Fisheye-Based 360 ∘ Multi-Person Gaze Estimation with Global-Local Feature Fusion by Zhuojiang CaiZhenghui Sun, Feng Lu State Key Laboratory of VR Technology and Systems, School of CSE, Beihang University {caizhuojiang, sunzhenghui, lufeng}@buaa.edu.cn (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-17161.txt.

Description

Cai, Sun, and Lu present GazeOnce360, a model that estimates the gaze of many people from one upward facing fisheye camera. Standard gaze systems use forward facing cameras with narrow views. Their setting covers a full 360 degree scene around a tabletop camera. They built MPSGaze360, a large synthetic dataset rendered in Unreal Engine with 3D gaze and eye landmark labels. The model uses rotational convolutions to handle fisheye distortion. A dual resolution design fuses global context with high resolution eye regions.

Catalog record

CREATORZhuojiang CaiZhenghui Sun, Feng Lu State Key Laboratory of VR Technology and Systems, School of CSE, Beihang University {caizhuojiang, sunzhenghui, lufeng}@buaa.edu.cn
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE42541
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.17161

Browse the full library · About the Tower of Babel