一种基于眼球模型的3D视线估计方法及系统
By constructing an eye model using a few-sample computation and sparse sampling strategy, the problem of high computational cost and inability to update the eye model in real time in existing technologies is solved, achieving high-precision and fast 3D gaze estimation and improving the stability and reliability of the system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHEJIANG UNIV
- Filing Date
- 2025-07-29
- Publication Date
- 2026-07-17
AI Technical Summary
Existing 3D gaze estimation methods based on eye models suffer from sliding errors, high computational costs, and the inability to update in real time, resulting in significant gaze estimation errors and making it difficult to meet the needs of real-time applications.
Employing a few-sample computation and sparse sampling strategy, the system constructs a multi-frame ellipse set, calculates the projection point of the eyeball center on the image plane, builds an eyeball model, identifies abnormal frames, and filters out bad frames caused by camera slippage.
It achieves high line-of-sight estimation accuracy with an error of no more than 1°, fast model fitting speed with a time of no more than 0.05ms, and improves system stability and the reliability of estimation results.
Smart Images
Figure CN120954077B_ABST