Zheng Xiu Juan, Li Zhan Heng, Chun Xin Yi, Yang Xiao Mei, Liu Kai
College of Electrical Engineering, Sichuan University, Chengdu 610065, China.
Math Biosci Eng. 2019 Nov 26;17(2):1396-1412. doi: 10.3934/mbe.2020071.
The eyeball distortions caused by eye diseases, such as myopia and strabismus, can lead to the deviations of eye-tracking data. In this paper, a model-based method with geometric solutions is proposed for gaze correction. The deviations of estimated gaze points are geometrically analyzed based on the individual eyeball models with considerations of the distortions caused by myopia and strabismus. A set of integrated geometric solutions is derived from the varied situations including the case of strabismus and the case of myopia and strabismus, and then used for gaze correction in eyetracking. The experimental results demonstrate that this model-based method is effective to reduce deviations in estimated gaze points, and can be used to correct the modeling error in eye-tracking. Moreover, the proposed method has the potential to provide a simple approach to correct the eyetracking data for various populations with eye diseases.
由近视和斜视等眼部疾病引起的眼球变形会导致眼动追踪数据出现偏差。本文提出了一种基于模型的几何求解方法用于注视校正。基于个体眼球模型，考虑近视和斜视引起的变形，对估计注视点的偏差进行几何分析。从包括斜视情况以及近视和斜视情况等多种情形中推导得出一组综合几何解，然后将其用于眼动追踪中的注视校正。实验结果表明，这种基于模型的方法能有效减少估计注视点的偏差，可用于校正眼动追踪中的建模误差。此外，所提出的方法有可能为患有眼部疾病的各类人群校正眼动追踪数据提供一种简单方法。
