基于瞳孔直径的自由立体观看视疲劳预测

Prediction of Autostereoscopic Display Visual Fatigue Based on Pupil Diameter

自由立体显示不需要用助视设备就能带给人们高临场感的视觉体验,但随之而来的却是不能忽视的视疲劳问题.提出了基于瞳孔直径的立体观看视疲劳的预测方法,采用方差分析确定对视疲劳有显著影响的因素,并进行视疲劳主客观评价实验.根据评价结果建立被试者的主观视疲劳程度、客观视疲劳程度分别与3D视频视差最大值、视频帧率、环境亮度、观看角度及被试者年龄构成的2个多因素回归模型.结果显示,此2个模型预测到的视疲劳值与主观、客观评价结果之间的线性相关系数均接近于1,表明此类模型评分接近人类视觉系统的评分,并能代替主观、客观评价从3D观看情形中快速且准确地预测立体观看视疲劳程度.

Autostereoscopic display brings people a high sense of presence without the use of visual aids,but the consequent problem of visual fatigue cannot be ignored.A prediction method of visual fatigue based on pupil diameter is proposed,and then the method of ANOVA is adopted to determine the factors that have a significant impact on visual fatigue.After that,the experiment on the subjective and objective evaluation of visual fatigue is carried out.According to the evaluation results,two multi-factor regression models are established,which represent the relationship between the subjective as well as objective visual fatigue of subjects and the maximum parallax of 3D video,frame rate,environmental brightness,viewing angle,and the age of subjects.The results show that the linear correlation coefficient between the visual fatigue predicted by the two models and the evaluation result is close to 1.It means that the score of the model is closer to that of the human visual system.This model can replace subjective and objective evaluation to predict the degree of visual fatigue rapidly and accurately from 3D viewing situation.