为准确反映人眼在观看立体视频过程中的视觉舒适程度, 提出了一种基于特征区域分割的立体视频视觉舒适度评价方法。首先自适应地分割提取出立体视频序列中的人眼的感兴趣区域, 以及相邻帧间的运动区域, 然后结合深度感知理论和空间合并技术, 对视觉舒适度这一特征量进行建模, 最后实现对立体视频视觉舒适度的评定。实验结果表明, 使用本文的方法对立体视频视觉舒适度进行评价, 可以很好的模拟人眼直接观测的视觉感受, 对于视觉舒适度的研究具有重要的参考价值。

First, adaptive segment and extraction of the interested area of human eye from the sequence of stereoscopic video is performed and the motion of adjacent frame view is captured. Then, depth perception is combined based on the theory with spatial pooling technology andthe characteristics of visual comfort is modeled. Finally, the assessment of stereoscopic video comfort is completed .It is experimentally shown that the method of assessing the stereoscopic video visual comfort can simulate the perception of human eye and it is an important reference to thethe study of visual comfort.