Visual comfort metric based on the motion characteristics of salient object in stereoscopic 3D video content

Abstract

The increasing interest in stereoscopic 3D contents has triggered a demand for building guidelines for the safe creation and viewing of stereoscopic 3D contents. As one of method for building safety guideline, it is very important to build a metric that is able to automatically quantify the degree of visual comfort. Among well-known causes of visual discomfort such as binocular disparity, temporal variation of binocular disparity and binoc-ular asymmetry, we deal with temporal variation of binocular disparity. Temporal variation of binocular disparity can arise from various reasons, including camera motion, convergence variation of toed-in camera, object motion, and so on. This paper investigates the visual discomfort induced by object motion. The major contribution of this paper is the de-velopment of a novel visual comfort metric based on the motion characteristics of object. The proposed visual comfort metric allows modeling motion characteristics of the atten-tional object by exploiting the concept of visual saliency. An extensive number of subjective experiments have been conducted to construct and validate our visual comfort metric. Experimental results show that the correlation coefficient is as high as 0.83.