Investigation of object thickness for visual discomfort prediction in stereoscopic images
Hosik Sohn, Yong Ju Jung, Seong-il Lee, Hyun Wook Park, Yong Man Ro · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2012
Visual discomfort assessment metric is of importance in the creation and viewing of stereoscopic 3D contents. This paper investigates the importance of considering object thickness as well as disparity magnitude to predict visual discomfort. Throughout the paper, we introduce the overall process to predict visual discomfort by analyzing the thickness of objects and their disparity magnitude in an image. Using natural stereoscopic images, we evaluate the contribution of object thickness to the prediction performance of visual discomfort. Experimental results demonstrate that the combined use of disparity magnitude and object thickness substantially improves the prediction performance of visual discomfort.