Simulation of Depth Interpolation from Surface Boundary in Binocular Viewing
Weifu Shi* and Masanori Idesawa**
*Nikon Corporation, 1-6-3, Nishi-oki, Shinagawa-ku, Tokyo, 140, Japan
**Graduate School of Information Sysyems, The University of Electro-Communications, 1-5-1, Chofugaoka, Chofu-shi, Tokyo, 182, Japan
With binocular viewing, the human visual system can perceive depth information of a 3-D object from several kinds of cues, such as binocular disparity, occlusion, optical flow, perspective, texture gradients, shading, and motion. Several years ago, in relation to the 3-D illusion, we found that surface depth can also be perceived even if no apparent visual stimulus occurs inside the surface boundary that provides the cues above except for binocular disparity given partially along the surface boundary. In addition, we found that, in some cases, shapes could be perceived in different forms from entirely the same surface boundary when the attention depth is changed or the attention point is presented in binocular viewing. We built a simple, preliminary mathematical model to simulate surface depth perception. It was based on the following hypotheses: activity potential sources are located along the perceived surface boundary; cells inside the column corresponding to the view volume enclosed by the surface boundary are activated by the summation of the activation power emitted from the activity potential sources; the surface then is perceived at the depth where the most activated cell is observed in depth. We then improved our mathematical model to introduce the effects of attention points. In this paper, we describe the mathematical model in detail and present simulation results for different surface shapes. We also discuss the influence on surface shape by changing the parameters of the mathematical model.