Mechanisms and Inaccuracies of 3D Spatial Perception by Visual System
Abstract
Purpose: The lack of clear ideas about how the human visual system forms the sensation of spatial depth forces us to experimentally
determine while shooting every scene whether it is possible to obtain the desired stereoscopic effect, and if so, under what parameters.
The aim of this article is to identify the features of visual perception of 3D spatial depth essential to resolve this problem. Results:
Theoretical and experimental studies have shown that at short observation distances the vision mechanism based on the stereoscopic
effect allows us to estimate distances to objects with lesser errors than the vision mechanism based on perspective transformations.
At long distances, the situation is reversed. An hypothesis was proposed and experimentally confirmed which explained how the
sensation of spatial depth appears while watching real scenes and stereoscopic images by "converging-diverging" eye movements. This led to the development of a method determining the optimal conditions of stereoscopic shooting giving the best stereo effect, as well as the technique of choosing the optimal parameters of image/video conversion from 2D into 3D. It was shown that the depth map of a stereoscopic image tolerates significant inaccuracies which do not lead to any noticeable distortion of the resulting image. Practical relevance: The obtained result makes it possible to correct the existing stereoscopic images and movies. When we know that the depth map of a stereoscopic image tolerates inaccuracies without any noticeable distortion, we can considerable simplify the procedure of converting 2D images into 3D.
