Object depth adjustment based on planar approximation in stereo images

Abstract

This paper presents a method for adjusting the 3D depth of an object in a pair of stereo images by utilizing a planar approximation in the 3D space. The proposed method first segments an object in the stereo images based on both color and disparity. The segmented object is approximated by a plane in the 3D space and the depth of the object is adjusted by moving the plane in the 3D space, resulting in a new pair of stereo images. The proposed method is evaluated with DQCQS (Double Stimulus Continuous Quality Scale) subjective test and the experimental results show the feasibility of the proposed approach.