Depth image improvement for 3D applications

Abstract

This paper uses the area-based canny edge detector to keep the boundaries intact while applying the boundary-based depth image improvement technique to the flat areas. The proposed method will provide a better sense of depth by addressing image quality deterioration, which may take place in acquiring depth images in a limited environment, through post-or pre-treatment. We expect such improvement in depth images to be applied to various 3D applications, such as depth-based 3D body gesture recognition/interaction, depth-based image rendering (DBIR), 3D TV. © 2011 IEEE.