Depth Estimation for 2D-to-3D Image Conversion Using Scene Feature

Abstract

In this modern era 3D supportive hardware popularity is increased but the demand for 3D contents and there availability is not matching. They are still dominated by its 2D counterpart hence there is need of 3D contents. While doing 2D-to-3D image or video conversion depth estimation is a key step and a bit challenging procedure. There are distinct parameters that can be considered during conversion like, structure from motion, defocus, perspective geometry, etc. Until now many researchers have been proposed different methods to close this gap by considering one or many parameters. In this paper for depth estimation, conversion using scene feature is used. Here color is chosen as a scene feature. Intensity information is used here to estimate depth image, hence RGB to HSV conversion is performed from which Value (V) deals with intensity information. RGB to HSV conversion is implemented on FPGA. The proposed method is prototyped on Spartan 3E FPGA based developing board and MATLAB.
DOI: 10.17762/ijritcc2321-8169.150688