Open AccessMotion depth generation using MHI for 2D‐to‐3D video conversion
Author(s) -
Gil J.,
Kim M.
Publication year - 2017
Publication title -
electronics letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 146
ISSN - 1350-911X
DOI - 10.1049/el.2017.1505
Subject(s) - motion estimation , computer vision , artificial intelligence , quarter pixel motion , computer science , bottleneck , rendering (computer graphics) , structure from motion , motion (physics) , motion field , stereoscopy , 2d to 3d conversion , block matching algorithm , depth map , computation , image (mathematics) , computer graphics (images) , algorithm , video processing , video tracking , embedded system
2D‐to‐3D conversion has been studied over past decades and integrated to commercial 3D displays and 3DTVs. Generally, depth cues extracted from a static image are used for generating a depth map followed by depth image‐based rendering for producing a stereoscopic image. Further, the motion has been considered as an important cue for motion depth estimation. In most works, motion estimation has relied on block‐based motion estimation, optical flows, and their variants even though they provide inaccurate data and high computation time, posing performance bottleneck. These problems by proposing Motion History Image ‐based motion depth estimation method are addressed. Experimental results show that the proposed method is not only faster than the conventional methods but also outperforms them in terms of motion depth estimation.