10 April 1996 Disparity estimation with object-contour information for synthesizing intermediate view images
Author Affiliations +
Abstract
To synthesize intermediate view images between a pair of left and right view images of an object, we propose a new method which estimates disparity with discontinuity at the object contour. Our method consists of four steps: (1) Two initial disparity maps based on the right or left view images are estimated integrating the various block size correlations. (2) Ill- corresponding areas in both of the initial disparity maps are detected with `confidence measures'. Then, the disparity of the il-corresponding areas is estimated from the boundary disparity and the object contours are detected using the Snake method for both view images. (3) The disparity maps obtained above are reestimated along the points of the object contours. (4) Using the reestimated disparity maps, two intermediate view images based on both view images are synthesized and integrated into the final intermediate view image. Our method can estimate the disparity of an occluded area with discontinuity at the object contour. Experimental results show the proposed method improves the quality of the intermediate view images, especially in the region of object contours, and its MSE is reduced to 30%.
© (1996) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Takeo Azuma, Kenya Uomori, Atsushi Morimura, "Disparity estimation with object-contour information for synthesizing intermediate view images", Proc. SPIE 2653, Stereoscopic Displays and Virtual Reality Systems III, (10 April 1996); doi: 10.1117/12.237434; https://doi.org/10.1117/12.237434
PROCEEDINGS
7 PAGES


SHARE
Back to Top