Disparity estimation and segmentation are vital tasks in stereo image coding and stereo vision matching.This paper proposes an algorithm for disparity estimation based on hierarchical MRF/GRF model and overlapped block matching (HMOM)

and for disparity segmentation and extraction using active contour model.By combining the advantages of variable size block matching

overlapped block matching and MRF model

the HMOM scheme estimates a more smoother and accurate disparity field with lower computational complexity.In terms of PSNR of the disparity compensated image

the proposed disparity estimation algorithm achieves about 2.5dB higher PSNR

as compared to conventional variable block matching.A smooth and consistent disparity field with sharp boundary can be obtained by pixel-wise disparity refinement on the HMOM disparity field

and the disparity contour of OOI (objects of interest) can be extracted by the active contour model.The resulting disparity field and corresponding disparity contours can further be used in stereo image coding and video object segmentation.