3D SMAP Algorithm

Abstract

This report provides an extension of the original Bouman and Shapiro’s SMAP paper for segmenting 3D image volume into regions of distinct statistical behavior. Specifically, we extend the structure of MSRF into 3D. The prior distribution is modified to satisfy the MRF property. In addition, an iterative strategy is proposed to refine the segmentation.

Publication
ECE, Purdue University
Date
Links

alt text