Title

Segmentation Of Neighboring Structures By Modeling Their Interaction

Abstract

The similarity of pixel intensities, presence of noise, existence of partial volume effect and so on make segmentation of neighboring structures in medical image a challenging task. In this paper, we present a novel approach for segmenting multiple neighboring organs simultaneously by modeling interaction between them. Our method is motivated by the observation that radiologists mark ambiguous boundaries by considering all surrounding anatomic structures. In the proposed interaction mode I, the connectedness, competition and repulsion between organs are analyzed. By quantitatively defining these interaction components, an energy functional is formulated and structures are obtained by minimizing the energy. An energy minimization algorithm based on multiway graph cuts is proposed, which obtains an approximation within a known factor of the global minimum. The promising experimental results on various medical images demonstrate the performance of our segmentation algorithm. © 2006 IEEE.

Publication Date

1-1-2006

Publication Title

Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition

Volume

2006

Number of Pages

77-84

Document Type

Article; Proceedings Paper

Personal Identifier

scopus

DOI Link

https://doi.org/10.1109/CVPRW.2006.187

Socpus ID

33845537584 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/33845537584

This document is currently not available here.

Share

COinS