Markov random field model for mammogram segmentation

A method for the segmentation of digital mammograms is proposed for the identification of tumors or other suspicious areas. The proposed technique uses a Markov random field combined with a Gaussian process for modelling an image. The 'iterative conditional modes' method is used to get an...

Full description

Saved in:
Bibliographic Details
Main Authors: Huai-Dong Li, Kallergi, M., Clarke, L.P., Wei Qian, Clark, R.A.
Format: Conference Proceeding
Language:English
Subjects:
Cancer
Gaussian processes
Image segmentation
Iterative algorithms
Iterative methods
Mammography
Markov random fields
Pixel
Radiology
Shape
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A method for the segmentation of digital mammograms is proposed for the identification of tumors or other suspicious areas. The proposed technique uses a Markov random field combined with a Gaussian process for modelling an image. The 'iterative conditional modes' method is used to get an approximate maximum a posterior solution to the segmentation problem. The performance of the proposed algorithm on thirteen mammograms indicated that the new technique has significant potential for assisting the diagnosis of tumors in digital mammography.
DOI:10.1109/IEMBS.1993.978426