Abstract
A graph based method is introduced to segment the pectoral muscles in screening mammograms. An adaptive pyramid is used to segment the mammogram into a number of components. Components forming the pectoral muscle are identified based on position, intensity, and shape. The boundary of the union of these components forms an initial boundary that is refined via an adaptive deformable contour model. The method is tested on 83 medio-lateral oblique mammograms from the Mini-MIAS database. Segmentation results are evaluated in terms of the proportion of correctly assigned pixels. Performance compares well with existing methods based on Hough transform and on Gabor
wavelets.
wavelets.
Original language | English |
---|---|
Title of host publication | Proceedings of the APRS Workshop on Digital Image Computing |
Editors | Brian C Lovell, Anthony J Maeder |
Place of Publication | St Lucia, Queensland |
Publisher | The University of Queensland |
Number of pages | 6 |
ISBN (Print) | 0-9580255-3-3 |
Publication status | Published - Feb 2005 |
Event | APRS Workshop on Digital Image Computing - Duration: 21 Feb 2005 → … |
Conference
Conference | APRS Workshop on Digital Image Computing |
---|---|
Period | 21/02/05 → … |