Abstract #1086
Automatic segmentation of breast images using clustering and dynamic programming
Jos Angel Rosado-Toro 1 , Tomoe Barr 2 , Marilyn T Marron 3 , Jean-Phillipe Galons 4 , Patricia Thompson 3 , Alison Stopeck 3 , Jeffrey Joel Rodrguez 5 , and Mara I Altbach 4
1
Electrical and Computer Engineering,
University of Arizona, Tucson, Arizona, United States,
2
Biomedical
Engineering, University of Arizona, Tucson, AZ, United
States,
3
Arizona
Cancer Center, University of Arizona, Tucson, Arizona,
United States,
4
Medical
Imaging, University of Arizona, Tucson, Arizona, United
States,
5
Electrical
and Computer Engineering, University of Arizona, Tucson,
Arizon, United States
A fully automated breast segmentation algorithm has been
developed to segment the breast anatomy using various
types of imaging pulse sequences. The segmentation first
finds the chest and breast pixels using a clustering
technique. Next it removes the chest pixels using a
dynamic programming technique on the vertical gradient.
Then it removes the skin pixels using a thinning
algorithm and finally it splits the two breasts using a
morphological technique. The performance of the
algorithm is evaluated on 202 breast imaging slices
using manually traced breast outlines as reference.
This abstract and the presentation materials are available to members only;
a login is required.
Join Here
