A Way to Improve Pectoral Muscle Isolation in Digital Mammograms

Main Article Content

Sara Dehghani
Mashallah Abbasi, Amir Masoud Rahmani

Abstract

In most mammography images taken from the side view, there is a muscle above, which its gray surface is almost like tumors and some denser breast tissues. It is worth noting that the tumors wouldn't exist within the muscles, but may be very close to it. So it is better to separate the muscle before processing the whole image. In this project we have separated the muscle from the outskirt tissue by assimilating statistical methods and techniques to increase the zone of dealt. This new method was accomplished on 60 images of Mini MIAS database. To evaluate the accuracy of separation, first the range of muscles is marked by a physician or radiologist and is considered as a gold standard. The percent of overlapping between identifying by the computer and golden standard indicates the system accuracy. The results report accuracy of 90%. The benefits of this method are simplicity of implementation and its high speed.

 

 

Keywords: Mammography, pectoral muscles, breast cancer, statistical methods, region growing methods

Downloads

Download data is not yet available.

Article Details

Section
Articles