Contrast Adjusting and K-Means Clustering to Isolate Tumors in Mammograms

Authors

  • Hussein Saadi Kareem, Rabab Saadoon Abdoon

Abstract

- Radiologists do a very important role in disease recognition based on the increasing utilization of medical images. Medical scanning devices such as mammography yield more images to patients which make the radiologist's work even more burdensome. It is, therefore, necessary to help the radiologist to obtain a better and faster diagnosis. In this work, two methods were proposed for the segmenting three mammography images. The first method is the contrast adjusting based technique by implementing spatial domain enhancement technique and the second method is K-means clustering in both methods utilized many morphological operations to isolate and extract breast tumors. The result of these two methods showed good procedures to isolate and extract breast tumors according to the consultation of the radiologist and have good agreement with the radiologist's delineation. The percent relative differences for the experimental images were reached from 0.8085 to 0.8338 % for the first technique and reached from 0.3731 to 0.7891 % for the second technique.                    

Published

2020-11-01

Issue

Section

Articles