Detection and Segmentation of Masses in Mammograms by The Rule Based Elimination Approach


Ture H., KAYIKÇIOĞLU T.

25th Signal Processing and Communications Applications Conference (SIU), Antalya, Turkey, 15 - 18 May 2017 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/siu.2017.7960440
  • City: Antalya
  • Country: Turkey

Abstract

In this study, a method was proposed that eliminated the non-suspicious salient regions for the detection and segmentation of masses in mammograms. Since suspicious regions are generally salient dense regions, the method firstly extracts the maximum regions of interest (ROIs) that have the optimum lifetime. Subsequently, these ROIs are segmented with the rule based elimination using morphological and intensity properties. The texture features taken from the suspicious regions are classified by Rus Boost method for detection of masses. The developed method has been tested on all mammograms, which includes mass, taken from the MIAS database. Experimental results demonstrate that the method achieves a satisfactory performance during the detection and segmentation of suspicious regions.