Classification of benign and malignant patterns in digital mammograms for the diagnosis of breast cancer

Verma, Brijesh and McLeod, Peter and Klevansky, Alan (2010) Classification of benign and malignant patterns in digital mammograms for the diagnosis of breast cancer. Expert Systems with Applications, 37 (4). pp. 3344-3351. ISSN 0957-4174

Abstract

The classification of benign and malignant patterns in digital mammograms is one of most important and significant processes during the diagnosis of breast cancer as it helps detecting the disease at its early stage which saves many lives. Breast abnormalities are often embedded in and camouflaged by various breast tissue structures. It is a very challenging and difficult task for radiologists to correctly classify suspicious areas (benign and malignant patterns) in digital mammograms. In the early stage, the visual clues are subtle and varied in appearance, making diagnosis difficult; challenging even for specialists. Therefore, an intelligent classifier is required which can help radiologists in classifying suspicious areas and diagnosing breast cancer. This paper investigates a novel soft clustered based direct learning classifier which creates soft clusters within a class and learns using direct calculation of weights. The feature space for suspicious areas in digital mammograms from same class patterns can have multiple clusters and the proposed classifier uses this fact and introduces a novel idea to create soft clusters for each available class and applies them to form sub-classes within benign and malignant classes. A novel learning process based on direct learning is introduced. The experiments using the proposed classifier have been conducted on a benchmark database. The results have been analysed using ANOVA test which showed that the results are statistically significant.


Statistics for USQ ePrint 26520
Statistics for this ePrint Item
Item Type: Article (Commonwealth Reporting Category C)
Refereed: Yes
Item Status: Live Archive
Additional Information: Permanent restricted access to Published version due to publisher copyright policy.
Faculty / Department / School: Current - USQ Other
Date Deposited: 08 Jan 2015 02:36
Last Modified: 13 Sep 2016 04:25
Uncontrolled Keywords: intelligent classifiers; neural networks; digital mammography
Fields of Research : 08 Information and Computing Sciences > 0801 Artificial Intelligence and Image Processing > 080109 Pattern Recognition and Data Mining
10 Technology > 1004 Medical Biotechnology > 100402 Medical Biotechnology Diagnostics (incl. Biosensors)
11 Medical and Health Sciences > 1103 Clinical Sciences > 110320 Radiology and Organ Imaging
Identification Number or DOI: 10.1016/j.eswa.2009.10.016
URI: http://eprints.usq.edu.au/id/eprint/26520

Actions (login required)

View Item Archive Repository Staff Only