Weaver Ant Colony Optimization-Based Neural Network Learning for Mammogram Classification

Weaver Ant Colony Optimization-Based Neural Network Learning for Mammogram Classification

A. Kaja Mohideen (Department of Applied Mathematics and Computational Sciences, PSG College of Technology, Coimbatore, Tamil Nadu, India) and K. Thangavel (Department of Computer Science, Periyar University, Salem, Tamil Nadu, India)
Copyright: © 2013 |Pages: 20
DOI: 10.4018/ijsir.2013070102
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Neural Networks (NNs) have been efficaciously used for classification purposes in medical domains, including the classification of microcalcifications in digital mammograms. Unfortunately, for a NN to be effective in a particular purview, its architecture, training algorithm and the domain variables selected as inputs must be amply chosen. In this paper, a novel Ant Colony Optimization (ACO) based learning approach with a modified architecture is proposed to speed up the learning phase of a Backpropagation Neural Network (BPN) classifier. The novel ACO simulates the behavior of weaver ants, known for their unique nest building behavior where workers construct nests by weaving together leaves using larval silk. The proposed Weaver Ant Colony Optimization (WACO) based Backpropagation Neural Network (WACO-BPN) is applied for classifying digital mammograms received from MIAS database. The performance is analyzed with Receiver Operating Characteristics (ROC) curve. The greater accuracy of 97% states the grander performance of the proposed neural network learning approach.
Article Preview

1. Introduction

Breast cancer is the most common diagnosed cancer among women. Mammography is helpful to detect breast cancer at the earliest stage to reduce the rate of mortality amongst women. However, detecting the early signs in the X-ray mammograms is significantly a challenging task, because such signs are hard to distinguish against the highly textured or dense. A detailed study about the various methods used for automatic detection and classification of microcalcification is given in (Cheng et al., 2003, 2006; Thangavel et al., 2005). Although mammogram encompasses useful evidence for the early detection of breast cancer, it is difficult for radiologists to make precise and steady judgments due to the huge amount of data and extensive screening. Consequently, about 10–30% cases are missed during the routine check (Cheng et al., 2003). With the assistance of Computer-Aided Diagnosis (CAD), the overall sensitivity from human observers can be improved by 10% on average, which affords a promising solution in such a context. The strain for the detection of Microcalcification Clusters (MCCs) is due to (i) small size but various shapes, (ii) low contrast and unclear boundary from surrounding normal tissue, etc. (Cheng et al., 2003; Sentelle et al., 2002). To solve such problems, a typical CAD system contains at least four stages including preprocessing, segmentation of MCCs, feature-based extraction of Regions of Interest (ROI), and classification.

The focus of the paper is on classification, number of techniques has been proposed in the literature using machine learning approaches to classify the microcalcifications as malign and benign. Neural networks have been widely used in the research of medical image classification, Jiang et al., (2010) provided a focused literature survey on recent neural network developments in CAD. Classification of microcalcification clusters from mammograms play essential roles in computer-aided diagnosis for early detection of breast cancer. Ayer et al., (2012) provided an overview of most commonly used neural network models, in CAD and CADx for mammography interpretation and biopsy decision making.

The two most commonly used techniques are Artificial Neural Network (ANN) and Support Vector Machine (SVM). Based on learning samples, the ANN outperform SVM when balanced learning is absent, whereas in case of imbalanced learning samples, the SVM performs better than ANN (Ren 2012) but they require prior configuration. The configuration has two phases: one is to identify the suitable structure for the network, and numerical weights for the neuron connections must be updated efficiently such that the classifier is as correct as possible. In this paper, we have designed a novel learning as well as novel architecture for improving the ANN performance.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 9: 4 Issues (2018): 1 Released, 3 Forthcoming
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing