Texture-based characterization of microcalcification clusters
provides a robust tool for the development of a computeraided
diagnosis (CADx) in mammography. Unlike its
counterpart, a shape-based approach, texture approach does
not require a microcalcification segmentation stage. This
paper presents a new texture-based CADx that uses a PSOSVM
embedded feature selection that is also an improvement
of previous PSO-KNN approach [11]. The proposed system
mainly consists of texture feature extraction and heuristic
parameter selection stages. The first stage characterizes
microcalcification (MC) clusters using 28 GLCM features.
The second stage of the system involves feature selection and
performance optimization of a kernel-based SVM classifier
using a PSO-SVM approach. This step is to heuristically
search for the most discriminative texture features and to find
the optimal SVM learning model that comprises of the
regularization and kernel parameters. Testing the proposed
parameter selection approach using MC clusters from mini-
MIAS datasets produced perfect classification accuracy and
demonstrated a promising performance of parameter selection
using PSO-SVM method.
Track
ISKE: Intelligent Systems and Knowledge Engineering
Conference
5th Mosharaka International Conference on Communications, Computers and Applications (MIC-CCA 2012)
Congress
2012 Global Congress on Communications, Computers and Applications (GC-CCA 2012), 12-14 October 2012, Istanbul, Turkey
Pages
7-12
Topics
Optimization of Computing Algorithms Knowledge-Based Systems
ISSN
2227-331X
DOI
BibTeX
@inproceedings{266CCA2012,
title={ An Improvement of Texture-Based Classification of Microcalcification Clusters in Mammography using PSO-SVM Approach },
author={Imad Zyout, and Ikhlas Abdel-Qader},
booktitle={2012 Global Congress on Communications, Computers and Applications (GC-CCA 2012)},
year={2012},
pages={7-12},
doi={}},
organization={Mosharaka for Research and Studies}
}
