期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2016
卷号：5
期号：3
页码：4483
DOI：10.15680/IJIRSET.2016.0503256
出版社：S&S Publications
摘要：Classification of medical images is an important issue in computer-assisted diagnosis. In this paper, aclassification scheme based on a kernel principle component analysis (KPCA) model ensemble has been proposed forthe classification of medical images.Common CT imaging signs of lung diseases (CISLs)are defined as the imagingsigns that frequently appear in lung CT images from patients and play important roles in the diagnosis of lung diseases.This paper proposes a new feature selection method based on Fisher criterion and genetic optimization, called FIG forshort, to tackle the CISL recognition problem. The ensemble consists of KPCA models trained using different imagefeatures from each image class, and a proposed product combining rule was used for combining the KPCA models toproduce classification confidence scores for assigning an image to each class. The effectiveness of the proposedclassification scheme was verified using nine categories of CISL images from CT scan. The combination of differentimage features exploits the complementary strengths of these different feature extractors. The proposed classificationscheme obtained promising results on the medical image sets.
