Hybrid Data Mining Approach for Image Segmentation Based Classification

Citation:
Panda, M., A. E. Hassanien, and A. Abraham, "Hybrid Data Mining Approach for Image Segmentation Based Classification", International Journal of Rough Sets and Data Analysis (IJRSDA), vol. 3, issue 2, 2016.

Abstract:

Evolutionary harmony search algorithm is used for its capability in finding solution space both locally and globally. In contrast, Wavelet based feature selection, for its ability to provide localized frequency information about a function of a signal, makes it a promising one for efficient classification. Research in this direction states that wavelet based neural network may be trapped to fall in a local minima whereas fuzzy harmony search based algorithm effectively addresses that problem and able to get a near optimal solution. In this, a hybrid wavelet based radial basis function (RBF) neural network (WRBF) and feature subset harmony search based fuzzy discernibility classifier (HSFD) approaches are proposed as a data mining technique for image segmentation based classification. In this paper, the authors use Lena RGB image; Magnetic resonance image (MR) and Computed Tomography (CT) Image for analysis. It is observed from the obtained simulation results that Wavelet based RBF neural network outperforms the harmony search based fuzzy discernibility classifiers.

Related External Link