Breast Cancer Diagnostic System Based on MR images Using KPCA-Wavelet Transform and Support Vector Machine

Mustafa Zuhaer AL-Dabagh, Dr. Firas H. AL-Mukhtar

k-mean clustering, Morphological operators, gray scale, Texture, symmetrical, kernel principle components analysis and Support vectors machine.

Automated detection and accurate classification of breast tumors using magnetic resonance image (MRI) are very important for medical analysis and diagnostic fields. Over the last ten years, numbers of methods have been proposed, but only few methods succeed in this field. This paper presents the design and the implementation of CAD system that has the ability to detect and classify the tumor of the breast in the MR images. To achieve this, k-mean clustering methods and morphological operators are applied to segment the tumor. The gray scale, Texture and symmetrical features as well as discrete wavelet transform (DWT) are used in feature extracted stage to obtain the features from MR images. Kernel principle components analysis (K-PCA) are also applied as a feature reduction technique and support vectors machine (SVM) are used as a classifier. Finally, the experiments results have confirmed the robustness and accuracy of proposed system

10.22161/ijaers.4.3.41