Design and implementation of a computer-aided diagnosis system for brain tumor classification

In this study, a two-stage CAD system was developed for the automatic detection and classification of brain tumor via magnetic resonance imaging (MRI). The systems both increase the diagnostic accuracy and reduce the time needed. This system first classifies brain tumor MRI as normal and abnormal images. In the second stage, the tumor type from abnormal MRIs is classified as benign (non-cancerous) or malignant (cancerous). With the proposed CAD; Feature extraction using K-means clustering, MRI image segmentation, discrete wavelet transform (DWT) and feature reduction by applying principal component analysis (PCA). In the second stage classification, a support vector machine (SVM) was used. Performance evaluation of the proposed CAD results were obtained using a non-standard MRI database. Keywords: Brain Tumor, DWT, PCA, SVM, Tumor Classification