Brain Tumor MRI Identification and Classification Using DWT, PCA, and KSVM

Abstract

Background Classification, segmentation, and the identification of the infection region in MRI images of brain tumors are labor-intensive and iterative processes. The optimum classification technique helps make the proper choice and delivers the best therapy. Despite several significant efforts and encouraging discoveries in this subject, precise segmentation and classification remain challenging tasks. Method In this study, we proposed a new method for the exact segmentation and classification of brain tumors from MR images. Initially, the tumor image is pre-processed and segmented by using the Threshold function for removing image noises. To minimize complexity and enhance performance used Discrete wavelet transformation (DWT) for getting the accurate in MR Images. Principal component analysis (PCA) are used to condense the feature vector dimensions of magnetic resonance images.Finally, for differentiate between benign and malignant tumor types, the Classification stage employs a pre-trained Support Vector Machine with several kernels, also known as a kernel support vector machine (KSVM). Result The efficacy of the suggested approach is also compared to that of other existing frameworks for segmentation and classification. Results demonstrated that developed approach is effective and quick, where as we obtained excellent accuracy and recognized the brain MR Images as normal and pathological tissues.