Brain Tumor Detection and Segmentation using CNN and Optimised Edge Features

Accurate identification of tumor using MRI scans of brain demands robust preprocessing techniques that effectively capture the intricate details of the skull area, especially in the sagittal and coronal planes. This study proposes a novel preprocessing approach tailored to enrich the visibility of the skull area in MRI images, thereby aiding in precise recognition of brain tumor and segmenting the tumor area. Leveraging Convolutional Neural Network (CNN) for classification and segmentation via thresholding, our method addresses the challenge posed by the broader skull area present in the saggital and coronal plane brain MRI images. By incorporating advanced preprocessing approaches, our suggested approach enhances the precision of identifying tumor and also minimize computational complexity. Experimental evaluations demonstrate the efficacy of the recommended method in accurately identifying brain tumors, underscoring its potential to assist clinicians in early diagnosis of tumor region and treatment planning.