Detection and Segmentation of Tumors in Brain MR Images

Abstract

Brain tumors arise from various complex factors, including genetic, environmental, immunological, and biochemical influences. They can be classified as primary or metastatic, differing in their origin and location. Brain tumors significantly impact the quality of life, leading to symptoms such as headaches, seizures, cognitive decline, and motor function impairment, depending on the tumor's size and location. Early diagnosis of brain tumors is crucial for improving quality of life. Timely detection allows for prompt treatment initiation, which can prevent tumor growth and the worsening of symptoms. Diagnosis typically involves neurological examinations, imaging examinations, tissue biopsies, and blood tests. In particular, MRI provides high-resolution images of the brain's detailed structure, clearly depicting the location, size, shape, and surrounding tissues of the tumor. This study proposes a method for detecting and segmenting brain tumors in MRI images, utilizing a dataset constructed for this purpose, named "BrainTumors_1.0.zip." Experimental results demonstrate that filtering the input images enhances image quality and enables accurate tumor detection. Future research will focus on enhancing algorithm generalization, diversifying the dataset, developing automated methodologies, and assessing clinical utility to establish an effective tool for the diagnosis and treatment of brain tumors.