Brain Tumor Detection Using MRI Image Processing Techniques

Medical image processing is highly developed and critical in the medical field, encompassing methods like X-rays, MRI, and CT scans. These techniques help to identify minute defects in the human body. Brain tumors, which disrupt normal brain function, are detected through MRI using segmentation, feature extraction, and classification, processes that are time-consuming and depend on the expertise of clinical professionals. Computer-aided technology addresses these challenges. This work aims to diagnose brain tumors accurately and quickly using MRI images. Image pre-processing is done using CLAHE and median filter to enhance contrast and eliminate noise. Thresholding is used for segmentation and a lightweight U-net is used for tumor segmentation, achieving high accuracy and efficiency without extensive training data. Finally, an MSD-CNN classifies the tumor, capturing both local and global features for better image recognition. Our proposed model outperforms over existing methods in terms of accuracy, robustness, and efficiency in brain tumor detection.