Brain Tumor and Types Detection Using MedicalImaging

Abstract

Brain tumors pose a major threat, so finding them early is key to better treatment and survival. Current ways to spot them involve radiologists checking MRI and CT scans by hand. This takes time, can be spotty, and might have mistakes. To fix these problems, this study shows off a system that uses deep learning to automatically find and sort brain tumors by checking medical images. This system puts image prep, cutting out parts, and sorting into one deep learning setup. First, MRI images are cleaned up to get rid of noise and look better. Then, a U-Net design is used to carefully cut out the tumor spots. After that, a CNN sorts the tumors into types like glioma, meningioma, and pituitary. The model is trained and checked using standard MRI datasets like BRATS. This makes sure it works well and is correct in both cutting and sorting jobs. Tests show this new method is over 97% correct. It cuts down on how long it takes to figure things out and lowers the amount of work needed by hand. The automatic setup makes diagnoses more trustworthy, shows where the tumor is, and helps radiologists make quicker, more right calls. By mixing AI with medical images, this work points out how deep learning can change healthcare diagnoses. This leads to early finds, smooth work, and better results for patients.