Computer-aided detection and classification of brain tumor using YOLOv3 and deep learning

The human brain has a complicated structure, making it difficult to diagnose any brain ailments, particularly when they affect important parts of the brain. A brain tumor with a high grade type instance is extremely hazardous and lethal. The traditional method for locating tumors in brain MRI (magnetic resonance imaging) pictures relies on human skill. An effective and trustworthy model for cancer tumor detection and classification will revolutionize the medical industry because people are fallible. In this study, a paradigm for identifying and categorizing tumors visible in brain MRI scans is put forth. The object detection algorithm YOLO-v3 is employed for detection. The third iteration of YOLO, YOLOv3 (you look only once), detects the target topic in a single pass and provides a confidence score. Using a convolutional neural network (CNN), the identified tumor is then divided into two classifications based on the World Health Organization's (WHO) grading system, low-grade and high-grade glioma. The performance of the proposed method is evaluated and the classification accuracies and results are compared with other state-of-the-art models. The proposed method using CNN and YOLOv3 has more potential to classify the brain tumor by achieving an overall accuracy of 97%.