Advancing Brain Tumor Diagnosis: A Machine Learning Approach for Accurate Detection and Classification

In this paper, a novel brain tumor detection algorithm is proposed using YOLOv5 and YOLOv8 models. Due to their remarkable speed-accuracy balance, these models are perfect for clinical applications requiring low latency and high dependability. Before training and optimizing the models by using the YOLO approach for tumour identification and localization, the framework first preprocesses MRI images to improve quality and standardise dimensions. On a mixed dataset, which comprises pituitary tumours, gliomas, meningiomas, and patients without tumors, YOLOv8 and YOLOv5 exhibited 98.9% and 12 ms per picture accuracy. These results demonstrate how well models based on YOLO work toward real-time, highly accurate tumour identification. By implementing the framework within a pipeline for diagnosis, it may alter early tumour diagnosis, and clinical practices can provide decisions much faster. To ensure more application in actual healthcare settings, future work will concentrate on growing the dataset, including multi-modal imaging, and optimizing models for edge deployment.