Brain Tumor Detection using YOLO v7 and YOLO v8

The architecture of the brain is incredibly complex, with multiple areas managing various parts of the nervous system. An adult’s health is greatly at risk from brain tumors because of their ability to severely impair organ function or possibly cause death due to their rapid growth of abnormal brain cells. Brain tumors can develop in any area of the skull or brain, including the protective layer around the brain and many other areas. Since these tumors can have a wide range of shapes, sizes, and locations, magnetic resonance imaging (MRI) is an essential tool for locating malignant tumors. MRIs are examined manually in order to diagnose brain tumors. However, manually identifying brain tumors is a challenging and time-consuming task that may result in errors. As a result, computer-aided brain tumor detection has been considered a crucial area of study in biomedical engineering. However, because tumor sizes and shapes vary enormously throughout subjects, developing a reliable brain tumor detector remains difficult. Thus, for the purpose of accurately identifying brain tumors in MRI scans, two popular deep learning detection algorithms—YOLO v7 and YOLO v8—were tested and compared in this study. From the results, YOLO v8 model has achieved highest validation accuracy of 96.50% with average precision and recall value of 95.9% and 93.3% respectively, while YOLO v7 has validation accuracy of 87.6% with average precision and recall value of 81.9% and 85.4% respectively.