abstract

In this study, we tackled the complex task of detecting brain tumors in MRI scans using a large dataset of brain tumor images. We found that fine-tuning a cutting-edge YOLOv7 model through transfer learning significantly enhanced its ability to detect gliomas, meningiomas, and pituitary tumors. Our deep learning model demonstrated promising results by accurately pinpointing the presence and exact location of brain tumors in MRI scans. Our approach achieved superior accuracy compared to conventional methods, achieving an impressive 99.5% accuracy in our evaluations. However, we recognize the need for further investigation and validation to ensure the efficacy of our method in detecting small tumors. The challenge of identifying small tumors underscores the ongoing necessity for research in brain tumor detection and continual improvement of our detection systems. By pursuing this path, we strive to advance diagnostic capabilities for both patients and medical professionals in the demanding fight against brain cancers.