Deep Convolutional Network Tumor Classification System in Brain MRI

Doctors' availability is one of the concerns of people who want to have an appointment with a specialist. Doctors do have a very tight schedule and cannot consistently have appointments, especially for doctors who specialize in brain tumors as they are not that much in numbers here in the Philippines. Brain tumors are a very rare case among individuals and the 10-year survival rate is about 31%, a very low survival rate compared to other diseases. YOLOv3 or a machine learning model using transfer learning and object detection is used as the approach to classify brain tumors. The detection model has an mAP value of 99%, according to the study's findings. The proposed system applies to hospitals, especially those that conduct brain MRIs to classify what type of brain tumor are we dealing with right after the test. The system has a 51% to 99% precision rate per frame