Leveraging Yolov3 and CNN Algorithms for Brain Tumor Detection in MRI Scans

• Published in: 2024 IEEE 4th International Conference on ICT in Business Industry & Government (ICTBIG) pp. 1 - 6
• Main Authors: Dhabliya, Dharmesh, Prabagaran, S., Sood, Gourav, K, Divya, Arora, Mithhil, V, Vivek
• Format: Conference Proceeding
• Language: English
• Published: IEEE 13.12.2024
• Subjects: Accuracy; Brain Computer Interfaces; Brain modeling; CNN; Computational modeling; Convolutional neural networks; Electrocardiography Probabilistic Neural Network; Hands; Image segmentation; Magnetic resonance imaging; MRI; Systematics; Tumors; YOLO
• DOI: 10.1109/ICTBIG64922.2024.10911683

The brain ranks high among the body's most vital organs, and scientists have examined it to learn more about human behavior. The body acts better when the brain is operating properly. Several things affect the brain's systematic function. We opted to research brain tumours, which are caused by aberrant brain cell growth. MRI is usually used for brain recognition. The inability to determine location is a major drawback. Finding the tools and techniques to identify, diagnose, and categorise the disease from the image is crucial. The proposed method is an MRI-based approach to identify and classify brain tumours. We utilise the YOLOv3 object detection technique. This approach may determine the target subject and confidence score in one pass. Additionally, CNN divides tumors into two groups: gliomas of low severity and high severity gliomas, according to the grading system used by the World Health Organisation. The model can reliably detect MRI tumours with 99.5%, 97%, and 98% accuracy, respectively. We achieved 98% categorisation accuracy by evaluating forty photographs, twenty each category. With Low Grade Glioma, one may acquire 100% precision at recall 97%, while with High Grade, 97% precision at recall 100%.