DETECTION AND CLASSIFICATION OF BONE TUMOUR FROM MRI-IMAGES USING DEEP LEARNING

Abstract

Bone tumour detection, a critical aspect of medical imaging, plays an essential role in diagnosing various bone-related diseases. Early and accurate identification of bone tumours can significantly improve treatment outcomes and patient survival rates. Traditionally, the detection of bone tumours from MRI images has been challenging due to the complexity of tumour shapes, variations in size, and the resolution of the images. Over the years, deep learning techniques have revolutionized the field of medical image analysis, providing highly accurate and automated solutions. In this project, an artificial intelligence (AI) model, using the YOLOv8 architecture, was developed to detect and classify bone tumours from MRI images. The model was trained on a dataset sourced from Kaggle, consisting of 1146 training images, 157 validation images, and 79 test images. The YOLOv8 model achieved an accuracy of 96%, demonstrating its high potential for real-time detection. Additionally, a web-based platform was created, allowing users to upload MRI images for tumour detection and receive instant results, with the added feature of user feedback for further system improvement. This system offers a practical and efficient solution for bone tumour detection, contributing to the field of medical diagnostics by enhancing early detection and diagnosis.