Knee Osteoarthritis Detection Using an Improved CenterNet With Pixel-Wise Voting Scheme

The primary objective of the project is to identify and diagnose knee osteoarthritis using knee images obtained through X-ray scans. X-ray images are a common and cost-effective modality for examining knee joint health, and the project focuses on leveraging these images for accurate detection of osteoarthritis. Current methods for knee osteoarthritis detection, which rely on image processing techniques, face challenges in terms of accuracy and precision. This technique involves the development of a specialized Center Net,A state-of-the-art object detection architecture. This Center Net is tailored with a pixelwise voting scheme, a, method that allows for the extraction of the features at a granular level. By customizing the Center Net in this way, the project aim to enhance the accuracy and reliability of knee osteoarthritis detection the model incorporates Dense Net 201 as a basic network for future extraction. By utilizing DenseNet201, the model aims to capture the most representative features for knee samples,enhancing the robustness of the feature extraction process.