Detection of Osteoarthritis in Knee Radiographic Images using Artificial Neural Network

Abstract

In this paper, the Osteoarthritis (OA) analysis in knee radiographic images using artificial neural networks (ANN) is considered. In Osteoarthritis, mobility is restricted and bones rub each other causing extreme pain in knee due to cartilage disintegration. The cartilage destruction is minimal in the initial stage of OA. It is observed that a small number of researchers have implemented identification and grading of Osteoarthritis utilizing their own datasets for experimentation. However, there is still need of automatic computer aided techniques to detect Osteoarthritis for early recognition. In this work, a dataset of 1650 radiographic images of knee joints of OA patients are collected from different hospitals and have been annotated by two different orthopedic surgeons as per the Kellgren and Lawrence (KL) grading system. To automate this grading procedure, the local phase quantization and multi-block projection profile features are computed from the images and then presented to artificial neural network to classify the images based on the KL grading of the severity of the disease. The classification accuracy of 98.7% and 98.2% with reference to surgeon-1 and surgeon-2 opinions, respectively, is achieved.

MedKnee: A New Deep Learning-Based Software for Automated Prediction of Radiographic Knee Osteoarthritis

“…The model achieved an accuracy of 95.20% on Medical Expert-I and 94.94% on Medical Expert-II. This minimal difference confirms the results obtained in [34]. To validate the results, the model was tested on 30 images and compared with the diagnosis of a rheumatologist.…”

Automatic Detection and Classification of Knee Osteoarthritis Using Hu's Invariant Moments

“…In order to provide improved performance for the denser regions, local phase quantization and projection profile features are computed and classified. The gradation was done using Artificial Neural Network and an accuracy of 98.7 and 98.2% as per the Expert I and II opinions, respectively, has been achieved (Gornale et al, 2019c). Lastly, the experiments utilizing multi-resolution wavelet filters with varying filter orders and decomposition levels have been performed and classification is done using decision tree classifier.…”