Osteoarthritis Detection and Classification from Knee X-Ray Images Based on Artificial Neural Network

Abstract

Osteoarthritis is a chronic and degenerative disorder of multifactorial etiology. Knee, hip and spine joints are majorly affected by the Osteoarthritis due to loss of articular cartilage, bone remodeling (due to accidents, for instance) and heavy weight-bearing on joints. In orthopedics, it is one of the most occurring disorder nowadays. Hence an effective computer-aided diagnosis (CAD) system is required to diagnose the Osteoarthritis. This paper presents a simple artificial neural network (ANN) based classification system to differentiate between healthy and affected knee X-ray images. In the proposed system, guided filter and adaptive histogram equalization techniques are respectively used for noise removal and image enhancement. Global thresholding-based segmentation technique is adapted to extract synovial cavity region from an image and curvature values (like mean, standard deviation, range, and skewness) are computed. To draw a fair conclusion, the experiments are conducted on real patient-specific images collected from local hospitals in India. By confirming the results, the proposed method accurately classifies the inputted image to their respective classes.