Osteoarthritis Classification Using Knee X-Ray Images Based on Hybrid Feature Fusion Framework

Osteoarthritis is the most common disease in the worldwide population targeting the knee, spine, neck, hand, hip, and almost all the joints of the human body. It is mostly affected by 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 disorders nowadays. In this paper the Knee Osteoarthritis classification using X-ray images. Initially, the input X-ray images are preprocessing. Secondly, features are extracted from these preprocessed images using techniques such as Grey-level co-occurrence matrix (GLCM), Gray-level run length matrix (GLRLM), and Histogram of oriented gradient (HOG). These extracted features are combining with feature fusion vector represented as GLCM + GLRLM + HOG. Lastly, multi-class classifiers are used Bayes Net, Naïve Bayes, Logistic, Random Forest, and Random Tree for the classification by using individual feature vectors as well as feature fusion vectors. The dataset used in this study from Kaggle and Mendeley websites. These are the two datasets available online, Dataset-I and Dataset-II contains the Knee X-ray images of Osteoarthritis. The Knee X-ray images consist of Grade 0, Grade 1, Grade 2, Grade 3, and Grade 4. This paper proposes a hybrid framework that combines feature extraction and classification to improve accuracy. The GLCM, GLRLM, and HOG are used for feature extraction techniques. Dataset-I gives good results by using the Naïve Bayes is 76.23%.