 This research paper proposes three methods for X-ray analysis and diagnosis of knee osteoarthritis, KOA. The first method utilizes two convolutional neural networks, CNNs, VGG19 and ResNet 101, to analyze the X-ray images and determine the severity of KOA. The second method combines the features of these two CNNs with principal component analysis, PCA, to further reduce the number of features and improve the accuracy of the diagnosis. Finally, the third method fuses the features of VGG19 and ResNet 101 with handcrafted features to achieve even higher accuracy. All three methods were tested on two different datasets and achieved high accuracy rates. This article was authored by Ahmed Khalid, Ebrahim Mohamed Senan, Khalil Al-Waji, and others.