Osteoarthritis (OA) is most typically a result of cartilageiaging. Particularly in rural India, ithe percentage of people affected due to OA is far larger. It could be due to obesity, injuries or hereditary problems. This paper focuses on various aspects of severity analysis done using segmented cartilage. The automatic detection of OA severity supported by KL grades corresponding to several stages has been  proposed by researchers and provides better results for analysis of the disease. Automatic segmentation approach using U-net, support vector machine-n classier is being implemented in this paper. The paper mainly focuses on all important issues related to segmentation in OA by testing on 100 images and training 355 images.