Vol:.(1234567890) La radiologia medica (2022) 127:398–406 https://doi.org/10.1007/s11547-022-01476-7 1 3 COMPUTER APPLICATION Automatic detection and classifcation of knee osteoarthritis using deep learning approach S. Sheik Abdullah 1  · M. Pallikonda Rajasekaran 1 Received: 11 April 2021 / Accepted: 25 February 2022 / Published online: 9 March 2022 © Italian Society of Medical Radiology 2022 Abstract Purpose We developed a tool for locating and grading knee osteoarthritis (OA) from digital X-ray images and illustrate the possibility of deep learning techniques to predict knee OA as per the Kellgren-Lawrence (KL) grading system. The purpose of the project is to see how efectively an artifcial intelligence (AI)-based deep learning approach can locate and diagnose the severity of knee OA in digital X-ray images. Methods Selection criteria: Patients above 50 years old with OA symptoms (knee joint pain, stifness, crepitus, and func- tional limitations) were included in the study. Medical experts excluded patients with post-surgical evaluation, trauma, and infection from the study. We used 3172 Anterior–posterior view knee joint digital X-ray images. We have trained the Faster RCNN architecture to locate the knee joint space width (JSW) region in digital X-ray images and we incorporate ResNet-50 with transfer learning to extract the features. We have used another pre-trained network (AlexNet with transfer learning) for the classifcation of knee OA severity. We trained the region proposal network (RPN) using manual extract knee area as the ground truth image and the medical experts graded the knee joint digital X-ray images based on the Kellgren- Lawrence score. An X-ray image is an input for the fnal model, and the output is a Kellgren-Lawrence grading value. Results The proposed model identifed the minimal knee JSW area with a maximum accuracy of 98.516%, and the overall knee OA severity classifcation accuracy was 98.90%. Conclusions Today numerous diagnostic methods are available, but tools are not transparent and automated analysis of OA remains a problem. The performance of the proposed model increases while fne-tuning the network and it is higher than the existing works. We will extend this work to grade OA in MRI data in the future. Keywords Osteoarthritis · Digital X-ray images · JSW · Faster R-CNN · AlexNet Introduction Osteoarthritis (OA) is the one type of arthritis that involves knee deformity, resulting in joint pain and knee functional disability [1]. Epidemiology studies show that OA is an important disability factor around the world [2]. Besides medication and physiotherapy, total knee joint replacement surgery remains the only choice to extend the life of patients with healthy futures [3]. We consider narrowing of joint space and bony spurs to be the key features of osteoarthritis [4, 8]. Usually, medical experts examine knee joint OA on knee plain radiographs, and it is very useful to assess the development of osteophyte, knee joint space width reduc- tion, subchondral geodes, and subchondral bone sclerosis. Kellgren-Lawrence (KL) grading system is the benchmark for OA evaluation based on the above factors. X-ray images are still important for diagnosing health maintenance and experts use digital X-ray images for the OA assessment. Researchers have performed much research work to locate and measure the JSW, but there are signifcant diferences between their results. Machine learning (ML) based com- puter-aided approaches may solve the diagnostic problem by the automatic diagnosis of knee OA severity[5]. In addition to that, the system of manual diagnosis of OA is slow and repetitive. There is a need for computer. * M. Pallikonda Rajasekaran mpraja80@gmail.com S. Sheik Abdullah sabdullah787@gmail.com 1 Department of Electronics and Communication, Kalasalingam Academy of Research and Education, Srivilliputhur, India