AI/ML Enabled Computer Aided Diagnosis System of Osteoarthritis by Using Medical Image Processing Techniques

Abstract

The most frequent cause of disability in older and affected younger generations is osteoarthritis. Younger people as well as older adults may develop OA due to a variety of factors, including poor diet, obesity, improper exercise, genetic disorders, a lack of calcium, and others. Among all the major issues, the weakening of bones is primarily due to the new generation's poor eating and sleeping habits. Modern methods like biochemical testing and other spectroscopic identification are more expensive, detrimental to healthy epidermal cells, and have an adverse impact on neurons. The main objective to get around this, all of the software that we are developing uses computer assisted diagnosis to find the cause using the image processing technique. Through the use of the database  input images (X-ray or MRI) has gathered for the proposed method, after which we have standardize the images while using image resizing and other techniques, and finally process the images further using pre-processing methods like image filtering and image enhancement techniques. Additionally, the CNN segmentation techniques has been  used on training and test data sets. Following the identification of the desired area in the input image, pertinent features  has been be extracted, and machine Leaning CNN classification techniques has been  used to determine whether the area is normal or abnormal. AC segmentation from MR images is a difficult task that has received a lot of attention. Highest Accuracy 75.64 % , Precision 71.87% and Recall 77.27 % has been obtain  using proposed CNN ML model in CAD tool system. Experimental Result has been obtain in  3xGTX1080Ti NVIDIA cards nvidia-docker to  install the software using Jupiter Notebook in Anaconda with   the Dockerfile provided.