Automatic joint detection and measurement of joint space width in arthritis

Arthritis is an inflammatory disease which causes erosion in bones or narrowing of joint space in various joints of the body. First symptom of this disease is seen in joints of hand finger and wrist joints thus making hand radiograph analysis extremely important. Lately Reading hand X-ray radiographic image to measure joint space width is very tedious and time consuming task for the radiologist since there are 14 joints in hand and also the structure of hand is complicated to carry out joint space width measurement and analysis. It has certain disadvantages like inaccuracy because of visual measurement and also variation from one reader to another, which can be overcome by automatic technique that can serve as a powerful aid for peoples suffering from disability due to pain, stiffness in joints. In this paper, Image processing based algorithm is developed to yield solution to two major problems joint detection and JSW measurement. The algorithm is divided into following steps, First image preprocessing is carried out using Gaussian filter. Second hand mask is extracted by separating foreground and background by using Otsu's binarization method. Third morphological thinning is applied to get thinned skeleton of binarized image. Fourth To detect joint location in original X-ray image Gabor filter is used. Fifth edge Finally of minimal joint space width is extracted and analyzed automatically. We have experimented 10 digital hand X-ray radiograph of resolution 2000pixels×2000pixels and calculated 120 readings of JSW of finger joints successfully.