Comparison between Sobel, Canny and Shen & Castan edge detection applied on bone joints / Mohammad Ashri Abu Hassan

Edge detection techniques are frequently used to find complex object boundaries by marking potential edge points corresponding to places in an image where rapid changes in brightness occur. Edge detection operators are based on the ideas that edge information in an image is found by looking at the relationship of neighborhood pixels. The edge is defined by a discontinuity in gray-level values. X-ray images generally show the location, size and structural changes such as cracks or fracture in bones. Many fine structural changes in bone x-ray images escape detection even by experienced radiologists due to the slight gray level changes in the images. This is due to the nature of x-ray images which are often fuzzy. Because of this radiologists or surgeons usually recommend the patients to go for MRI and CT scan for further investigation in suspicious cases. But MRI and CT scan are more expensive than x-ray images. However by implementing appropriate image enhancement techniques on the low cost digitized x-ray films, the cortical (bone wall) outline of the bone can be seen clearer. In this research three edge detection methods, namely Sobel, Canny and Shen & Castan were implemented on bone joints to enhance the edges of the bone cortical outline of bone joint images. A prototype system was developed based on the three methods. Results obtained are compared by using Single Line Histogram Method (SLHM) and comparison of the Operator Performance to determine the best method most suitable to enhance edges of the joints in the x-ray images.