Bone edge detection in hand radiographic images

A method for bone segmentation in hand radiographic images using an intensity based edge detection scheme is presented. Bone segmentation is an intermediate level processing stage for an automated vision system for skeletal assessment of children. An inherent problem with hand radiographic images is the uneven contrast that is manifested by an increasing intensity gradient from the distal phalanges to the carpal region. The intensity based edge detection approach exploits knowledge of boundary region intensities to remove extraneous edges and maintain bone edges. This approach has been tested on hand X-ray images from various age groups and results have shown a marked improvement in bone segmentation.< >