A new method for identification and classification of human wrist bones in CT

A system designed for the purpose of segmenting and classifying the bones present in computed tomography (CT) images of the human wrist is described. The classification scheme presented is new in its treatment of the cost penalty associated with an incorrect classification. The analysis of an image is carried out in two stages: the segmentation stage and the classification stage. The segmentation stage uses the fundamental image structures, such as edges, to locate the areas of the image that are to be examined as possible candidates for the bones in the wrist scan. The classification stage considers knowledge of the structures of the human wrist to determine which of the candidates found in the segmentation stage correspond to the actual bones. The desired result is a structural description of the given scene, with the contents located and identified. Preliminary results are judged promising