A knowledge based segmentation method applied to medicalradiographs

摘要

The problem of identifying an individual's skeletal age from aradiographic image of the hand and wrist is of importance toradiologists. From a radiograph a person's growth and development can beassessed and examined for abnormal conditions. At present, a person'sage is estimated either by visual comparison of radiographs using anatlas compiled by Greulich and Pyle (1959) or by combining visual andtextual descriptions as detailed by Tanner and Whitehouse (1975). Bothmethods are time consuming and results are often subjective, see Cole(1988). Computer analysis of skeletal age would overcome the aboverestrictions. Three separate levels of processing may be identified toachieve this goal; bone boundary detection, bone identification andclassification. The subjects of the paper are (a) detection of boneboundaries using closed contour edge detectors to segment the image and(b) clustering to form regions representing the bone detail.Segmentation methods which rely on first derivative edge detection haveshown to be ineffective due to the occurrence of gaps in the resultingedge maps. One approach is to prevent such gaps from occurring byimproving the method of detecting edges. Three solutions to the problemof closed contour edge detection are proposed: an anisotropic Laplacianof a Gaussian operator; a space limited Laplacian of a cosine operator;and a gradient tracking scheme