Standalone functional CAD system for multi-object case analysis in hepatic disorders

A new algorithm able to automatically diagnose the presence of the hemangioma areas in the hepatic ultrasonographic image is proposed. The algorithm uses a new multi-object approach which decomposes the image into three biological regions: a normal hepatic area, a hemangioma area and other areas. The de-noising process is efficiently accomplished for both Gaussian and Rayleigh noise distributions. Furthermore, a segmentation technique, based on gray level intensity analysis and the Moore-Neighbor contour tracing algorithm for a robust differentiation of the hemangioma area are employed. This new proposed technique is almost fully automatic, fast, and simple and its results are satisfactory.