Likelihood function analysis for segmentation of mammographic masses for various margin groups

The purpose of this work was to develop an automatic boundary detection method for mammographic masses and to observe the method´s performance on different four of the five margin groups as defined by the ACR, namely, speculated, ill-defined, circumscribed, and obscured. The segmentation method utilized a maximum likelihood steep change analysis technique that is capable of delineating ill-defined borders of the masses. Previous investigators have shown that the maximum likelihood function can be utilized to determine the border of the mass body. The method was tested on 122 digitized mammograms selected from the University of South Florida´s Digital Database for Screening Mammography (DDSM). The segmentation results were validated using overlap and accuracy statistics, where the gold standards were manual traces provided by two expert radiologists. We have concluded that the intensity threshold that produces the best contour corresponds to a particular steep change location within the likelihood function.