An assessed digital mammography segmentation algorithm used for content-based image retrieval

In a previous work, we presented a comprehensive validation analysis to evaluate the performance of three existing digital mammography segmentation algorithms against manual segmentation results produced by two expert radiologists. In that study it was concluded that the region growing combined with maximum likelihood (RGCwML) model yielded not only the best accuracy, specificity, percent error and algorithm ranking, but also the greatest ratio of average computer to observer agreement and average inter-observer agreement (WI'). It was also noted that the upper limit of the 95% confidence interval (CI) was greater than 1.0 and thus each individual observer is a reliable member of the group. These studies are especially important for the development of computer-aided diagnosis (CAD) systems for cancer; equally important is the ability to retrieve "similar" images (mammograms) from a standing database. A framework for a new digital mammography content-based image retrieval system (DMCBIR) is discussed in this communication