Abstract
Abstract : The goal of this project is to develop a computer-aided diagnosis (CAD) system for interval change analysis of lesions on mammograms. An important component of the CAD system is the multistage regional registration technique for identifying corresponding microcalcification clusters on temporal pairs of mammograms. In the first stage, an initial search region was estimated on the prior mammogram based on the lesion location on the current mammogram. In the second stage the search region was refined. In the third stage the lesion was detected within the search region. In the first stage we compared the regional registration method (RRM) to the use of linear and nonlinear warping techniques for the initial estimation of the lesion location. 390 temporal pairs of mammograms were used for evaluation. The average distance between the estimated and the true lesion centroids on the previous mammogram after the initial stage was 8.5%+/- 6.2mm.2mm for RRM and 9.0 +/- 6.7mm for the best of the warping techniques. The RRM method outperformed the warping techniques. In the second step, automated detection of microcalcification cluster within the search region is performed. Using our current cluster detection program with standard thresholds, 69.4% (50/72)TP with 0.21 false positives (FP) were detected within the search region. Using a high-sensitivity threshold, 84.7% (61/72) TP with 0.75 FP were detected.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call