부산대학교 도서관

Establishing spatial correspondence between features visible in x-ray mammograms obtained at different times has great potential to aid assessment of change in the breast indicative of malignant changes and facilitate their quantification. The literature contains numerous non-rigid registration algorithms developed for this purpose, but quantitative estimation of registration accuracy is limited. In this paper we describe a novel validation method which simulates plausible mammographic compressions of the breast using an MRI derived finite element model. By projecting the resulting known 3D displacements into 2D and simulating x-ray mammograms from these same compressed MR volumes, we can generate convincing images with known 2D displacements with which to validate a registration algorithm. We illustrate this approach by computing the accuracy for a non-rigid registration algorithm applied to mammographic test images generated from three patient MR datasets.