X-Ray Mammogram Registration: A Novel Validation Method

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 and facilitate its quantification. The literature contains numerous non-rigid registration algorithms developed for this purpose, but quantitative estimation of registration accuracy is limited. We describe a novel validation method which simulates plausible mammographic compressions of the breast using an MRI derived finite element model. Known 3D displacements are projected into 2D and test images simulated from these same compressed MR volumes. In this way 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 mammograms simulated from three patient MR datasets.