Rigid registration algorithm based on the minimization of the total variation of the difference map

Image registration is broadly used in various scenarios in which similar scenes in different images are to be aligned. However, image registration becomes challenging when the contrasts and backgrounds in the images are vastly different. This work proposes using the total variation of the difference map between two images (TVDM) as a dissimilarity metric in rigid registration. A method based on TVDM minimization is implemented for image rigid registration. The method is tested with both synthesized and real experimental data that have various noise and background conditions. The performance of the proposed method is compared with the results of other rigid registration methods. It is demonstrated that the proposed method is highly accurate and robust and outperforms other methods in all of the tests. The new algorithm provides a robust option for image registrations that are critical to many nanoscale X-ray imaging and microscopy applications.

Automated summary

This study proposes a new method for image rigid registration, using the total variation of the difference map as a dissimilarity metric. The method demonstrates high accuracy and robustness in different noise and background conditions, outperforming other methods in all tests. This is significant for image registrations in nanoscale X-ray imaging and microscopy applications.