lcreg - Efficient registration of large 3D images
Rigid and affine registration of large scalar 3D images is an import step for both medical and non-medical image processing. The distinguishing feature of lcreg is its capability to efficiently register images that do not fit into system memory. lcreg is based on the optimisation of the local correlation similarity measure  using a novel image encoding scheme fostering on-the-fly image compression and decompression .
Tutorial and samples
Contact and support
Many thanks to Karl-Heinz Kunzelmann for his support, many helpful discussions and for making dental test images available. This work benefited from the use of ITK-SNAP, bcolz, numpy scipy and cython. The University of Applied Sciences, Augsburg, in particular the Faculty of Computer Science supported this project by granting sabbatical leaves. Special thanks to Gisela Dachs, Andreas Gärtner, Evi Köbele, Stefan König, Dominik Lüder, Thomas Obermeier and Sigrid Podratzky for acquiring test images and for keeping computers up and running.
 T. Netsch, P. Rösch, A. v. Muiswinkel and J. Weese:
Towards Real-Time Multi-Modality 3-D Medical Image Registration. Eight IEEE International Conference on Computer Vision, ICCV (2001) 718-725,
 P. Rösch and K.-H. Kunzelmann: Efficient 3D rigid Registration of Large Micro CT Images. International Journal of Computer assisted Radiology and Surgery 13 (Suppl. 1) (2018) 118–119,