Computer Methods and Programs in Biomedicine (2005) 78, 101—114 Registration of MR and CT images of the liver: comparison of voxel similarity and surface based registration algorithms Wen-Chi Christina Lee a , Mitchell E. Tublin c , Brian E. Chapman b, * a Department of Bioengineering, University of Pittsburgh, 749 Benedum Hall, Pittsburgh, PA 15261, USA b Departments of Radiology and Bioengineering and Center for Biomedical Informatics, University of Pittsburgh, Imaging Research, 300 Halket Street, Pittsburgh, PA 15213-3180, USA c Department of Radiology, University of Pittsburgh, 200 Lothrop Street, Pittsburgh, PA 15213, USA Received 4 April 2004; received in revised form 10 December 2004; accepted 30 December 2004 KEYWORDS Magnetic resonance imaging; Computed tomography; Liver; Image registration Summary The purpose of this work was to determine the feasibility and efficacy of retrospective registration of MR and CT images of the liver. The open-source ITK In- sight Software package developed by the National Library of Medicine (USA) contains a multi-resolution, voxel-similarity-based registration algorithm which we selected as our baseline registration method. For comparison we implemented a multi-scale surface fitting technique based on the head-and-hat algorithm. Registration accu- racywasassessedusingthemeandisplacementofautomaticallyselectedpointland- marks. The ITK voxel-similarity-based registration algorithm performed better than the surface-based approach with mean misregistration in the range of 7.7—8.4mm for CT—CT registration, 8.2mm for MR—MR registration, and 14.0—18.9mm for MR— CT registration compared to mean misregistration from the surface-based technique in the range of 9.6—11.1mm for CT—CT registration, 9.2—12.4mm for MR—MR reg- istration, and 15.2—19.0mm for MR—CT registration. © 2005 Elsevier Ireland Ltd. All rights reserved. 1. Introduction The evolution of imaging and computing technol- ogy has created the opportunity to perform quan- titative analysis of medical images. Tomographic imaging techniques such as computed tomogra- phy (CT) and magnetic resonance imaging (MRI) ∗ Corresponding author. E-mail address: chapmanbe@upmc.edu (B.E. Chapman). have reduced the amount of structural noise in im- ages [1] facilitating the ability to isolate individual anatomical structures within the images. Multiple contrast mechanisms can be exploited to create images with different depictions of anatomy and pathology either by using multiple modalities (e.g. CT, PET and MRI) or by exploiting different con- trast mechanisms within a modality (e.g. T1 and T2 MRI imaging). Concurrently, improvements in com- puterhardwarehaveprovidedthecomputingpower needed for rapid manipulation of medical images, 0169-2607/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.cmpb.2004.12.006