Open source image registration: the elastix toolbox

Year of award: 2024

Grantholders

  • Prof dr Marius Staring

    University of Leiden Medical Center, Netherlands

Project summary

Image registration is the task of finding the spatial relationship between two or more images, and is abundantly used in biomedical image processing. The applicants have previously developed an open source tool for performing image registration, dubbed Elastix. While being one of the most popular image registration toolboxes,efforts are needed to sustain the software, better connect to (novice) users and the other software they use, and improve runtime performance. In a currently running CZI-funded project (EOSS4-0000000028) we have made substantial foundational improvements to Elastix so that almost all functionality has become available via the library interface and accessible to biologists via a Napari plugin. Building on the success of EOSS4-0000000028, we propose the following. A. FOUNDATIONAL. In this task we implement support for providing landmark files as input to the Elastix library interface. B. INTEROPERABILITY. Here, we further improve the friendliness of Elastix to other software. C. PERFORMANCE. In this task we will adopt threadpools to reduce threading overhead throughout Elastix. D. MAINTENANCE is an important part of software development. E. OUTREACH. We will create and present demos and a website on image registration resources. F. THE NEXT LEAP. We will make Elastix available via a web browser.