Computer Vision News - November 2020

7 aspect to allowing the registration of multiple image sequences of the same subject taken during one scanning session, specifically aligning cardiac 2D long axis images with the 3D short axis volume. With this sort of cardiac data, it is also important to ensure the time information is preserved so that images are rendered at the correct point in the cardiac cycle. ” Eric points out that “Eidolon has a primary focus on cardiac images, and it uses the MIRTK toolkit for image registration and motion tracking, so registration specifically is the alignment of long axis images to the short axis of the same subject [...] MIRTK is also used for motion tracking of the short and long axis images. This information can be used to estimate mechanical properties like strain, deform a segmentation image or mesh over time to calculate other properties like volumes, or to generate time-dependent meshes to use with cardiac simulations. Typically, cardiac images have time specified in milliseconds since end diastole so this allows the calculation of biomarkers over the cardiac cycle which can be used to assess heart function.” The words from Eric highlight the main use of Eidolon - cardiac images - and its close connection to MIRTK, the research-focused image processing toolkit developed at the BioMedIA research group, which provides some of the underlying processes available on the platform. One absolutely amazing thing about Eidolon is the level of attention and details employed to build a GUI that can help the user visualize the data as well as possible. To open it, once the source folder is downloaded, one just needs to activate the Python environment with all the required packages and run the following command: Registration Tool Comparison Its main developer, Eric Kerfoot, Software Architect at King’s College London, told us that Eidolon “was developed as a visualization tool focused on rendering image and mesh data together for viewing biophysics simulation results.” He also added that “One important aspect of development was to ensure images were rendered in their correct positions in space as defined by their transforms, this has to work correctly for images loaded from a range of file types. Ensuring image spatial information was correctly used was an important . /run.sh foo.py