A group in Vienna has established a centralized hub geared toward accelerating total-body PET AI research – namely, a website called enhance.pet.
An acronym for “Enabling New Horizons for Advanced Networking, Code-sharing, and Education,” enhance.pet is a platform for anybody interested in molecular imaging research, specifically total-body PET imaging, according to the initiative’s collaborators.
The platform hosts a bevy of released and soon-to-be-released data analytics tools, such as the following:
- Moose (multiorgan objective segmentation), a data-centric AI algorithm designed to streamline systemic total-body research by generating precise multilabel organ segmentations
- Falcon (fast algorithms for motion correction), a fully automatic motion correction algorithm tailored for dynamic total-body or whole-body PET imaging across various vendors, tracers, and organs
- Ocelot (diffeomorphic registration for voxel-wise anomaly tracking), a tool designed to detect patient anomalies by leveraging a normative database for precise reference PET/CT comparisons
In addition, the platform links users with leading global educational initiatives, such as the Open Kinetic Modeling Initiative led by Guobao Wang, PhD, of the University of California, Davis, and the Australian National Total-body PET webinar series.
The website was officially launched during the EANM 2023 meeting in Vienna as part of a presentation by Lalith Kumar Shiyam Sundar, PhD, of the University of Vienna’s Quantitative Imaging and Medical Physics (QIMP) group.