Antscan uses a synchrotron-powered CT pipeline to quickly generate high-detail X-ray scans, producing roughly 3,000 images per scan and yielding hundreds of 3D models of ants.

The 3D models reveal internal anatomy—muscles, nervous and digestive systems, and stingers—at micrometer resolution, with raw data and an integrated viewer made openly accessible.

Led by Evan Economo at the University of Maryland, with KIT collaborators, Antscan combines a high-throughput workflow, robotics, and AI to digitize 800 ant species into interactive 3D models.

The project offers an open-access library of raw 3D data and visualization tools, accelerating discovery and cross-disciplinary collaboration in biology, morphology, and education.

Overall, Antscan represents a significant step toward an open, scalable library of life for 3D morphology, enabling broad access and new applications in science and education.

The Nature Methods paper on Antscan provides a blueprint for large-scale quantification of organismal form and discusses its implications for biodiversity research.

Economo envisions wide-ranging applications of digitized biodiversity data—from labs and classrooms to Hollywood—and plans to expand specimen scanning while continuing AI work with UMD students.

The project’s published workflow outlines how high-throughput data can transform the study of morphology at scale.

Antscan stands as a transformative, interdisciplinary effort to digitize global ant biodiversity with broad implications for science, education, and media.

AI-driven pose estimation is being developed to convert contorted scan poses into lifelike poses, enhancing realism for research, education, and entertainment.

All raw data are openly available, and an online viewer lets users explore each model, democratizing access to high-resolution morphological data for researchers, educators, and artists.

Antscan is a freely accessible database derived from thousands of ants across hundreds of genera, forming an open-source library of 3D ant scans and morphological data.