Particle Accelerator and AI Collaboration Digitizes Ant Biodiversity at Unprecedented High-Throughput Speeds

University of Maryland researchers use a particle accelerator and AI to scan 800 ant species, creating a high-resolution digital library of insect anatomy.

By: AXL Media

Published: Mar 11, 2026, 10:55 AM EDT

Source: Information for this report was sourced from University of Maryland

Revolutionizing the Speed of Biological Digitization

The field of insect morphology has long been hampered by the slow pace of high-resolution imaging, often requiring hours to capture a single specimen. A breakthrough collaboration between the University of Maryland and the Karlsruhe Institute of Technology has fundamentally altered this timeline by integrating particle physics with biological research. By utilizing a synchrotron accelerator and specialized robotics, the "Antscan" project managed to compress what would normally be six years of continuous laboratory work into just seven days. This massive leap in efficiency allows for the creation of extensive digital archives that can be shared globally across scientific and educational platforms.

The Mechanics of High-Speed Micro-CT Scanning

To achieve such rapid data collection, the research team utilized intense X-ray beams generated by a synchrotron, which are far more powerful than standard laboratory CT scanners. During the process, a robotic sample changer rotated ethanol-preserved specimens and replaced them every 30 seconds, creating a conveyor belt of biological data. This high-throughput method produces massive stacks of two-dimensional images that are later reconstructed into three-dimensional models. The magnification levels reached in this process allow for the inspection of features at the micrometer level, providing a level of detail that surpasses conventional microscopy.

Artificial Intelligence and Pose Correction

One significant challenge in scanning preserved insects is that specimens are often fixed in unnatural or distorted positions. To resolve this, computer science students at the University of Maryland developed AI tools specifically designed for automated pose estimation. This technology digitally manipulates the raw scan data to reposition the ants into lifelike stances, mirroring how they would appear in their natural habitats. This computational layer is essential for making the models useful for behavioral studies and public engagement, transforming static, awkward scans into interactive and biologically accurate digital organisms.