Introducing Catalyst Vision: AI-Powered Analysis for Advanced Energy Materials
Catalyst Vision
Our human-in-the-loop, deep learning software enables rapid semantic segmentation of nanocatalyst features, providing the critical statistical data needed to link material structure to performance.
Developing next-generation catalysts for advanced energy applications, such as hydrogen production, is a key objective in materials science. A catalyst's effectiveness and longevity are intrinsically linked to its nanoscale structure—the size, shape, and arrangement of its active particles. However, tracking how this structure evolves during operation presents a significant challenge, often requiring countless hours of manual image analysis. This slow, subjective process hinders our ability to quickly understand degradation mechanisms and design more robust materials.
To address this challenge, we have developed Catalyst Vision, an open-source software package designed to automate and accelerate this analysis. Catalyst Vision employs an effective, human-in-the-loop deep learning workflow, allowing researchers to rapidly segment and quantify nanostructures from large electron microscopy datasets with minimal manual input. By turning complex images into actionable data, this tool is intended to accelerate the discovery of structure-property relationships and enable the rational design of next-generation energy materials.