AI-TOMO: Accelerated materials characterisation by AI and X-ray tomography

Purpose and goal:

The purpose of the project is to accelerate characterisation of materials structures by AI and X-ray tomography to facilitate development of new sustainable products and processes of the future. The goal is to develop AI algorithms for fast, effective segmentation and quantification of 3D and 4D X-ray tomography data. The developed algorithms will support materials development in the participating companies and provide a foundation for supporting the wider materials industry.

Expected results and effects:

The tools will be developed to streamline the analysis process, delivering faster and more precise results that enhance decision making and foster innovation. This initiative aims to optimize utilization of synchrotron experiments and encourage more industrial researchers to leverage AI, tomography, and synchrotron facilities for product development. The hope is to significantly enhance the speed and quality of product development, thereby boosting the competitiveness of Swedish industry.

Approach and implementation:

The project will use existing 3D and 4D structures as well as acquire new data using synchrotrons or lab-tomographs available at the partners for fiber materials, ganular and porous binary materials. These will be annotated and used for training, testing and validation of AI models. AI models based on SGMs or CNNs for 3D, 4D and live segmentation of these structure classes and transfer models will be developed and validated. The tools will be packed and implemented at MAX IV.