Visualization Toolkit for Nanoscale Imaging Data

High brightness X-ray sources such as synchrotrons are used by thousands of researchers across the United States for cutting-edge research. Data collected in these facilities is complex, often multi-channel and multi-modal. Visualizing such datasets is extremely difficult, often requiring substantial human effort to select suitable display parameters that highlight the complex relationships within this data. The inability to quickly and accurately visualize this data significantly reduces the impact of scientific discovery at high brightness user facilities. To enable scientists to accurately and efficiently visualize nanoscale imagining data by (1) simultaneously visualizing multi-channel and multi-modal two-dimensional and three-dimensional microscopy/imagining datasets; (2) providing visualization parameters that autonomously adjust to end- user needs; (3) generating meaningful first impressions about data; (4) learning associations among structures within multi-channel and multi-model imagery; and (5) connecting to existing current microscopy data management efforts within the Department of Energy, allowing machine learning algorithms to learn optimal display parameters from shared datasets. The small business will develop a set of coordinated software tools that can simultaneously visualize multi-channel and multi-model two-dimensional and three-dimensional microscopy/ imagining datasets. For the Phase I feasibility study, the small business will focus on visualizing multi-channel and multi-modal X-ray computed fluorescence tomography data. Target customers include federally funded research and development centers and other independent research institutions and light source user facilities utilizing high-brightness X-ray sources. The Basic Energy Science programĺs user facilities including Advanced Photon Source, Advanced Light Source, and National Synchrotron Light source II are all target facilities, with over 1,000 annual users per facility. The proposed software and machine learning models are applicable to nanoscale and microscopy use cases at 42 national laboratories and over 250 academic research institutions utilizing high- brightness X-ray sources, as well as electron microscopes as a secondary customer base.