Topic 01a - Pan3D: Open Source Scalable and Reproducible Scientific Workflows for 3D Data Analytics

The scientific community has been building complex systems and consequently designing and developing scientific workflows, the common elements or stages of acquisition, integration, reduction, model development, visualization, and publication. These workflows also typically demand data and model sharing and steps to reproduce a particular finding with the team and outside collaborators. Traditional software lacks the necessary collaboration features to sup- port a multi-faceted team. A scientific area with particular needs for workflow improvement is the three-dimensional (3D) scientific computing domain, which requires dealing with more complex and larger 3D and 4D data structures, algorithms, and niche software. A large volume and variety of potentially big 3D datasets mean that workflows also need to address scalability. Additionally, existing 3D software solutions lack an integrated environment that simultaneously facilitates data management and large data visualization. Finally, like commercial 1D and 2D data analytical offerings, the 3D scientific community needs to adopt state-of-the-art web and cloud technologies to create rich, scalable, and shareable workflows that increase functionality and productivity. To address the shortcomings of the existing 3D workflows to serve, share, consume, and visualize 3D data and provide an interactive, reproducible, and integrated 3D computing environment, we propose to build Pan3D. Pan3D will provide a framework and modular system that will support scientific workflows for 3D datasets. It will support accessing and transforming 3D data from remote servers and visualizing data interactively for algorithm or modeling development. Pan3D will enable users to use existing open-source tools to perform large data processing asynchronously on the cloud. Out of the box, Pan3D will support data transformations and interoperability between modern data structures needed for algorithm development and visualization, data ingest, access, and cataloging, an interactive client and server-side 3D visualization. Pan3D will bridge the gap between widely used 3D visualization and analytics data formats to maximize data interoperability and productivity. This work will further our open-source platforms for data integration, algorithm development, and visualization while providing a turn-key platform for scientific communities. Such platforms provide Kitware with significant business opportunities, as open software platforms lend themselves to a services business model where the company can provide technology integration services with customers and collaborators to build valuable, competitive products and services.