Studies smart data structures and prediction optimization with provable guarantees for spectral and geometric processing

Collaborators: Dr. Manfred Auer (UCB), Dr. Rohit Bhargava (UIUC) and Dr. Pradeep Ravikumar

An in-depth understanding of the cancer microenvironment and progression requires an integrated bio-imaging approach that describes and emphasizes the integrated nature of the biological processes, such as baseline descriptions of cells and tissues or their changes during development or pathogenesis. Such changes typically include cellular metabolism, 3D architecture/ultrastructure and localization of constituents, such as proteins and other macromolecules. The aim is to visualize the progression of cancer in higher detail and more comprehensively than has been possible until now. Correlating imaging techniques such as chemical imaging (CI) and 3D electron microscopy (3DEM) provides an opportunity for studying the progression of cancer, in a much more detailed way than previous methods. In order to take full advantage of the new imaging techniques, we focus on developing computational tools for segmenting, classifying, interpreting and correlating the multi-modal cancer imaging data, then visualizing the results in a human-readable way. We also propose to create, adapt and further develop the computational infrastructure needed for comprehensive bioimaging that will overcome respective current bottlenecks in both compositional and architectural bioimaging, and to integrate compositional, architectural and localization information into a spatial data infrastructure framework that allows superposition and annotation of spatiotemporal image data.