Computational Imaging for Micro- and Nano-structures in Materials ScienceMS11

In microscopy, the proliferation of digital data resulting from an increasing number of sensors, and automated data capture, has created opportunities for rich characterization of structures in materials. Whereas traditional characterization involved small numbers of 2-D images that were hand analyzed, modern techniques can capture image sequences involving hundreds of individual images, leading to characterization over "large volumes," as compared with traditional approaches. Multiple sensors have allowed for multimodal data collection; robotics has allowed for automated data preparation and collection; and increasing storage capacities have lead to large datasets. The availability of this type of data, which primarily involves imaging modalities, has created an opportunity for modern Computational Imaging (CI) methods to be applied to obtain vast improvements in the quality of micro- and nano-structure analyses as well as their complexity. Traditionally, quantitative analysis of microscope imagery has relied upon forward modeling for testing hypotheses. CI, with its emphasis on inversion methods, represents a largely unexplored methodology in this field. This symposium brings together CI researchers from Imaging Science, Materials Science, and other fields who are applying techniques such as statistical image processing, model-based regularization, sparsity, denoising methods, optics, and the integration of the physics of structure formation towards advancing the science of microscopy.

Physics-based Regularization for Denoising Polycrystalline Material Images
Jeff Simmons (Air Force Research Laboratory)
Incorporating Physical Constraints and Regularization in Min-cut/Max-flow Graph Partitioning for Segmentation and Clustering in Materials Imaging
Stephen Niezgoda (The Ohio State University)
Convolutional sparse coding based regularizers for tomographic inverse problems
Singanallur Venkatakrishnan (Oak Ridge National Laboratory)
Regularized Image Reconstruction for Nonlinear Diffractive Imaging
Ulugbek Kamilov (Washington University in St. Louis)
SLADS: Fast Dynamic Sampling using Machine Learning
Charles Bouman (Purdue University)
Sparse Sampling in Scanning Electron Microscopes
Kurt Larson (Sandia National Laboratories, US Department of Energy)
Jeff Simmons (Air Force Research Laboratory)
Brendt Wohlberg (Los Alamos National Laboratory)
computational imaging, computed tomography, image reconstruction, inverse problems