Super-Resolution Tensor Image Recovery from Adaptive Sub-Sampling

Friday, January 25, 2019 - 13:00 to 14:30
Room 120, New CS

CS/IACS Seminar - All are welcome!

Speaker: Professor Chandrajit Bajaj, Department of Computer Science and Institute for Computational and Engineering Sciences (ICES), UT Austin

Super-Resolution Tensor Image Recovery from Adaptive Sub-Sampling

Abstract: Hyperspectral images (HSI), acquired as spatio-spectral broadband (visible to infra-red) light intensity tensors of order 3, are naturally voluminous. They capture material reflection or refraction intensities for multiple spectral bands (e.g., a few thousand) and so HSI's are easily two or three orders of magnitudes larger than color multispectral images (MSI). Current HSI sensors, however, have limited spatial resolution yielding poorer quality imagery. This work leverages the spectral sparsity of an input HSI and MSI pair, and computationally generates a digitally enhanced super-resolution image (SRI) recovery algorithm. The SRI recovery additionally utilizes only a minimal, adaptive sub-sampling of the HSI and MSI, making our technique scalable to extremely large HSI (MSI) images. We provide accuracy vs. speed tradeoffs with recovery guarantees, based on spectral sparsity bounds, and incoherency measures. Additionally, experiments on various data domains demonstrate the practical usefulness of our adaptive super-resolution technique.

This is joint work with Tianming Wang.

Short Bio: Chandrajit Bajaj is a Professor of Computer Science and the director of the Center for Computational Visualization in the Institute for Computational and Engineering Sciences (ICES) at the University of Texas at Austin. Bajaj holds the Computational Applied Mathematics Chair in Visualization. He is also an affiliate faculty member of Mathematics, Bio-medical Engineering, the Institute of Cell and Molecular Biology and Neurosciences. He currently serves on the editorial boards for the International Journal of Computational Geometry and Applications, the ACM Computing Surveys, and the SIAM Journal on Imaging Sciences. He is a fellow of the American Association for the Advancement of Science (AAAS), the Association of Computing Machinery (ACM), the Institute of Electrical and Electronic Engineers (IEEE), and the Society of Industrial and Applied Mathematics (SIAM).


Computed Event Type: 
Event Title: 
Super-Resolution Tensor Image Recovery from Adaptive Sub-Sampling