ECE Seminar: Sundaresh Ram

Sundaresh Ram PhD Candidate University of Arizona Faculty host: Vivek Goyal Light refreshments will be available outside of PHO 404/428 at 9:45 am. Title: Sparse modeling for solutions to inverse imaging problems Abstract: Sparse presentations of signals have drawn a considerable amount of interest in the recent years. The assumption that natural signals, such as images, admit a sparse decomposition over a redundant dictionary leads to efficient algorithms for handling such sources of data. Standard sparse representation however does not consider the intrinsic and geometric structure present in the data, thereby leading to sub-optimal results. Using the concept that a signal is block sparse in a given basis—i.e., the non-zero elements occur in clusters of varying sizes—in this talk, we will first present a novel and efficient framework for learning sparse presentation modeling of natural images, called graph regularized block sparse representation (GRBSR). Next, we apply the proposed GRBSR towards two applications of image restoration. In the first application, we will present how the GRBSR can be used to learn a dictionary-based local regression model for super-resolving a single low-resolution image without any external training image sets. In the second application, we will present a novel multiresolution exemplar-based image inpainting technique, where we will show how a GRBSR-based learnt dictionary can be used to generate new pixels in missing regions of the image for removing objects from the image and make it look visually plausible to the human eye. Results validating the performance of the GRBSR-based dictionary learning technique for evaluating single image super-resolution and image inpainting methods in natural images will be presented. Bio: Sundaresh Ram is a Ph.D. Candidate in the Department of Electrical and Computer Engineering, University of Arizona, Tucson, Arizona, USA. He received the B.E. degree in Electrical and Electronics Engineering from Anna University, Chennai, India in 2007, and the M.S. degree in Electrical and Computer Engineering at University of Arizona in 2010. His research interests include signal/image/video processing and analysis, computer vision, machine learning, computational imaging, and inverse problems, and applications of these methods in the biological sciences, conventional and medical imaging and data science. A current focus of his research has been on solving problems involving adaptive imaging, sparse representations, compressive sensing, and data (image and video) recovery. He is a member of the IEEE and SIAM.