Dr. Saiprasad Ravishankar
Michigan State University
Date/Time: Friday, March 25th, 4:30 pm
Zoom Meeting ID: 945 8994 6946 (Passcode: 896970)
Abstract: In this talk, we present approaches for learning sparsity-based and deep neural network-based regularizers for inverse problems in imaging. Our methods allow learning robust models from very limited training data. We will first present an approach for model-based image reconstruction (MBIR), where the sparsity-based regularizer’s parameters are learned in an unsupervised manner from a very small set of high-quality images. We pre-learn a union of sparsifying transforms that can cluster image patches into multiple groups, with a specific transform well-matched to each group. When incorporated in the regularizer in MBIR, the learned transforms provide much better image reconstructions in CT compared to conventional filtered back projection and non-adaptive regularization methods, especially at low X-ray doses. We then develop supervised learning of sparsity-promoting regularizers, where the parameters of the regularizer are learned to minimize reconstruction error on a paired training set. Training involves a challenging bi-level optimization problem with a nonsmooth lower-level objective. We derive an expression for the gradient of the training loss using the implicit closed-form solution of the lower-level variational problem given by its dual problem and provide an accompanying gradient descent algorithm (dubbed BLORC) to minimize the loss. Our experiments on 1D signals and natural images show that the gradient computation is efficient, and BLORC learns meaningful operators for signal reconstruction. Finally, we develop a new approach for unified supervised-unsupervised (SUPER) learning of regularizers that combines classical MBIR optimization and unsupervised transform learning regularization together with supervised deep learning in a common formulation. We provide multiple interpretations of the resulting scheme from fixed point iteration analysis or bi-level training optimization and show that with limited training data, it provides much better CT image reconstructions at low X-ray doses than the constituent supervised or unsupervised schemes. We also show a variation of the framework involving dictionary-based blind or on-the-fly learning and deep supervised learning for MR image reconstruction from limited data. The results show that the patient-adaptive features captured by the blind dictionary-based approach complement the deep network learned (population-adaptive) features to provide significantly better reconstructions with limited training data.
About the Speaker: Saiprasad Ravishankar is currently an Assistant Professor in the Departments of Computational Mathematics, Science and Engineering, and Biomedical Engineering at Michigan State University. He received the B.Tech. degree in Electrical Engineering from the Indian Institute of Technology Madras, India, in 2008, and the M.S. and Ph.D. degrees in Electrical and Computer Engineering in 2010 and 2014, respectively, from the University of Illinois at Urbana-Champaign, where he was then an Adjunct Lecturer and a Postdoctoral Research Associate. Since August 2015, he was a postdoc in the Department of Electrical Engineering and Computer Science at the University of Michigan. He was a Postdoc Research Associate in the Theoretical Division at Los Alamos National Laboratory from August 2018 to February 2019. His interests include signal and image processing, biomedical and computational imaging, machine learning, inverse problems, and large-scale data processing and optimization. He has received multiple awards, including the Sri Ramasarma V Kolluri Memorial Prize from IIT Madras and the IEEE Signal Processing Society Young Author Best Paper Award in 2016 for his paper “Learning Sparsifying Transforms” published in the IEEE Transactions on Signal Processing. A paper he co-authored won the best student paper award at the IEEE International Symposium on Biomedical Imaging (ISBI) in 2018, and other papers were award finalists at the IEEE International Workshop on Machine Learning for Signal Processing (MLSP) in 2017 and ISBI 2020. He is currently a member of the IEEE Computational Imaging Technical Committee.