Events

Jan 24, 2018
12:00 PM
149 13th Street (Building 149), Room 2204

Abstract:

The data-driven learning of signal models including dictionaries, sparsifying transforms, low-rank models, tensor and manifold models, etc., is of great interest in many applications. In this talk, I will present my research that developed various efficient, scalable, and effective data-driven models and methodologies for signal processing and imaging. I will mainly discuss my work in the recently developed field of transform learning. Various interesting structures for sparsifying transforms such as well-conditioning, double sparsity, union-of-transforms, incoherence, rotation invariance, etc., can be considered, which enable their efficient and effective learning and usage. Transform learning-driven approaches achieve high-quality results in applications such as image and video denoising, and X-ray computed tomography or magnetic resonance image (MRI) reconstruction from limited or corrupted data. The convergence properties of the learning-based algorithms will be briefly discussed. I will also present recent work on efficient synthesis dictionary learning in combination with low-rank models, and demonstrate the usefulness of the resulting LASSI method for dynamic MRI. The efficiency and effectiveness of the methods proposed in my research may benefit a wide range of additional applications in imaging, computer vision, neuroscience, and other areas requiring data-driven parsimonious models. Finally, I will provide a brief overview of recent works and some future pathways for my research. This will include topics such as physics-driven deep training of image reconstruction algorithms, data-driven learning of undersampling patterns in compressed sensing-type setups, online data-driven estimation of dynamic data from streaming, limited measurements, etc. 

 

About the Speaker: 

Saiprasad Ravishankar received the B.Tech. degree in Electrical Engineering from the Indian Institute of Technology Madras, in 2008. He received 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 an Adjunct Lecturer in the Department of Electrical and Computer Engineering during Spring 2015, and a Postdoctoral Research Associate at the Coordinated Science Laboratory until August, 2015. Since then, he has been a postdoctoral researcher in the Electrical Engineering and Computer Science Department at the University of Michigan. His research interests include signal, image and video processing, biomedical and computational imaging, data-driven methods, machine learning, signal modeling, inverse problems, compressed sensing, dictionary learning, and large-scale data processing and optimization. He has over 1200 Google Scholar citations and 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 for his paper “Learning Sparsifying Transforms” published in IEEE Transactions on Signal Processing. He has organized several sessions at IEEE conferences and workshops on up and coming research themes including one in ISBI 2018 on smart imaging systems.

Jan 31, 2018
12:00 PM
149 13th Street (Building 149), Room 2204

 

Abstract:

Partners institutions have a robust clinical informatics infrastructure supporting the research enterprise.  Over the past several years new tools have been developed to facilitate regulated access to medical image data collected during the conduct of routine clinical care.  The visionary leadership and research efforts in the Departments of Radiology at Partners institutions together with a high volume of patients yields a large volume of medical images in the institutional archives acquired with parameters comparable to the best quality research scans.  This talk will review the tools available to all Partners faculty and staff to identify, access and work with data extracted from the electronic healthcare records including the medical images. The newest of the suite of tools is the Partners Clinical Image Bank, which is a user friendly portal that enables interactive analysis and exploration of valuable image repositories.  An exemplar project that uses ADC values from brain MR images to identify neonatal hypoxic ischemic encephalopathy lesions will be used to illustrate the clinical potential of these tools. The Clinical Image Bank is in a period of expansion and the characteristics of new image registries that would be of greatest value will be discussed.

 

About the Speaker:

Randy L. Gollub is Professor of Psychiatry and Associate Director of Translational Research in the Neuroimaging Research Program at Massachusetts General Hospital. She is a recognized leader in the development and application of advanced neuroimaging technologies to understand the pathophysiology of neuropsychiatric disorders including neonatal hypoxic ischemic encephalopathy, chronic pain and schizophrenia. She is currently working on translating these advances into clinical radiology practice to improve patient care through the use of large-scale imaging informatics approaches.