CSE DSI Machine Learning Seminar with Christopher Metzler (University of Maryland)

An ML-based approach to imaging through scattering media

Imaging through scattering is arguably the most important open problem in optics. If one could overcome scattering, one could (a) see through tissue to observe "biology in action" at cellular scale; (b) see through fog, smoke, and inclement weather to safely navigate in adverse conditions; (c) see through the atmosphere and allow ground-based telescopes to outperform James Webb for a fraction of the cost; and (d) see through thin fiber bundles to enable minimally invasive endoscopy.

This talk first describes our recently-developed guidestar-free approach to imaging through scattering and other optical aberrations; neural wavefront shaping (NeuWS). NeuWS integrates maximum likelihood estimation, measurement modulation, and neural signal representations to reconstruct diffraction-limited images through strong static and dynamic scattering media without guidestars, sparse targets, controlled illumination, nor specialized image sensors. We will then describe how NeuWS's performance can be further improved using "end-to-end" machine learning.

Chris Metzler is an Assistant Professor in the Department of Computer Science at UMD, where he leads the UMD Intelligent Sensing Laboratory. He is a member of UMIACS and has a courtesy appointment in the Electrical and Computer Engineering Department. His research develops new systems and algorithms for solving problems in computational imaging and sensing, machine learning, and wireless communications. His work has received multiple best paper awards; he recently received ARO Early Career Program, NSF CAREER, and AFOSR Young Investigator Program awards; and he was an Intelligence Community Postdoctoral Research Fellow, an NSF Graduate Research Fellow, a DoD NDSEG Fellow, and a NASA Texas Space Grant Consortium Fellow.