FIP Virtual Student Seminar "Computational 3D imaging, from optical coherence tomography and microscopes to your smartphone"

This event has passed.

Wed, 03/03/2021 - 12:00 to 13:00

Kevin Zhou

Presenter

Kevin Zhou, BME Ph.D. Candidate, Duke University


KEVIN ZHOU SEMINAR TALK 

Computation has become a powerful, even indispensable ally of modern imaging. Combined with an intimate understanding of the underlying physical mechanisms of image formation, computational imaging allows us to fuse multiple measurements to uncover hidden information about the sample under investigation. In this talk, I will present several applications of computational imaging from my research, a common theme among which is combining multi-angle measurements to obtain high-resolution 3D reconstructions. In particular, I will first present optical coherence refraction tomography (OCRT), a new technique that introduces angular diversity to optical coherence tomography (OCT), which not only enhances the resolution and reveals new structural features, but also estimates a quantitative refractive index (RI) map of various biological samples. Next, I will show that we can modify the OCRT computational reconstruction algorithm to reconstruct a 3D surface profile from a sequence of smartphone camera images acquired from multiple views under freehand motion. We achieve tens-of-micron accuracy over multi-centimeter fields of view on a variety of biological and non-biological samples. Finally, I will extend this concept and demonstrate snapshot 3D profilometric imaging using a gigapixel-scale, multi-aperture microscope, thus opening the door to real-time 3D monitoring of dynamic samples at simultaneously high resolution and large fields of view.

Kevin Zhou is a sixth-year PhD student, coadvised by Profs. Joseph Izatt and Warren Warren and works closely with Profs. Sina Farsiu and Roarke Horstmeyer. He is supported in part by an NSF Graduate Research Fellowship. His research interests are broadly in computational imaging, optical imaging, and machine learning.