[Photobears] Dissertation Talk: Computational Hyperspectral Microscopy for Bioimaging

[Neerja Aggarwal]

Please see below. 

Sent from iPhone, please excuse brevity and typos.

---------- Forwarded message ---------
From: Neerja Aggarwal <nee...@berkeley.edu>
Date: Fri, Apr 25, 2025 at 9:00 AM
Subject: Dissertation Talk: Computational Hyperspectral Microscopy for Bioimaging
To: <eecs-a...@lists.eecs.berkeley.edu>

 

Title: Dissertation Talk: Computational Hyperspectral Microscopy for Bioimaging

Speaker: Neerja Aggarwal

Advisor: Laura Waller

Date: Wednesday,  May 7th, 2025

Time: 10:00am - 11:00am PT

This is a hybrid event held in person and virtually over Zoom.
Location (In-person): 510 Soda Hall
Location (Zoom): https://berkeley.zoom.us/my/neerja

Reception afterwards ~11am-12 pm - please RSVP for food head count.

Abstract: Hyperspectral imaging involves detecting the spectrum (intensity vs wavelength) of light emitted at each point in space. It has applications in biology and medicine such as imaging of live cells and interferometry to see inside tissue. Computational imaging involves the codesign of both optics and algorithms together to beat traditional limitations. In this work, we present three imaging systems for various bioimaging applications that benefit from computational imaging to improve spectral imaging performance.

In the first project, we redesigned a traditional spectrometer for spectral domain optical coherence tomography, an eye imaging technique. We used a diffuser instead of a grating to diffract light reducing the size of the device and solved a simple inverse problem to reconstruct the spectrum. In the second project, we developed a compact imager that enables snapshot hyperspectral imaging on a traditional benchtop fluorescence microscope. We used a diffuser to multiplex light onto a spectral filter array on an image sensor and used compressed sensing to solve for more voxels in the hyperspectral data cube than pixels on the sensor. Our experimental reconstruction results include fluorescent biological samples such as labeled cells and bioassay beads.  In the final project, investigated a new idea of spatial frequency scanning of the hyperspectral datacube. We adapted a Fourier ptychography microscopy system for spectral imaging using a filter array and demonstrated reconstruction with limited measurements using priors in simulation. We hope this work can be expanded into an experimental setup for digital pathology.

More details: https://events.berkeley.edu/events/event/298526-dissertation-talk-computational-hyperspectral

--

--

Neerja Aggarwal

Ph.D. Candidate

Computational Imaging Lab

UC Berkeley EECS

[Neerja Aggarwal]

This is now! 

Sent from iPhone, please excuse brevity and typos.