About

Shwetadwip Chowdhury, Ph.D., is a researcher specializing in computational imaging technologies. As of the information provided, he is a postdoctoral researcher in the Electrical Engineering and Computer Science (EECS) department at UC Berkeley, working with Professor Laura Waller in the Computational Imaging Laboratory. His research focuses on designing imaging systems and algorithms that integrate hardware and software to develop next-generation optical imaging technologies. These technologies utilize computational techniques to enhance optical hardware, enabling significant improvements in imaging resolution, speed, throughput, and simplicity. His current projects include multi-dimensional imaging and big-data inversion for optical super-resolution and diffraction tomography. Dr. Chowdhury earned his Ph.D. in 2017 from Duke University, where he was part of the Biomedical Engineering Department's Biophotonics Group under the supervision of Professor Joseph Izatt. His doctoral thesis explored structured illumination microscopy as a solution for multimodal 3D super-resolution, which laid the foundation for his postdoctoral work. He planned to transition to an assistant professorship at the University of Texas at Austin in February 2021, where he intended to establish a research lab and recruit graduate students and postdoctoral researchers.

Research topics

• Optics
• Physics
• Computer Science
• Materials science
• Algorithm
• Computational physics
• Mathematical analysis
• Geometry
• Mathematical optimization
• Nanotechnology
• Mathematics

Selected publications

• Angular scattering data (scattering phantom and various biological samples)

  Texas Digital Library (University of Texas) · 2026-01-30

  datasetOpen accessSenior author

  This dataset contains angular scattering measurements (complex-field measurements) of a scattering phantom, C.elegans, intestinal organoids and zebrafish embryo. Parameters of the angle-scanning imaging system and parameters used for MSBP-based inverse scattering are also included.<br><br> Matlab code to run the MSBP-based inverse-scattering to recover the sample's 3D RI can be found on Github: https://github.com/ut-cwo/Inverse-scattering-in-biological-samples-via-beam-propagation<br><br> If this dataset is useful to you, please consider citing: Kim, Jeongsoo, et al. "Inverse-scattering in biological samples via beam-propagation." bioRxiv (2025).

  PublisherDOI

• Inverse-scattering of absorptive samples via beam propagation

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-04-24

  articleOpen accessSenior authorCorresponding

  Abstract Inverse-scattering methods enable label-free, quantitative visualization of a sample’s three-dimensional (3D) refractive index (RI), providing intrinsic and volumetric morphological contrast without exogenous labels. This is achieved by developing computational frameworks that reconstruct the sample’s 3D RI from a series of scattering measurements acquired under different data-capture conditions. Recent advances have demonstrated successful 3D RI reconstructions in multiple-scattering samples using angle-varying illuminations; however, these studies have primarily focused on non-absorptive samples. Here, we extend the multi-slice beam propagation (MSBP) inverse-scattering framework to reconstruct complex-valued RI, encompassing both the sample’s conventional RI (real part) and absorptivity (imaginary part). We show that reconstructing complex-valued RI makes the inverse problem ill-posed under angle-varying illumination alone, and that incorporating measurement diversity from both angle-varying illumination and sample defocus is necessary to ensure stable and accurate convergence. Experimental demonstrations were conducted on 1) dyed microsphere samples to characterize accuracy of reconstructed RI and absorptivity; and 2) diverse absorptive scattering samples to demonstrate biological utility. These results represent an important step for label-free volumetric imaging in biological tissue, which typically exhibits both scattering and absorption.

  PublisherDOI

• Inverse-scattering in biological samples via beam-propagation

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-08-22 · 2 citations

  preprintOpen accessSenior authorCorresponding

  Multiple scattering limits optical imaging in thick biological samples by scrambling sample-specific information. Physics-based inverse-scattering methods aim to computationally recover this information, often using non-convex optimization to reconstruct the scatter-corrected sample. However, this non-convexity can lead to inaccurate reconstructions, especially in highly scattering samples. Here, we show that various implementation strategies for even the same inverse-scattering method significantly affect reconstruction quality. We demonstrate this using multi-slice beam propagation (MSBP), a relatively simple nonconvex inverse-scattering method that reconstructs a scattering sample's 3D refractive-index (RI). By systematically conducting MSBP-based inverse-scattering on both phantoms and biological samples, we showed that an amplitude-only cost function in the inverse-solver, combined with angular and defocus diversity in the scattering measurements, enabled high-quality, fully-volumetric RI imaging. This approach achieved subcellular resolution and label-free 3D contrast across diverse, multiple-scattering samples. These results lay the groundwork for robust use of inverse-scattering techniques to achieve biologically interpretable 3D imaging in increasingly thick, multicellular samples, introducing a new paradigm for deep-tissue computational imaging.

  PublisherOA PDFDOI

• Maskless and on-chip LED-array microscope with spatially-varying angle calibration for centimeter-scale phase imaging

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-09-10 · 1 citations

  preprintOpen accessSenior authorCorresponding

  Abstract On-chip imaging with LED-array-based angled illumination offers a cost-effective approach for large field-of-view (FOV) phase imaging. However, it faces two main challenges: (1) twin-image ambiguity can degrade phase reconstruction. While mask-based modulation can help, it adds system complexity due to fabrication and alignment requirements; and (2) the illumination angle from each LED varies across large FOVs, and can degrade centimeter-scale phase reconstruction without calibration. Here, we present a computational framework to jointly achieve mask-free on-chip phase imaging and adaptive calibration of spatially-varying illumination angles. The sensor FOV is divided into subregions within each of which LED illumination is approximated as planar. LED illumination angles for each subregion are initialized geometrically. Phase retrieval is then performed within each subregion by constraining the reconstruction with a soft optical transparency prior while simultaneously refining angle estimates. Reconstructed phase maps are merged to produce a high-quality, large-FOV phase image. We demonstrate this approach by achieving centimeter-scale on-chip phase imaging (up to 2.7 × 1.7 cm 2 ) with micron-level resolution across various biological tissue sections.

  PublisherOA PDFDOI

• Correction of Inter-frame Translational Motion in Optical Diffraction Tomography

  2025-01-01

  articleSenior author

  We propose a 3D space-time inverse-scattering framework to correct translational sample-motion occurring during data collection in optical diffraction tomography. By jointly solving for the sample’s 3D refractive-index (RI) alongside its frame-to-frame translational motion, our method achieves motion-corrected 3D RI reconstruction.

  PublisherDOI

• Demonstrating the Utility of a Pipeline for Microwave Breast Image Analysis: A Two-Factor Variation Study

  2025-07-13

  article

  Microwave imaging (MWI) is a promising non-invasive technique for breast cancer screening, utilizing differences in tissue dielectric properties to produce images. This study demonstrates the potential of a pipeline that enables quantitative, high-throughput analysis of reconstructed images. Microwave measurements were simulated in the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$1-5 \text{GHz}$</tex> frequency range using a simplified breast model, which includes a skin layer encasing homogeneous healthy tissue, with a tumor target embedded in the background. An 8 -antenna array was used to simulate data acquisition while the dielectric properties of the background as well as the tumor location were simultaneously and randomly varied. To conduct the simulation, models were constructed in HFSS, and computations were performed on Compute Canada's supercomputer, enabling efficient processing within a short timeframe. Image reconstruction was achieved using the open-source delay-and-sum (DAS) radar-based imaging algorithm MERIT. Metrics such as signal-to-noise ratio (SNR) and location error were applied to evaluate image quality. Preliminary results demonstrate the robustness and accuracy of the pipeline, successfully reconstructing tumors across ten scenarios with varying background permittivity and tumor locations. The findings underscore the importance of accurate tissue property modeling, as adjusting the input permittivity to account for the skin layer significantly improved image quality. Additionally, we highlight the effect of tumor position on reconstruction accuracy. In conclusion, this work provides valuable insights into the evaluation of MWI systems, introducing tools and metrics for testing performance and robustness in unknown scenarios.

  PublisherDOI

• 3D on-chip refractive index microscopy with multi-slice beam propagation

  2025-01-01

  articleSenior author

  We present a novel on-chip lens-free imaging approach that collects scattering measurements via angular illumination of a sample placed directly on a sensor. Multi-slice beam propagation model is reformulated to reconstruct high-resolution large field-of-view 3D refractive-index from these on-chip measurements.

  PublisherDOI

• Adaptive angle-calibration for wide field-of-view on-chip phase imaging with LED array

  2025-01-01

  articleSenior author

  We introduce an adaptive angle-calibration framework for mask-less on-chip phase imaging, where an LED array is used to deliver angled illuminations. Our framework accounts for spatially-varying illumination angles produced by each LED across expansive fields-of-view, enabling wide field-of-view phase imaging.

  PublisherDOI

• Computational bio-imaging via inverse scattering

  2025-03-19

  article1st authorCorresponding

  Optical imaging is a major research tool in the basic sciences, and is the only imaging modality that routinely enables non-ionized imaging with subcellular spatial resolutions and high imaging speeds. In biological imaging applications, however, optical imaging is limited by tissue scattering to short imaging depths. This prevents large-scale bio-imaging by allowing visualization of only the outer superficial layers of an organism, or specific components isolated from within the organism and prepared in-vitro. I present recent developments in our lab that design inverse-scattering methods to computationally unscramble the effects of scattering in optically thick samples. I will specifically discuss 1) novel computational microscope system designs that enable novel methods for data collection; and 2) the design and practical implementation of large-scale computational nonlinear and nonconvex frameworks that enable robust inverse-scattering. Real-world bio-imaging will be demonstrated on multiple-scattering organisms popularly used in the basic-sciences.

  PublisherDOI

• Breast Cancer Screening: Impact of Antenna Array Configurations on Microwave Imaging Quality

  2025-04-15

  article

  Microwave imaging (MWI) is a promising noninvasive technique for breast cancer screening, utilizing differences in dielectric properties between tissues to generate images. This study focuses on developing a robust imaging pipeline to optimize the design of specialized antenna arrays, enhancing measurement efficiency and reducing image reconstruction time. Simulations were conducted in the 1-5 GHz frequency range using a simplified breast model consisting of a single layer of skin surrounding a homogeneous healthy tissue, with a tumor target embedded within the background material. To test our pipeline, we used different antenna distributions for data acquisition. Additionally, we varied the dielectric properties of the background, the number of tumors, and their size and location. To streamline the simulation workflow, the models were built in HFSS, and simulations were processed on Compute Canada's supercomputer, allowing for efficient computation in a short time frame. Images were reconstructed using a delay-andsum (DAS) radar-based imaging algorithm called MERIT. Metrics such as signal-to-noise ratio (SNR) and location error were then used to assess image quality. Our preliminary findings indicate that while non-equally spaced antennas show potential for detecting targets near the sensors, equally spaced configurations consistently yield superior image quality. This work lays the groundwork for strategic diagnostic hardware design, aiming to improve MWI accuracy through refined antenna array configurations.

  PublisherDOI

Recent grants

• Structured Illumination Computational Microscopy with UV Surface Excitation (MUSE) for Multispectral Super-Resolution Histology

  NIH · $146k · 2018–2020

Frequent coauthors

• Joseph A. Izatt

  Duke University

  15 shared

• Laura Waller

  University of California, Berkeley

  14 shared

• Adam Wax

  Duke University

  9 shared

• David Ren

  University of California, Berkeley

  9 shared

• Will J. Eldridge

  Duke University

  6 shared

• Christophe Arnoult

  Inserm

  5 shared

• Regina Eckert

  Jet Propulsion Laboratory

  5 shared

• Nicole Repina

  University of California, Berkeley

  5 shared

Labs

• Shwetadwip Chowdhury LabPI

Education

• PhD, Biomedical Engineering

  Duke University

  2016

Awards & honors

• Fellow of the Jack Kilby/Texas Instruments Endowed Faculty F…
• NIH Ruth L. Kirschstein NRSA Postdoctoral Fellow
• NSF CAREER Awards
• Scialog: Advancing BioImaging Award
• Frontiers of Imaging Grant