Varun A. Kelkar



Algorithmic Systems Group, Analog Garage
Analog Devices, Inc.

I am a Research Scientist in the Algorithmic Systems Group within the Analog Garage, which is the research incubator of Analog Devices, Inc. My research is about computational imaging and sensing algorithms.

I defended my PhD in Electrical and Computer Engineering from University of Illinois at Urbana-Champaign in May, 2023. I was a part of the Computational Imaging Science Lab, advised by Prof. Mark A. Anastasio.

My PhD research was about developing ML-based algorithms for solving ill-posed imaging inverse problems, with applications to medical imaging. I have also worked on compressive microscopy, phase retrieval, physics of optical tomography, statistical optics and objective assessment of image quality.

I completed my undergrad (B.Tech.) in Engineering Physics from Indian Institute of Technology Madras in July 2017, and my MS in ECE from UIUC in August 2019.

I have done research internships at Mitsubishi Electric Research Labs (summer 2022), Algorithmic Systems Group at Analog Garage, Analog Devices Inc. (summer 2019), and LIGO 40m Lab, Caltech (summer 2016). I worked on various signal processing and imaging problems throughout these internships.

news

Apr 5, 2023 I gave a talk about our work on evaluating generative models in the 6th Health Data Analytics Workshop hosted at University of Illinois! Recording available here.
Jul 14, 2022 I was invited to give a talk on our work on assessing GANs at the AAPM Annual Meeting tutorial session on Assessment of Deep-Learning Technologies in Medical Imaging! Link to the talk here.
Jun 5, 2022 I was invited to give a talk and lead a session at the Gordon Research Seminar on Imaging Science for early career researchers.
Feb 1, 2022 Our paper on quantifying hallucinations in image reconstruction was covered by several leading newsletters including IEEE Spectrum and within a Forbes article on AI bias. Read the paper here.
Mar 22, 2021 I was awarded the Oak Ridge Institute and FDA fellowship to evaluate GANs for medical imaging applications. Read more here.

selected publications

  1. IEEE TMI
    Assessing the ability of generative adversarial networks to learn canonical medical image statistics
    Varun A Kelkar, Dimitrios S Gotsis, Frank J Brooks, Prabhat KC, Kyle J Myers, Rongping Zeng, and Mark A Anastasio
    IEEE transactions on medical imaging 2023
    PDF Video
  2. ICML
    Prior Image-Constrained Reconstruction using Style-Based Generative Models
    Varun A. Kelkar, and Mark A. Anastasio
    In Proceedings of the 38th International Conference on Machine Learning (ICML) 2021
    PDF Video Code
  3. IEEE TMI
    On hallucinations in tomographic image reconstruction
    Sayantan Bhadra*, Varun A. Kelkar*, Frank J. Brooks, and Mark A. Anastasio
    IEEE Transactions on Medical Imaging 2021
    PDF Code
  4. SPIE JMI
    Impact of deep learning-based image super-resolution on binary signal detection
    Xiaohui Zhang*, Varun A Kelkar*, Jason Granstedt, Hua Li, and Mark A Anastasio
    Journal of Medical Imaging 2021
    PDF
  5. IEEE TCI
    Compressible latent-space invertible networks for generative model-constrained image reconstruction
    Varun A Kelkar, Sayantan Bhadra, and Mark A Anastasio
    IEEE Transactions on Computational Imaging 2021
    PDF
  6. Nat. Photon.
    Harmonic optical tomography of nonlinear structures
    Chenfei Hu*, Jeffrey J Field*, Varun Kelkar, Benny Chiang, Keith Wernsing, Kimani C Toussaint, Randy A Bartels, and Gabriel Popescu
    Nature Photonics 2020
    PDF