Kunal Gupta

I am a PhD candidate in the CSE department at UC San Diego advised by Manmohan Chandraker. I work primarily in 3D computer vision.

In the past, I have twice interned at Adobe w/ Kalyan Sunkavalli, Vladimir Kim and at NVIDIA Research w/ Stan Birchfield . I have also worked as a staff researcher at UC San Diego School of Medicine with Fransisco Contijoch and was a visiting researcher in the Bio Robotics Lab with YU Haoyong at National University of Singapore.

I received my M.S. degree in Computer Science from UC San Diego and B.Eng. degree in Electrical and Electronics Engineering from BITS Pilani, where I was advised by Surekha Bhanot.

My research is supported by the Qualcomm Innovation Fellowship 2023.

Email  /  CV  /  Google Scholar  /  GitHub  /  LinkedIn

profile photo
Note: If you wish to ask me questions concerning a master's or Ph.D. in CSE at UCSD, please see the FAQs before contacting me. Feel free to send me a reminder if you don't hear from me within 7-10 days.
Research

I'm interested in 3D Computer Vision, particularly in inverse rendering, shape deformation, and representation learning with applications to autonomous driving, robotics, augmented reality, and medical imaging. Lately, I have been exploring the applications of LLMs in inverse rendering. Read an overview of my research here.

MCNeRF: Monte Carlo Rendering and Denoising for Real-Time NeRFs
Kunal Gupta, Miloš Hašan, Zexiang Xu, Fujun Luan, Kalyan Sunkavalli, Xin Sun, Manmohan Chandraker, Sai Bi
ACM SIGGRAPH ASIA 2023

paper / project page / arXiv / code

A general Monte Carlo-based method to accelerate the rendering of any NeRF representation.

Neural Jacobian Fields: Learning Intrinsic Mappings of Arbitrary Meshes
Noam Aigerman, Kunal Gupta, Vladimir Kim, Siddhartha Chaudhuri, Jun Saito, Thibault Groueix
ACM SIGGRAPH 2022   (journal track)

Paper / code

A framework for learning to deform meshes in a highly detail-preserving manner, without being tied to a specific mesh

DiFiR-CT: Distance Field Representation resolves motion during Computed Tomography
Kunal Gupta, Brendan Colvert Zhennong Chen and Francisco Contijoch
Medical Physics 2023

paper / project page / arXiv / code

Motion resolved computed tomography by combining neural implicit representations and differentiable rendering

Neural Mesh Flow: 3D Manifold Mesh Generation via Diffeomorphic Flows
Kunal Gupta, Manmohan Chandraker
NeurIPS 2020   (Spotlight)

project page / arXiv / code / Master's Thesis

A network that generates manifold meshes enabling photo-realistic physically based renderings and physics simulation

Force Control of Robotic Walker
Bachelor's Thesis

Novel Stable Gait Criteria (SGC) is proposed for safe and intuitive human-robot interaction and rehabilitation.

Talks
Random Walks and Sampling at MC Lab Discussions, September 2020
ELBO and KL-Divergence Tutorial at ERL Discussions, December 2018

Template stolen from Jon Barron