Photo of Yash Garg

Yash Garg

Ph.D. Student, Electrical & Computer Engineering
University of California, Riverside

CV Scholar GitHub LinkedIn

Hi! I am a Ph.D. student in Electrical and Computer Engineering at the University of California, Riverside, co-advised by Prof. Amit K. Roy-Chowdhury and Prof. M. Salman Asif. Before starting my Ph.D. in 2023, I completed an M.S. (research) in ECE at UC Riverside, and a B.Tech. in Electrical and Electronics Engineering at VIT Vellore.

My research sits at the intersection of 3D computer vision and multi-modal learning. I work on 3D human pose and shape estimation under occlusion, scene representations such as Gaussian Splatting and NeRFs, and, more recently, evaluating and improving the physical reasoning abilities of vision-language models. I also have experience with parameter-efficient and low-rank methods for multi-task and multi-domain learning.

I have interned at Mitsubishi Electric Research Laboratories (MERL) and Motional.

  • 3D Human Pose & Shape
  • Gaussian Splatting / NeRF
  • Vision-Language Models
  • Physical Reasoning
  • Multi-Modal Learning
  • PEFT / Low-Rank Methods
  • Unsupervised Domain Adaptation

News

Publications

* denotes equal contribution. See Google Scholar for the full list.

ICCV
2025

VOccl3D: A Video Benchmark Dataset for 3D Human Pose and Shape Estimation under Real Occlusions

Yash Garg, Saketh Bachu, Arindam Dutta, Rohit Lal, Sarosij Bose, Calvin-Khang Ta, M. Salman Asif, Amit K. Roy-Chowdhury

IEEE/CVF International Conference on Computer Vision (ICCV), 2025

Paper Project & Dataset

OJSP
2025

Factorized Tensor Networks for Multi-Task and Multi-Domain Learning

Yash Garg, Nebiyou Yismaw, Rakib Hyder, Ashley Prater-Bennette, M. Salman Asif

IEEE Open Journal of Signal Processing (OJSP), 2025

Paper

WACV
2025

STRIDE: Single-video based Temporally Continuous Occlusion-Robust 3D Pose Estimation

Rohit Lal, Saketh Bachu, Yash Garg, Arindam Dutta, Calvin-Khang Ta, Dripta S. Raychaudhuri, Hannah Dela Cruz, M. Salman Asif, Amit K. Roy-Chowdhury

IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025

Paper arXiv Code

IEEE
TIP

POSTURE: Pose Guided Unsupervised Domain Adaptation for Human Body Part Segmentation

Arindam Dutta, Rohit Lal, Yash Garg, Calvin-Khang Ta, Dripta S. Raychaudhuri, Hannah Dela Cruz, Amit K. Roy-Chowdhury

Under review, IEEE Transactions on Image Processing (TIP)

arXiv

PLoS ONE
2021

eARDS: A Multi-Center Validation of an Interpretable Machine Learning Algorithm of Early Onset Acute Respiratory Distress Syndrome (ARDS) among Critically Ill Adults with COVID-19

Yash Garg*, Lakshya Singhal*, Philip Yang*, Azade Tabaie, A. Ian Wong, …, Rishikesan Kamaleswaran

PLoS ONE, 2021

Paper

Experience

Mitsubishi Electric Research Laboratories (MERL)

Jun 2026 – Sep 2026

Research Intern · Manager: Anoop Cherian

  • Contributed to the design and development of a synthetic, game-engine-simulated dataset for evaluating physical reasoning.
  • Benchmarked state-of-the-art open- and closed-source vision-language models on physical reasoning. Collaboration ongoing, targeting a peer-reviewed publication.

Motional

Jun 2025 – Sep 2025

ML Intern · Manager: Patrick Buehler

  • Built a multi-modal vehicle-light classifier fusing image sequences with vehicle metadata (acceleration, yaw), reducing error by 14.4% over the image-only baseline.
  • Distilled a CLIP-based teacher into a deployment-ready student that is 61.8% faster and 55% smaller while retaining over 97% of its accuracy.

University of California, Riverside

Sep 2021 – Present

Graduate Student Researcher · Prof. Amit K. Roy-Chowdhury & Prof. M. Salman Asif

  • Robust semantic segmentation for 3D Gaussian Splatting that lifts noisy, partial 2D masks into 3D via feature-guided Gaussian expansion (+8.06 mIoU over the best baseline; under review).
  • VOccl3D: a video benchmark for 3D human pose and shape under real occlusions, built with SMPL, Gaussian Splats/NeRFs, Blender and Unity (ICCV 2025).
  • STRIDE: unsupervised, temporally continuous 3D human pose estimation under occlusion (WACV 2025).
  • Unsupervised human tracking, silhouette and pose extraction under occlusion for an IARPA biometric recognition project.
  • Factorized Tensor Networks: a low-rank PEFT architecture for multi-task and multi-domain learning (IEEE OJSP 2025).

Ageye Technologies, Bangalore

Nov 2020 – May 2021

Computer Vision Intern

  • Developed an edge-detection model for overlapping leaves (92% accuracy) and a pipeline computing leaf orientation, count and centers, deployed in indoor farms.

Emory University

Apr 2020 – Nov 2020

Research Intern · Prof. Rishikesan Kamaleswaran

  • Developed an interpretable ML model for early prediction of ARDS in critically ill COVID-19 patients (AUROC 0.89), resulting in a co-first-author PLoS ONE paper.

Education

Ph.D., Electrical and Computer Engineering

Sep 2023 – Present

University of California, Riverside · Advisors: Prof. Amit K. Roy-Chowdhury, Prof. M. Salman Asif

M.S. (Research), Electrical and Computer Engineering

Sep 2021 – Aug 2023

University of California, Riverside · Advisor: Prof. M. Salman Asif

B.Tech., Electrical and Electronics Engineering

Aug 2016 – Jun 2020

Vellore Institute of Technology (VIT), Vellore

Service & Awards

Reviewer

  • CVPR 2024, 2025, 2026
  • ICCV 2025
  • ECCV 2026
  • ACCV 2026
  • 3DV 2027
  • ICIP 2025, 2026
  • IEEE Transactions on Computational Imaging (TCI) 2025
  • IEEE Open Journal of Signal Processing (OJSP)

Awards

  • Dean's Distinguished Fellowship, UC Riverside
  • GVSDP Scholarship, VIT Vellore
  • $10,000 Microsoft Azure research credit grant for pandemic research