Varun Jampani

About

Varun Jampani is a leading researcher in 3D computer vision, graphics, and robotics, with a focus on point cloud registration, pose estimation, and probabilistic modeling. His most influential work, **DeepGMR** (2020), with 247 citations, revolutionizes point cloud registration by learning latent Gaussian Mixture Models, enabling robust alignment under large transformations, noise, and time constraints—a critical advance for autonomous systems and 3D reconstruction. Jampani also introduced **Implicit-PDF** (2021), a non-parametric representation for probability distributions on the rotation manifold, tackling single-image pose estimation with uncertainty quantification and handling symmetric objects with multiple correct poses. This work addresses long-standing challenges in vision and robotics. His contributions bridge deep learning and geometric reasoning, offering practical solutions for real-world applications. With over 250 total citations, Jampani’s research is widely recognized for its impact on 3D perception, and his innovative approaches continue to inspire advancements in registration and pose estimation, making him a key figure in the field.

Research Focus

Key Achievements

Top Papers

1. 1
   DeepGMR: Learning Latent Gaussian Mixture Models for Registration

   247 citations · 2020
2. 2
   DeepGMR: Learning Latent Gaussian Mixture Models for Registration

   16 citations · 2020
3. 3
   Implicit-PDF: Non-Parametric Representation of Probability Distributions on the Rotation Manifold

   7 citations · 2021

Key Collaborators

Contact & Links