News

Tue, 03/24/2020

Check out our #CVPR2020 paper on Certifiably Optimal 3D Shape Reconstruction from 2D Landmarks in a single image. Camera-ready version here:

https://arxiv.org/abs/1911.11924.

 

Fri, 01/24/2020

Super excited to share our new work: “TEASER: Fast and Certifiable Point Cloud Registration”

Paper: https://arxiv.org/pdf/2001.07715.pdf

Code: https://github....

Tue, 01/21/2020

Our paper "Graduated Non-Convexity for Robust Spatial Perception: From Non-Minimal Solvers to Global Outlier Rejection" has been accepted to publish at the IEEE Robotics and Automation Letters (RA-L)! The final...

Thu, 11/28/2019

In our new arXiv preprint: In Perfect Shape: Certifiably Optimal 3D Shape Reconstruction from 2D Landmarks, we present Shape*, the first certifiably optimal non-minimal...

Wed, 10/16/2019

I am glad to release "Hands-on Tutorial on Global Optimization in Matlab", which demonstrates the power of semidefinite relaxations in globally solving non-convex computer vision problems that used to be known as NP-hard. This tutorial uses rotation averaging and pose graph optimization as an...

Mon, 07/22/2019

Our arXiv paper "A Quaternion-based Certifiably Optimal Solution to the Wahba Problem with Outliers" is accpetd to ICCV 2019 (International Conference on Computer Vision) for an oral...

Thu, 06/20/2019

TEASER (Truncated least squares Estimation And SEmidefinite Relaxation), our polynomial-time solution for robust registration with extreme outlier rates, is featured in ...