Stacks Image 0

Kai Ni


I am currently a senior architect at Baidu IDL, and I got my Ph.D. from College of Computing, Georgia Institute of Technology. My research interests include computer vision, robotics, and machine learning. More specifically, I work on SfM/SLAM problems as well as place recognition. Frank Dellaert is my advisor. Check out the Research page, the Publications page, and the Software page to find out more.

Joining Baidu

I joined the Baidu Institute of Deep Learning (IDL) in November 2013, where I will be leading the effort on 3D vision and robotics including large-scale 3D mapping.

idl

3DIMPVT & IROS papers

There are two papers coming: “HyperSfM” in 3DIMPVT 2012 and “Robust Optimization of Factor Graphs by Using Condensed Measurements” in IROS 2012. The details can be viewed here.

grandcanal-pt-tree

Automated SLAM Evaluation Workshop

On July 1st, Michael, Grisetti, and I will organize the workshop on Automated SLAM Evaluation in the 2011 Robotics: Science and Systems Conference at USC. Submissions are open now!

workshop_logo

Thesis Defense

I passed my thesis defense today. Thanks my thesis committee (Dr. Marc Pollefeys, Dr. Eric N. Johnson, Dr. Irfan Essa, Dr. Henrik I. Christensen, and Dr. Frank Dellaert) and my fellow students for the effort and the support! Meanwhile, I have started working at Microsoft as a researcher.


logo images

Visibility Learning, ICRA 2011

Our recent work with Pablo F. Alcantarilla has been accepted in ICRA 2011. The paper is about learning to predict which 3D features from a large environment model will be detected in a camera image, which can cut down tremendously on computational cost in localizing the camera, e.g., in robot or phone localization.

shapeimage_2