About
The principal aim of robot perception is to infer useful information from the environment to enable
precise, reliable, robust and safe robot execution.
Perception is universal, and serves the core to recent 3D vision technologies, e.g., from virtual
reality (VR) and augmented reality (AR) equipments to endoscopic surgical devices.
My work is to develop novel theoretical tools to solve perception in challenging scenarios.
In specific, I have been conducting research on:
-
perception in nonrigid scenes (e.g., in surgical devices),
-
techniques to solve derived optimization instances,
-
geometric methods and cameras (e.g., the rolling-shutter camera).
My research effort has resulted in several first-author publications on top rank journals in
robotics, vision, and artificial intelligence, including
T-RO 2021,
T-PAMI 2022,
IJCV 2022, and
IJRR 2023.
Research interests:
I enjoy fundamental research a lot aiming for permanent results that can live longer than ourselves.
Orthogonally, I appreciate applied research that can improve human well-beings in the forthcoming
years.
I am broadly interested in geometric methods, optimization techniques, applied graph theory, sensor
fusion, and machine learning technologies.
Aside from research, I enjoy traveling, visiting museums, and trying local food and drinks.
I am a big fan of medieval buildings.
I like hiking after daily work.
Publication
The list of publications where I workded as the leading author (as both the first-author and the corresponding author):
Journal
KernelGPA: A globally optimal solution to deformable SLAM in closed-form
Bai, Fang, Kanzhi Wu, Adrien Bartoli
International Journal of Robotics Research (IJRR), 43(4):456–484, 2024
|
The Proxy Step-size Technique for Regularized Optimization on the Sphere Manifold
Bai, Fang, Adrien Bartoli
IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 45(5):6428-6444, 2022
|
Procrustes analysis with deformations: A closed-form solution by eigenvalue decomposition
Bai, Fang, Adrien Bartoli
International Journal of Computer Vision (IJCV), 130(2):567-593, 2022
|
Sparse pose graph optimization in cycle space
Bai, Fang, Teresa Vidal-Calleja, Giorgio Grisetti
IEEE Transactions on Robotics (T-RO), 37(5):1381-1400, 2021
|
Robust incremental SLAM under constrained optimization formulation
Bai, Fang, Teresa Vidal-Calleja, Shoudong Huang
IEEE Robotics and Automation Letters (RA-L), 3(2):1207-1214, 2018
|
Conference
KernelGPA: A deformable SLAM back-end
Bai, Fang, Adrien Bartoli
Robotics: Science and Systems (RSS) XVIII, New York City, USA, 2022
|
Scanline homographies for rolling-shutter plane absolute pose
Bai, Fang*, Agniva Sengupta*, Adrien Bartoli*
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 8993-9002, New Orleans, USA, 2022
|
Change of optimal values: A pre-calculated metric
Bai, Fang
IEEE International Conference on Robotics and Automation (ICRA), pages 8295-8301, Paris, France, 2020
|
Predicting objective function change in pose-graph optimization
Bai, Fang, Teresa Vidal-Calleja, Shoudong Huang, Rong Xiong
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 145-152, Madrid, Spain, 2018
|
Incremental SQP method for constrained optimization formulation in SLAM
Bai, Fang, Shoudong Huang, Teresa Vidal-Calleja, Qingling Zhang
14th International Conference on Control, Automation, Robotics and Vision (ICARCV), pages 1-6, Phuket, Thailand, 2016
|