Hi! Welcome to Li Guan's home page.
Currently, I am a senior research manager at Meta Reality Lab Research, trying to bridge the gaps between the physical and digital worlds.
I received my Ph.D. degree in May 2010. My dissertation title is "Multi-view Dynamic Scene Modeling". My advisor was Prof. Marc Pollefeys. Since then I have worked in GE Global Research, Amazon, Zillow, Wormpex AI Research on different computer vision and machine learning projects, such as Amazon Go, Zillow 3D Home, and Wormpex Scene Reconstruction Platform.
My research interests include 3D reconstruction, camera network calibration, robotic SLAM, multi-sensory multi-modality fusion, optimization, video segmentation, human-computer interaction, machine learning (including deep learning) and GPU acceleration.
2026/02/03 We are hosting a multi-modal spatial sensing workshop SoS, in conjunction with CVPR this year, in Denver CO, USA.
2025/12/30 Our research paper AirGlove: Exploring Egocentric 3D Hand Tracking and Appearance Generalization For Sensing Gloves, has been accepted by ICASSP 2026. We propose an adversarial loss to penalize appearance difference when training new glove pose tracking models, so that we can leverage already collected large datasets about bare-hands or previous glove designs to reduce the amount of data collection and speed up the model realization.
2025/11/30 Our research paper Real-Time Hand Pose Tracking Using 6-axis IMUs, has been accepted by HRI 2026. Our framework aims to track hands reliably without magnetometers, so that the solution can work in varying magnetic field. The performance is achieved by using a new ML + EKF framework to correct the notorious gyro and accelerometer drift issues.
2024/12/11 Our research paper Glissando-Net: Deep Single vIew Category Level Pose eStimation ANd 3D Reconstruction has been accepted by IEEE PAMI, to simultaneously estimate the pose and reconstruct the 3D shape of objects at the category level from a single RGB image, with the capability of handling partial occlusions.
2024/06/18 5th OmniCV workshop is going to happen this summer alongside CVPR, in Seattle USA.
2023/07/13 Our research paper "Implicit Autoencoder for Point-Cloud Self-Supervised Representation Learning", co-authored with Siming Yan, Zhenpei Yang, Haoxiang Li, Chen Song, Hao Kang, Gang Hua, and Xiqing Huang got accepted to ICCV 2023 in Paris France.
2023/02/15 4th OmniCV workshop is going to happen this summer alongside CVPR, in Vancouver Canada. We have a list of amazing speakers lined up, and three challenges you can try out. Hope to see you there!
2022/06/25 I am going to join Meta as a research manager working on sensor fusion and tracking in July 2022. Super excited!
2022/06/20 3rd OmniCV workshop was successfully held in conjunction with CVPR 2022 at New Orleans, LA, USA. Over 150 As attendees physically and online. A huge success! Thanks to my co-organizers Varun Ravi Kumar Senthil Yogamani Shubhankar Borse Marc Eder Jan-Michael Frahm Pierre Moulen Jonathan Horgan Ciarán Eising Stefan Milz. See you next year!