Photo

Chen Liu

chenliu [at] wustl (dot) edu

[CV], [Google Scholar], and [GitHub]

I am a Applied Research Scientist in Facebook Reality Labs Research. I defended my Ph.D. thesis on structured indoor modeling in April, 2019. My Ph.D. advisor at Washington University in St. Louis is Prof. Yasutaka Furukawa. My research interests include 3D vision and scene understanding. I am particularly interested in geometry reasoning using learning techniques.


Publications

Floor-SP: Inverse CAD for Floorplans by Sequential Room-wise Shortest Path
In International Conference on Computer Vision (ICCV), 2019
[Project Page]
PlaneRCNN: 3D Plane Detection and Reconstruction from a Single Image
In Computer Vision and Pattern Recognition (CVPR), 2019
oral presentation
[Arxiv] [Project Page]
FloorNet: A Unified Framework for Floorplan Reconstruction from 3D Scans
Chen Liu*, Jiaye Wu*, Yasutaka Furukawa
(* indicates equal contribution.)
In European Conference on Computer Vision (ECCV), 2018
[Arxiv] [Project Page] [Code and Data] [Supp.]
PlaneNet: Piece-wise Planar Reconstruction from a Single RGB Image
In Computer Vision and Pattern Recognition (CVPR), 2018
spotlignt presentation
[Arxiv] [PDF] [Project Page] [Supp.] [Code and Data]
Raster-to-Vector: Revisiting Floorplan Transformation
In International Conference on Computer Vision (ICCV), 2017
[Paper] [Project Page] [Supp.] [Code and Data]
Deep Multi-Modal Image Correspondence Learning
arXiv:1612.01225, 2016
[Arxiv] [Project Page] [Supp.]
Multi-way Particle Swarm Fusion
Chen Liu*, Hang Yan*, Pushmeet Kohli, Yasutaka Furukawa
(* indicates equal contribution.)
arXiv:1612.01234, 2016
[Arxiv] [Project Page]
Layered Scene Decomposition via the Occlusion-CRF
In Computer Vision and Pattern Recognition (CVPR), 2016
spotlignt presentation
[PDF] [Project Page] [Code] [Supp.]

Talks

PlaneNet: Piece-wise Planar Reconstruction from a Single RGB Image
June 2018, CVPR spotlight presentation, video (1:05:30-1:09:30)

Layered Scene Decomposition via the Occlusion-CRF
June 2016, CVPR spotlight presentation, video

Layered Scene Decomposition via the Occlusion-CRF
April 2016, Midwest Vision Workshop talk

Reviewer

CVPR'(18,19), ECCV'18, 3DV'(17,18), ACCV'18

Other projects

MASC: Multi-scale Affinity with Sparse Convolution for 3D Instance Segmentation
A byproduct of another project, which happened to be the state-of-the-art method for 3D instance segmentation evaluated on the ScanNet benchmark (only for a few days).
[Paper] [Code]
Image-based Paper Pop-up Design
A fun project I worked on during a summer exchange program in lovely Taiwan.
SIGGRAPH poster, 2014
[Paper]

Abount Me

I ranked 1st in 5,000m race in the 2014's annual sports meet at USTC.

Jianzi (Freestyle shuttlecock), table tennis, paper craft.