Projective shape analysis (PSA) dataset



The goal of this work is to provide data for performing imperfect 3D shape segmentation by utilizing knowledge gained from images/projection data. To build these datasets, we have collected 11 sets of shapes, which include labeled 2D shapes and imperfect 3D shapes.


All data can be downloaded via the following links.


Category #Parts Labeled 2D Shapes Imperfect 3D shapes
#Labeled photos Photo snapshot #Tested shapes Our segmentions Shape snapshot
Chair 4 910 10 Results
Truck 3 400 5 Results
Vase 4 200 5 Results
Bike 5 181 6 Results
Guitar 3 20 3 Results
Robot 4 174 3 Results
Stroller 6 398 3 Results
Lamp 3 343 6 Results
Table 2 452 3 Results
Pavilion 3 60 4 Results
Fourleg 5 397 4 Results



If you use this dataset in your own research, please cite the following paper:

  • Yunhai Wang, Minglun Gong, Tianhua Wang, Daniel Cohen-Or, Hao Zhang, Baoquan Chen, Projective Analysis for 3D Shape Segmentation, ACM Transactions on Graphics (Proc. SIGGRAPH Asia), vol. 32, n. 6, 2013.

  • Credits

    We would like to thank Tianhua Wang, Bo Hua, Chaoran Fan and the artists in SIAT: Jiacheng Ren and Jinfeng Ou, have contributed to creating the labeled 2D photos; 3D Warehouse, and the users of SketchUp for making these imperfect 3D shapes available; Chao-Hui Shen for some of point cloud data.


    Shape COSEG dataset


    Please send email to cloudseawang [at] if you have any questions.