Direkt zum Inhalt springen
Computer Vision Group
TUM Department of Informatics
Technical University of Munich

Technical University of Munich



Deformable 3D Shape Matching

This dataset consists of a collection of 3D shapes undergoing nearly-isometric and within-class deformations. In particular, we provide two shape classes ("kid" and "fat kid") under different poses, where the same poses are applied to both classes. Filenames within each class are ordered by deviation from isometry, i.e. the last shape has the largest deformation with respect to the null pose. All shapes in the dataset are given in OFF format, have around 60k vertices and consistent triangulations.

Ground-truth: The shapes have compatible vertex ordering and thus the ground-truth correspondence is the identity. We are also including a text file which provides the symmetric ground-truth correspondence, that is, giving for each vertex its symmetric match on the same shape (self-isometry).

Download dataset

If you use the dataset, please cite the following paper:

Dense Non-Rigid Shape Correspondence Using Random Forests
E. Rodola, S. Rota Bulo, T. Windheuser, M. Vestner, and D. Cremers
In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2014


  • Several shapes have self-intersections, but in fact no topological changes (i.e. no merging of the triangles) of the mesh graph take place. There exists an extended version of this dataset with topological merges.
  • All shapes were created with DAZ Studio.

Rechte Seite

Informatik IX
Chair of Computer Vision & Artificial Intelligence

Boltzmannstrasse 3
85748 Garching info@vision.in.tum.de

Follow us on:
CVG Group DVL Group



Rick Szeliski (University of Washington) will give a talk in the TUM AI lecture series on Jan 28th, 5pm! Livestream


Frank Dellaert (Georgia Tech) will give a talk in the TUM AI lecture series on Dec 17th, 4pm! Livestream


Jon Barron (Google) will give a talk in the TUM AI lecture series on Oct 22nd, 9pm! Livestream


We have five papers accepted to 3DV 2020!


Our effcient deep network architectures form the AI engine of the project Slow Down COVID-19 at Harvard.