Informatics, TU Vienna

MPII Building Model as Data for Your Research

In this talk I present the project of the virtual reconstruction of Max-Planck-Institute for Informatics (MPII) building in Saarbruecken.


The resulting geometric, reflectance, and illuminaire data can be used for various research activities including computer graphics, computational and algorithmic geometry, computer vision, databases, etc. A significant advantage of this data set is the availability and accessibility of the real building as a reference for other data acquisitions and validations. The resulting data set will be provided for public non-commercial research purposes to facilitate the research advancement of algorithms working with the real data sets. The level of detail of the building model goes up to the furniture installed in the building. The total model complexity is tens of millions of triangles. I describe the effort to create this data set, different problems met during the project, and I show the example images rendered by several algorithms. I also briefly mention the possible algorithmic challenges for this data set.


Vlastimil Havran is an adjunct professor in the department of computer graphics and interaction at the Czech Technical University in Prague. Prior he has joined the university in 2006, he was a post-doc and senior researcher at the Max-Planck-Institute for Informatics. He received a PhD for the dissertation on ray shooting algorithms from the department of computer science in Czech Technical University in Prague in 2001.