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

Technical University of Munich



Intrinsic3D Dataset

Intrinsic3D: High-Quality 3D Reconstruction by Joint Appearance and Geometry Optimization with Spatially-Varying Lighting

Robert Maier1,2 Kihwan Kim1 Daniel Cremers2 Jan Kautz1 Matthias Nießner2,3
1NVIDIA 2Technical University of Munich 3Stanford University

IEEE International Conference on Computer Vision (ICCV) 2017

We introduce a novel method to obtain high-quality 3D reconstructions from consumer RGB-D sensors. Our core idea is to simultaneously optimize for geometry encoded in a signed distance field (SDF), textures from automatically-selected keyframes, and their camera poses along with material and scene lighting. To this end, we propose a joint surface reconstruction approach that is based on Shape-from-Shading (SfS) techniques and utilizes the estimation of spatially-varying spherical harmonics (SVSH) from subvolumes of the reconstructed scene. Through extensive examples and evaluations, we demonstrate that our method dramatically increases the level of detail in the reconstructed scene geometry and contributes highly to consistent surface texture recovery.

Intrinsic3D Dataset

The following dataset contains RGB-D sequences and reconstructed 3D models of five different scenes. We captured the RGB-D data using a Structure.io depth sensor (640x480px) and an iPad color camera (1296x968px). Please refer to the respective publication when using this data.


For each scene of the Intrinsic3D dataset, we provide the respective RGB-D sequence as well as the fused and refined 3D models (PLY binary format). Each RGB-D sequence contains:

  • Color frames (frame-XXXXXX.color.png): RGB, 24-bit, PNG
  • Depth frames (frame-XXXXXX.depth.png): depth (mm), 16-bit, PNG (invalid depth is set to 0)
  • Camera poses (frame-XXXXXX.pose.txt): camera-to-world pose computed by VoxelHashing
  • Camera calibration (colorIntrinsics.txt and depthIntrinsics.txt): color and depth camera intrinsics (default factory calibration).

For reference, the RGB-D data is provided in the same format as specified here. The Intrinsic3D source code on github can directly read and process sequences in this format.




Tomb Statuary



Unless stated otherwise, all data in the Intrinsic3D Dataset is licensed under a Creative Commons 4.0 Attribution License (CC BY 4.0).

Export as PDF, XML, TEX or BIB

Conference and Workshop Papers
[]Intrinsic3D: High-Quality 3D Reconstruction by Joint Appearance and Geometry Optimization with Spatially-Varying Lighting (R. Maier, K. Kim, D. Cremers, J. Kautz and M. Niessner), In International Conference on Computer Vision (ICCV), 2017. ([slides] [poster] [dataset] [code]) [bibtex] [pdf]
Powered by bibtexbrowser
Export as PDF, XML, TEX or BIB

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



TANDEM (CoRL 2021) received the Best Demo Award at 3DV 2021! Congrats to Lukas Koestler, Nan Yang, and Niclas Zeller!


Bernt Schiele (Max Planck Institute for Informatics) will give a talk in the TUM AI lecture series on June 10th, 3pm! Livestream

French-German Machine Learning Symposium

French-German Machine Learning Symposium

The French-German Machine Learning Symposium brought together some of the leading ML researchers from France and Germany to give us a glimpse into their research and engage in discussions on the future of machine learning. The list of speakers includes Yann LeCun, Cordelia Schmid, Jean-Bernard Lasserre, Bernhard Schölkopf, and many more! For the full program please visit the webpage.


Ron Kimmel (Technion - Israel Institute of Technology) will give a talk in the TUM AI lecture series on May 6th, 3pm! Livestream