KinectFusion Capture Tool and Example Datasets

Introduction

The paper "When Can We Use KinectFusion for Ground Truth Acquisition?" has been published in the

Workshop on Color-Depth Camera Fusion in Robotics (https://sites.google.com/site/iros2012ws/)

at the International Conference on Intelligent Robots and Systems 2012 (http://www.iros2012.org).

Participating Groups

Data

See below to download the object meshes and data used in the paper.

The used capture tool and KinectFusion implementation is a pre-production version of the tools which are now integrated into the official Microsoft Kinect SDK, available here: http://www.microsoft.com/en-us/kinectforwindows/develop/developer-downloads.aspx

Update Jan 2014:

Thanks to Mohamed Chafik Bakkay of the Research Team on Intelligent Systems in Imaging and Artifcial Vision (SIIVA), Laboratory RIADI in Tunisia
the Kinect Raw Data is now also available in the OpenNI .oni file format.

Acknowledgements

The IROS paper was authored by the following people:
Stephan Meister (HCI, University of Heidelberg)
Pushmeet Kohli (Microsoft Research)
Shahram Izadi\inst (Microsoft Research)
Martin Hämmerle (LiDAR Research Group - GIScience; Institute of Geography,University of Heidelberg)
Carsten Rother (Microsoft Research)
Daniel Kondermann (HCI, University of Heidelberg)

This work has been partially funded by the Intel Visual Computing Institute, Saarbrücken (IVCI) and is part of the Project "Algorithms for Low Cost Depth Imaging".
We thank Prof. Susanne Krömker, Anja Schäfer and Julia Freudenreich of the Visualization and Numerical Geometry Group (Interdisciplinary Center for Scientific Computing, University of Heidelberg) and the Heidelberg Graduate School of Mathematical and Computational Methods for the Sciences (HGS MathComp) We also thank Markus Forbriger, Larissa Müller and Fabian Schütt of the LiDAR Research Group (University of Heidelberg) for their support. Additional thanks to Mohamed Chafik Bakkay of the Research Team on Intelligent Systems in Imaging and Artifcial Vision (SIIVA) - Laboratory RIADI, Tunisia for converting our data

Dataset

Data and links for the IROS 2012 paper "When Can We Use KinectFusion for Ground Truth Acquisition?" This data includes the paper, supplemental material as well as scans and mesh data for the objects described in the paper. The Kinect capture tool as well as KinectFusion are now integrated into the official Microsoft Kinect SDK, available here: http://www.microsoft.com/en-us/kinectforwindows/ Evaluation data and meshes used for the comparisons can be found in IROS2012_data.zip The raw Kinect and Scanner data in contained in HCI_IROS2012_RawKinectData_*.zip (Split up into 2 files due to download size limitations) More elaborate documentation of the data formats will follow shortly. Binary compatibility with the latest version of the SDK has not been established yet. Files in the OpenNI (http://www.openni.org/) .oni format are also available. Please fill out the form below. A download link will be sent to your e-mail. Thanks!

top