Datasets

RGB-D Object Dataset

The RGB-D Object Dataset is a large dataset of 300 common household objects recorded using a Kinect style 3D camera.

This dataset was introduced in:

A Large-Scale Hierarchical Multi-View RGB-D Object Dataset. K. Lai, L. Bo, X. Ren, and D. Fox. ICRA, 2011.

Please cite the above publication if you use this dataset.

Visit Dataset Website

Urban Scenes 3D Point Cloud Dataset

The dataset consists of two parts:

1) 3D Point Clouds of 10 labeled real-world urban scenes: Velodyne Data

2) 3D Point Clouds of 274 Google SketchUp models of objects commonly found in urban scenes: Google SketchUp Data

This dataset was used in evaluating the object recognition technique described in:

Object Recognition in 3D Point Clouds Using Web Data and Domain Adaptation. K. Lai and D. Fox. IJRR, 2010.

3D Laser Scan Classification Using Web Data and Domain Adaptation. K. Lai and D. Fox. RSS-09.

Please cite the above publications if you use this dataset.

Acknowledgements

We thank Albert Huang and the MIT DGC Team for providing us with the urban scenes velodyne data.