Kinect RGBD Dataset

Kinect RGBD Dataset for Category Modeling

This RGBD category dataset consists of RGBD images containing about 900 objects. These RGBD images are collected in different environments, indoors and outdoors. There are seven large categories, such as basket, bucket, bicycle, scanner, fridge, notebook PC, sprayer, dustpan, and platform lorry. Each category has a large number of objects. The RGBD image is in the size of 640×480. Objects inside a category usually have different textures, and they are placed in complex environments with different translations and rotations. Moreover, objects within some category are of large difference in their local structures and sizes. For example, bicycles for men have beams, while those for women do not. Small bicycles are usually with simpler structures, and compared to other parts, the wheel radius changes most in size among different bicycles.

This RGBD object dataset is only able to be used for non-commercial purpose (research or education). If you use it in your research, please cite the following paper and acknowledge it in your technical publications:

Category Modeling from just a Single Labeling: Use Depth Information to Guide the Learning of 2D Models
Quanshi Zhang, Xuan Song, Xiaowei Shao, Huijing Zhao, and Ryosuke Shibasaki
Proc. of IEEE International Conference on Computer Vision and Pattern Rcognition (CVPR), 2013.

Please visit to download the dataset, and visit to download the attributed relational graphs and graph templates generated from these images.

We also share the labeling of initial graph templates used in “Q. Zhang, X. Song, X. Shao, H. Zhao, R. Shibasaki, Learning Graph Matching for Category Modeling from Cluttered Scenes, ICCV 2013”. If you are interested in these graph templates, please contact Quanshi Zhang (Website)
 If you have any further problem or question about it, please feel free to contact Quanshi Zhang (Website) via the following email: zqs1022/at/


The following image shows some selected examples in this category dataset.