Indoor Skeleton Tracking Dataset

Description

This database contains skeleton tracking information obtained by the Asus Xtion pro live sensor in combination with OpenNI. It contains 3 different indoor environments: 90 days of tracking data from a living room, 74 days of tracking data within a kitchen as well as 20 days of tracking data within an office environment. The dataset has been used for evaluation in [1].

Use

The dataset is freely available for non-commercial research use. Please also cite our paper [1] when using the dataset for your research.

Technical Details

The dataset contains 3 folders, one for each scenario. The tracking information is provided within csv files (e.g. 1.csv), where one day of tracking information is saved per CSV file. Corresponding ground plane estimation results are provided in textfiles (e.g. 1_groundPlane.txt), containing the estimated ground plane parameters a, b, c and d for the 1st day. The csv files include the following information:

FrameNr, Timestamp, FPS (can be ignored)

Moreover, tracking information (x, y, z, confidence) provided by OpenNI of the following skeleton joints is available:

Center of mass, left and right hip, spine, shoulder center, head, left and right knee, left and right foot, left and right shoulder, left and right elbow, left and right hand

References

[1] Planinc R., Kampel M. “Human Centered Scene Understanding based on 3D Long-Term Tracking Data”, Proceedings of the Iberian Conference on Pattern Recognition and Image Analysis (IbPria), to appear, Santiago de Compostela, Spain, June 2015.