IAS-Lab People Tracking dataset


The IAS-Lab People Tracking dataset is a dataset targeted to people detection and tracking applications and composed of RGB-D video sequences collected in an indoor environment with a Pioneer P3AT robot equipped with a Microsoft Kinect sensor. RGB images are recorded at 640 x 480 pixel resolution, while depth data are saved at 160 x 120 pixel resolution. Both streams have been saved at 30 Hz and the robot odometry is also available. Sequences have been collected both while keeping our platform as static and while moving it on a predefined path. For both cases, we acquired videos in three different scenarios of increasing difficulty:
- no obstacle is present, people move with simple (linear) trajectories;
- no obstacle is present, people move with complex trajectories and interact with each other;
- obstacles are present, people move with complex trajectories and interact with each other.
Every video sequence extends over about 750 frames, thus the total dataset includes 4671 frames, 12272 instances of people and 26 tracks that have been manually annotated on the RGB image and that constitute the ground truth.
The minimum distance between people is 0.2 m while the minimum people-object distance is 0.05 m.



To download the dataset as ROS bags containing synchronized RGB-D stream and robot pose, click here.



If you use this dataset, please cite the following works:

F. Basso, M. Munaro, S. Michieletto and E. Menegatti. "Fast and robust multi-people tracking from RGB-D data for a mobile robot". In Proceedings of the 12th Intelligent Autonomous Systems (IAS) Conference, Jeju Island (Korea). 2012. pp. 265-276.

M. Munaro, F. Basso, and E. Menegatti. "People tracking within groups with RGB-D data". In Proceedings of the International Conference on Intelligent Robots and Systems (IROS), Algarve (Portugal), pp. 2101-2107, 2012.

M. Munaro and E. Menegatti. Fast RGB-D people tracking for service robots. Journal on Autonomous Robots, Springer, vol. 37, no. 3, pp. 227-242, ISSN: 0929-5593, doi: 10.1007/s10514-014-9385-0, 2014.


Ground truth

A file with the suffix '_gt2D' is written for every video. This file contains a row for every rgb image reporting all the annotated persons with the following syntax:
timestamp: [bbox1], [bbox2], ..., [bboxN]


timestamp: timestamp of the rgb image
[bbox...] = [id x y width height]
id: track ID
x, y: image coordinates of the top-left corner of the person bounding box
width, height: width and height of the person bounding box.


Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. Copyright (c) 2013-2014 Matteo Munaro.