CMU-UAH Dataset

CMU:
This data is free for use in research projects.
You may include this data in commercially-sold products, but you may not resell this data directly, even in converted form.
If you publish results obtained using this data, we would appreciate it if you would send the citation to your published paper to jkh+mocap@cs.cmu.edu,
and also would add this text to your acknowledgments section:

The data used in this project was obtained from mocap.cs.cmu.edu.
The database was created with funding from NSF EIA-0196217.

UAH:
This data is free for use in research projects.
You may include this data in commercially-sold products, but you may not resell this data directly, even in converted form.

Description:
Two of the main goals of using this dataset are to train accurate
models with different pedestrian dynamics and to test the
feasibility and limits of the proposed methods in an extensive
way under ideal conditions. For that purpose, this dataset is based on a high frequency
and low noise dataset published by Carnegie Mellon University. On the one side, the high
frequency of the dataset helps the algorithms to properly learn
the dynamics of different activities and increases the probability of finding a similar test observation in the trained data
without missing intermediate observations. On the other side, low noise models improve the prediction when working
with noisy test samples. The dataset contains 3D coordinates of 41 joints located along the pedestrian bodies 
gathered at 120 Hz while people are simulating typical pedestrian activities.

The CMU dataset contains a considerable number of different activities, including human interaction of several subjects.
We pre-selected and filtered the available sequences according
to the following criterion: only sequences including walking, starting, stopping and standing activities without orientation changes of the subjects are selected. This way, a total
of 490 sequences composed of 302,470 pedestrian poses from
31 different subjects were extracted. Hereafter, this set of
sequences will be named as CMU-UAH dataset.
After this extraction, the CMU-UAH dataset was divided
into 8 subsets. The first division was based on the orientation
of typical crossing activities, i.e. left-to-right and right-to-left.
The second one was based on the type of activity, i.e.
walking, starting, stopping and standing. Those sequences
with more than one activity were cropped into subsequences
with only one action. 

It is worth remarking that each pedestrian observation is
composed of pose.

In each sequence the user can find the position of each joints in millimeters, the framerate and the instant of time when the observation were capture with respect to the init of the video.

Event-Labelling Methodology: 
A starting activity is defined as the action that begins when
the pedestrian moves one knee to initiate the gait and ends
when the foot of that leg touches the ground again. A walking
activity is defined as the action that happens after a starting
activity and before a stopping activity. Moreover, a stopping
activity is defined as the action that begins when a foot is
raised for the last step and finishes when that foot treads
the ground. Finally, standing activities are defined as the
actions that happen after stopping activities and before starting
activities. These criteria were adopted because these transitions
are easily labeled by human experts, thus enabling the creation
of reliable ground truths.

Papers:
R. Quintero, J. Almeida, D. F. Llorca, M. A. Sotelo, "Pedestrian Path Prediction using Body Language Traits", IEEE Intelligent Vehicle Symposium, pp. 317-323, 2014.

R. Quintero, I. Parra, D. Llorca, M. Sotelo, "Pedestrian path prediction based on body language and action classification", Intelligent Transportation Systems (ITSC) 2014 IEEE 17th International Conference, pp. 679-684, Oct 2014.

R. Quintero, I. Parra, D. F. Llorca, M. A. Sotelo, "Pedestrian Intention and Pose Prediction through Dynamical Models and Behaviour Classification", Intelligent Transportation Systems (ITSC) 2015 IEEE 18th International Conference on, pp. 83-88, 2015.

R. Quintero, I. Parra, J. Lorenzo, D. Fernndez-Llorca, M. A. Sotelo, "Pedestrian intention recognition by means of a Hidden Markov Model and body language", Intelligent Transportation Systems (ITSC) 2017 IEEE 20th International Conference on, pp. 1-7, 2017.

Mnguez, R. Q., Alonso, I. P., Fernndez-Llorca, D., and Sotelo, M. . Pedestrian path, pose, and intention prediction through gaussian process dynamical models and pedestrian activity recognition. IEEE Transactions on Intelligent Transportation Systems (2018).

