Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
OpenMoves is a toolbox that processes OpenPTrack person tracking data in real-time to generate high-level features and perform classification and recognition of spatiotemporal patterns of movement.
It aims to provide data about how people move, not just where they are.
OpenMoves is being developed by UCLA REMAP.
The lead developer is Sam Amin.
Please navigate through the pages on the right for further information.
Digital technology for recording and summarizing human movement is increasingly sophisticated and accessible – from commercial motion capture used for computer-generated characters in film and television to inexpensive, surveillance-oriented technologies used to make even home webcams capable of detecting people. A single RGBD (red-green-blue-depth) camera can be used to create interactive games, artworks, and other gesture- and movement-driven applications. The OpenPTrack system is an open source tool for using many such imagers for real-time person tracking over large environments; it is oriented towards arts, culture, and education applications.
Most of these systems, especially in the open source world, focus on capturing movement. With the notable exception of the analysis of GPS traces for location-based applications, supporting the interpretation of such motion data over time is less well-studied, even though many fields stand to benefit from pattern detection and feature calculation for human movement data.
OpenMoves is a new toolset which we are developing for generating higher-level features and performing learning and recognition on human movement trajectories, such as those gathered by OpenPTrack. The current set of features extracted and learning performed fall into four different categories and were chosen with live performances, educational applications, and interactive artworks in mind; however, we feel they are general enough to be applicable to other domains as well. The framework is meant to be modifiable, so as to work with systems other than OpenPTrack, and its output is published to the network for consumption.
If you use OpenMoves, please cite:
- Sam Amin and Jeff Burke. 2018. OpenMoves: A System for Interpreting Person-Tracking Data. In Proceedings of the 5th International Conference on Movement and Computing (MOCO '18). ACM, New York, NY, USA, Article 13, 4 pages.