Skip to content

Y-vision is a real-time robust crowd tracking framework

Notifications You must be signed in to change notification settings

jnblanchet/Y-Vision

Repository files navigation

Y-Vision

Y-Vision is a real-time robust crowd tracking proof-of-concept framework which combines RGB+D data from multiple kinects. A short demo is available here:

Y-Vision youtube demo

Features

  1. RGB and depth specific features extraction, detection and fusion.
  2. HOG features and blobs from connect component labeling are used as features.
  3. Bayesian model for pedestrian detection.
  4. Tracking using fast branch and bound association.
  5. Ground tracking using simple plane estimation (requires calibration)
  6. TCP interface with simple client API.
  7. Data set collection tool
  8. Efficient Kinect data capture

Packages

  1. Y-Vision is a C# pipeline for pedestrian detection and tracking.
  2. CollectionTool allows for massive data collection in the wild using a Kinect , useful for machine learning.
  3. Y-TcpServer is a TCP server that streams info on the detected pedestrians in real time.
  4. Y-TcpClient and Y-API are self-contained libraries allowing easy integration in other applications (e.g. unity)
  5. Y-TestApi2d listens and displays pedestrian detection events.
  6. Y-CalibrationBoard is a GUI calibration environment to help setup the Kinect rig.
  7. Y-Emulator is a development tool that simulates pedestrians for when no Kinect are available.
  8. Y-MachineLearning contains the Bayesian model for the pedestrian detection
  9. Y-UnitTests is the test package.
  10. Y-Visualization is a set of helpers to display the results and debug.

License

Y-Vision is licensed under the GNU 3.0 GENERAL PUBLIC license.

About

Y-vision is a real-time robust crowd tracking framework

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages