ucrscholar / pykinecttools Goto Github PK

pyKinectTools
Author: Colin Lea

These instructions are sparse and likely to change. If you would like to use this code the best thing is to email me ([email protected]). I am happy to help

---Dependencies---
numpy, scipy, scikit-learn, scikit-image, visvis, opencv

---Installation instructions---

** Python data libraries:
If you use a python virtual environment:
pip install numpy scipy matplotlib ipython scikit-learn scikit-image pyside visvis

** OpenCV
(Only used for Optical Flow in Histogram of Oriented Optical Flow -- I'm trying to get rid of this dependence!)
If on OSX/Ubuntu(12.04+)/Windows see: https://docs.google.com/document/d/1TsPRI1g_iXQmsCs1VPkLm-9M0T1n724H-wAo7QSNpMY/edit
Otherwise build from source

---Algorithms---
Background subtraction: Adaptive Mixture of Gaussian, Static median/mean model
Feature extraction: Histogram of Oriented Optical Flow, Cuboids, Geodesic Extrema
Simple person tracking
Basic graph Algorithms
Belief Propagation (tree-based)
Iterative Closest Point
Laplacian Eigenmaps manifolds
PCA-based gesture recognition
Superpixels wrapper [deprecated -- use the one in skimage instead]
Others...

Recording with Kinect:
Data capture program
Video/Skeleton data player (w/ user controller)

Depth image utilities (e.g. converting from depth->pointcloud)
chalearn annotated dataset reader