kaikai707 / gesturerecognition

Kinect 手势识别。Gesture recognition with Microsoft Kinect. (Bachelor Dissertation Project)

Gesture Recognition based on Kinect

Just for GRADUATION.

Aims

• Skin detection. ✔︎
• Obtain one hand's contour. ✔︎
• Distinguish hand from head. ✔︎
• Mark finger tips and palm's positions. ✔︎
• Use K-Means algorithm to detect finger tips' positions. ✖︎
• Use bounding polygon to detect finger tips' positions. ✔︎
• Remove points on arm. ✖︎
• Follow finger tips and palm's positions change.✔︎
• Use finite-state machine to define and recognize hand gesture.✖︎
• Improve precision of finger tips and palm's positions.✖︎
• Be able to follow finger tips's movement.✖︎
• Improve performance, avaliability and usability.✖︎

Requirements

• Unix/Linux, Microsoft Windows®
• OpenNI 1.x, OpenNI 2.x (Currently OpenNI 2.x only support Windows® Kinect SDK)
• OpenCV >= 3.0.0
• Microsoft® Kinect
• Python2

Build, Install and Run

$ virtualenv .venv
$ source .venv/bin/activate
$ ./setup.py build
$ ./setup.py install
$ gestrecog

ScreenShots

Thanks

• Directed by Associate Prof. Qingge Ji
• Open Source Computer Vision Library

License

BSD 3-Clause License