Skip to content

Latest commit

 

History

History
30 lines (17 loc) · 945 Bytes

README.md

File metadata and controls

30 lines (17 loc) · 945 Bytes

CNN-based hand gesture interface

Hand gesture interface for Desktop PC and Raspberry Pi. The system is intended to recognize 10 hand gestures. Based on such poses you can control some devices such as a drone, mobile robots, screens among others.

The systems work in real-time. The Convolutional Neural Networks (CNNs) were trained using Caffe framework. OpenCV libraries were used. In order to obtain a better computation performance, the system was implemented using C++ language. The systems are intended to work on Linux systems, however, can be compiled for Windows or other Operating System.

In order to modify the code, you should compile with:

$ make

Finally, execute the created file. Example:

$ ./handgesture_detrack

License

GNU General Public License, version 3 (GPLv3).

About

By: Dennis Núñez-Fernández

Website: http://dennishnf.com