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
GNU General Public License, version 3 (GPLv3).
By: Dennis Núñez-Fernández
Website: http://dennishnf.com