cnn-hand-gesture-interface
cnn-hand-gesture-interface copied to clipboard
Hand gesture interface for Desktop PC and Raspberry Pi.
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