A real-time command system based on hand gesture recognition

Jufeng Yang, Jing Xu, Mingda Li, Dingzhen Zhang, Congchao Wang

Abstract: Based on the idea of natural human-computer interaction, we design and implement a command system based on static hand gesture recognition. Firstly, the system continuously captures images by camera connected to the computer, and then determines whether there exists an available hand area based on the HSV (Hue-Saturation-Value) color model. It takes advantage of Fourier descriptors and K nearest neighbor classification to distinguish four different gestures. Finally, Microsoft Office PowerPoint can be controlled by the command corresponding to the identified gestures. Experiment results show that the system can meet the requirements of practical applications.