基于BP神经网络的手势识别系统

Gesture recognition system based on BP neural network

给出了采用ADXL335加速度传感器来采集五个手指和手背的加速度三轴信息,并通过ZigBee无线网络传输来提取手势特征量,同时利用BP神经网络算法进行误差分析来实现手势识别的设计方法。最后,通过Matlab验证,结果表明,该系统在测试中识别率较高,系统稳定。

A design of gesture recognition is provided, which use ADXL335 acceleration sensor to collect the triaxial acceleration information of five fingers and the back of the hand, pick up the gesture characteristic quantity through ZigBee wireless network transmission, and use BP neural network algorithm for error analysis. Finally, verification by MATLAB simulation shows that the system has high recognition rate in the test, and it is stable.