The Recognition of Visual Evoked Potential Based on Wavelet Transformation and BP Neural Network

The brain evoked potentials (BEP) are related directly to series of diseases and physical states. It is helpful to prevent and diagnose the brain diseases by recognizing the evoked potential. In this paper, we use the traditional average method and the wavelet transformation technology to wipe off the noise and extract the feature for the instantaneous visual evoked potential (VEP) at first. Then we recognize the feature with BP neural network so as to produce the control signal for the brain computer interface (BCI). Experiments show that the wavelet transformation can remove the noise and extract the feature effectively and the BP neural network can recognize the VEP comparatively well and truly. The averaged recognition accuracy achieves 85% which makes a base for the analysis and disposal of the evoked potential.