Detection Method of Targets in Videos Using NonInvasive Brain-Computer Interface

A novel video-target recognition method is proposed in this paper by directly translating the Electroencephalogram (EEG) signals of an operator when watching the video. In order to explore the neural signatures of the video-target recognition, we use the continuous wavelet transform (CWT) to analyze the EEG signals before and after a video-target recognition. It is found that the amplitude of EEG signals increases significantly in the θ (4 ~ 8Hz) and α (8 ~ 13Hz) bands when the operator recognizes a video target. Then we use the time-frequency features to construct the classifier. The experimental results show that the average area under the curve (AUC) of the classification model can reach 0.8413, which shows that the proposed method has good performance. This method can be used as a supplement to the existing machine intelligence based methods of video-target detection.