Abstract: Focusing on the shortcoming of blink detection based on Electrooculography（EOG）?technique, the article designs a novel Human-Computer Interaction（HCI） model based on continuous eye-blink, which mainly combines IPPG and the signal processing method to realize eye-blinking pulse signal extraction from facial video sequences. Further, the parameters for time/space analysis window and filter are selected and adjusted, for eye-blink pulses detection and related control command generation. On the basis of this, a set of blinking control multimedia player system is established to verify the effectiveness of the proposed algorithm and HCI model. The experimental results show that the stability of this method is better than that of the traditional EOG method in the jamming environment. In the system testing, the accuracy of blink control command is 92.95%, which shows a good performance for practical human-computer interaction.

Key words: facial video, G channel, Electrooculography（EOG）, eye blink signal, human-computer interaction control

摘要： 针对基于眼电图（Electrooculography，EOG）技术的眨眼检测方法存在的不足，设计了一种基于连续眨眼检测的人机交互新模式。主要结合图像光学体积描记技术（Image Photoplethysmography，IPPG）和信号处理方法，实现对视频序列中眨眼脉冲信号的提取，并选择合适的时/空分析窗口和滤波参数，对窗内连续的眨眼信号进行检测和控制命令转换，在此基础上建立了一套眨眼控制多媒体播放系统，以验证所提算法和人-机交互模式的有效性。实验对比显示，该方法在干扰环境下的眨眼检测稳定性明显优于传统EOG方法。系统实测中，眨眼控制命令的准确率达92.95%，体现出良好的人-机交互性能。

关键词: 面部视频, G通道, 眼电图, 眨眼信号, 人-机交互控制