THE PSYCHOLOGICAL STUDY OF BLINK ACTIVITIES USING AN AUTOMATED VIDEO IMAGE ANALYZING SYSTEM

本研究的目的是开发一个自动视频图像分析系统来检测眨眼，并检验该系统提取的眨眼参数。这个系统由三个阶段组成。首先，用摄像机拍摄彗星图像，并将其数字化为黑白。其次，每隔1/60秒将上眼睑运动触发的黑色部分位移输入计算机。第三，在数据去噪后，根据以下4个标准进行眨眼检测，即1)眨眼持续时间和幅度在当前值之间，2)眨眼幅度的上眼睑重开率在预设值以上，3)眨眼幅度的上眼睑持续时间在预设值以下，4)上眼睑运动加速度在预设值以下。大多数眨眼的检测方法与观察眼睛的方法相同，因为识别的标准可以通过参数设置。此外，该测量方法与EOG方法非常相似，表明了该测量方法的有效性。这个自动化系统使我们能够相当容易地测量眨眼开始的时间、眨眼波形和眨眼持续时间。该系统可轻松提取眨眼频率、眨眼时间分布、眨眼延时、眨眼间隔、眨眼持续时间和眨眼波形等眨眼参数。新的自动瞬变分析系统将产生更多的瞬变参数，因此这类瞬变研究将得到进一步推广。

The purpose of this research was to develop an automated video image analyzing system to detect eye blinks and to exemine what eye blink paramenters were extracted by this system.This system consists of three phases. First, comea images were taken by video camera and were digitized into black and white. Second, displacement of the black part triggered by upper eyelid movement was input into the computer at intervals of 1/60 second. Third, after noise was removed from the data, eye-blinks were detected on the basis of following 4 criteria, that is, 1) blink duration and amplitude was between present values, 2) the upper eyelid reopening rate by blink amplitude was over the preset rate, 3) the rate of blink duration by amplitude was under the preset values, 4) the moving acceleration of upper eyelid was under the preset values. Most blinks were detected in the same way as in observing the eye, since the criteria of the identification could be set with the parameters. Moreover, this measurement closely resembled that of the EOG method, which suggests the validity of this measure. This automated system enabled us to measure the time of blink onset, blink wave-form, and blink duration with considerable ease.Blink parameters, such as blink rates, temporal distribution of blinks, blink latencies, inter blink intervals, blink duration, and blink wave-forms were extracted from this system at ease. The new automated blink analyzing system will produce further blink parameters, so blink researches of this type will be spread furthermore.