中国大学生虚拟仿真动态表情库的建立与评价

目的:建立一套标准化的中国大学生虚拟仿真动态表情库。方法:采集筛选40名大学生的6种不同动态表情共240段,对其进行标准化加工处理。在北京地区4所高校招募23名大学生,对动态表情的类型、愉悦度和唤醒度进行内部一致性信度、评分者信度检验。结果:建立了包含愤怒9段、厌恶1段、害怕1段、高兴38段、悲伤19段、惊讶39段的表情库,其动态表情的认同率均大于60%。愉悦度与唤醒度的Cronbach α系数分别为0.96、0.76;评分者信度显示被试在愉悦度与唤醒度(W=0.54、0.21,均P<0.001)上具有一致的评价。结论:本研究建立的中国大学生虚拟仿真动态表情库有较好的信度,可作为虚拟现实及情绪研究的实验材料。

Establishment and evaluation of the Virtual Simulation Dynamic Facial Expression Database of Chinese University Students

Objective:To establish and evaluate the Virtual Simulation Dynamic Facial Expression Database of Chinese University Students.Methods:Six kinds of dynamic expressions from forty Chinese university students,totaling 240 segments,were recruited and screened.The standardized processing was carried out on these expressions.Twenty-three subjects were recruited from 4 universities in Beijing and were assessed for the category,pleasure and arousal of these dynamic expressions.Based on participants’ assessments,a consistency reliability,scorer reliability,repeated measurements,and independent sample t-test were analyzed.Results:The established dynamic facial expression database included 9 angry expressions,1 disgust expression,1 fear expression,38 happy expressions,19 sadness expressions and 39 surprise expressions.The identification rates of dynamic expressions were above 60%.The internal consistency coefficients of both pleasure and arousal were 0.96 and 0.76.The scorer reliability showed that the subjects had a consistent evaluation on the degree of pleasure(W=0.54,P<0.001) and the degree of arousal(W=0.21,P<0.001).Conclusion:It suggests that the Virtual Simulation Dynamic Facial Expression Database of Chinese University Students has a good reliability,which could be used in facial expression-related studies in future.