基于面部表情识别的突发事故应激情绪反应差异

doi:10.16265/j.cnki.issn1003-3033.2026.01.0868

摘要/Abstract

摘要：

为探究突发事件下不同类型群体情绪表达特征,深入分析性别、气质类型和事故环境对群体应激状态情绪的影响,基于面部表情技术设计应激状态情绪刺激试验。选取137名被试观看5类事故视频,采用Facereader软件采集6种基本情绪强度数据,通过非参数检验分析变量差异性,并构建基于唤醒度动态特征的k-means聚类模型。结果表明:女性悲伤与恐惧强度显著高于男性,而愤怒强度较低;多血质个体情绪反应最强烈,黏液质个体唤醒度调节速度最快;建筑火灾场景诱发的恐惧强度最显著;男性在应激状态下的唤醒度自我调节能力显著优于女性。

关键词: 面部表情, 突发事故, 应激状态, 情绪反应, k-means聚类

Abstract:

In order to explore the emotional expression characteristics of different types of groups under emergencies, and to deeply analyze the influence of gender, temperament type and accident environment on the stress state emotions of the groups, a stress state emotional stimulation test based on facial expression technology was designed. A cohort of 137 participants was exposed to five categories of accident videos. Facial expression data for six basic emotions were collected using FaceReader software, with the Kruskal-Wallis test employed to analyze differences across gender, temperament types, and accident scenarios. A k-means clustering model was further constructed based on arousal dynamic features. The results show that female participants exhibit significantly higher intensities of sadness and fear, whereas males show stronger anger responses. Sanguine individuals demonstrate the most pronounced emotional reactivity, while phlegmatic types achieve the fastest arousal modulation. Fear responses are most pronounced in building fire scenarios. Males outperform females in arousal self-regulation capacity.

Key words: facial expression recognition, sudden accident, stress state, emotional response, k-means clustering

中图分类号:

X912安全生理学

孟筠青, 符运莲, 高滨, 邱敬渊. 基于面部表情识别的突发事故应激情绪反应差异[J]. 中国安全科学学报, 2026, 36(1): 191-198.

MENG Junqing, FU Yunlian, GAO Bin, QIU Jingyuan. Differences in stress-induced emotional responses to sudden accidents based on facial expression recognition[J]. China Safety Science Journal, 2026, 36(1): 191-198.

图/表 15

图1

表1

表2

图2

图3

图4

表3

图5

表4

表5

图6

表6

表7

图7

表 8

参考文献 17

[1]	BARRETT L F, ADOLPHS R, MARSELLA S, et al. Emotional expressions reconsidered: challenges to inferring emotion from human facial movements[J]. Psychological Science in the Public Interest, 2019, 20(1): 1-68. doi: 10.1177/1529100619832930 pmid: 31313636
[2]	谢晶, 方平, 姜媛. 情绪测量方法的研究进展[J]. 心理科学, 2011, 34(2): 488-493.
	XIE Jing, FANG Ping, JIANG Yuan. Advances in emotion measuring methods[J]. Journal of Psychological Science, 2011, 34(2): 488-493.
[3]	谢黎, 史丰硕, 康晓管. 情绪测量技术综述[J]. 大众标准化, 2023(18): 31-34.
[4]	袁晓芳, 周垚, 刘成, 等. 应急场景下矿工生理指标评估的试验研究[J]. 煤炭工程, 2020, 52(5): 167-171. doi: 10.11799/ce202005035
	YUAN Xiaofang, ZHOU Yao, LIU Cheng, et al. Experimental study on evaluation of miners' physiological indexes in emergency scenarios[J]. Coal Engineering, 2020, 52(5): 167-171. doi: 10.11799/ce202005035
[5]	赵鹏飞, 贺阿红. 应急状态下情绪对个体心理变化的影响研究[J]. 矿业安全与环保, 2016, 43(6): 107-111.
	ZHAO Pengfei, HE Ahong. Study on impact of emotion on individual psychological changes under emergency state[J]. Mining Safety & Environmental Protection, 2016, 43(6): 107-111.
[6]	陈娜, 蒋正彪, 李新建, 等. 虚拟火灾场景构建及个体应激生理反应测评[J]. 消防科学与技术, 2018, 37(5): 698-702.
	CHEN Na, JIANG Zhengbiao, LI Xinjian, et al. The construction of virtual fire scene and the assessment of physiological reaction of individual stress[J]. Fire Science and Technology, 2018, 37(5): 698-702.
[7]	栗婧, 朱经政, 申彤. 面向公众的地震应激水平与影响因素评估[J]. 清华大学学报:自然科学版, 2024, 64(11): 1893-1901.
	LI Jing, ZHU Jingzheng, SHEN Tong. Assessment of earthquake stress levels and influencing factors for the public[J]. Journal of Tsinghua University:Science and Technology, 2024, 64(11): 1893-1901.
[8]	王树祎, 毛亮亮, 吴新阳, 等. 基于虚拟现实的灾害场景应激反应诱发方法[J]. 清华大学学报:自然科学版, 2023, 63(6): 960-967.
	WANG Shuyi, MAO Liangliang, WU Xinyang, et al. Method of stress reaction induction in disaster scenarios based on virtual reality[J]. Journal of Tsinghua University:Science and Technology, 2023, 63(6): 960-967.
[9]	杨超. 表情分析技术在可用性评估中的应用研究[D]. 杭州: 浙江理工大学, 2015.
	Yang Chao. The application research of facial expression analysis method in usability evaluation:facereader as an example[D]. Hangzhou: Zhejiang Sci-Tech University, 2015.
[10]	刘萌萌, 荣莉莉, 郄子君. 风险感知视角下公共风险事件负面情绪演化建模与仿真研究[J]. 管理学报, 2025, 22(2): 366-375.
	LIU Mengmeng, RONG Lili, QIE Zijun. Research on the modeling and simulation of evolution of negative emotions in public risk events from the perspective of risk perception[J]. Chinese Journal of Management, 2025, 22(2): 366-375.
[11]	TERRACCIANO A, MERRITT M, ZONDERMAN A B, et al. Personality traits and sex differences in emotion recognition among African Americans and Caucasians[J]. Annals of the New York Academy of Sciences, 2003: DOI: 10.1196/annals.1280.032.
[12]	薛艳, 朱达清, 冯程, 等. 重大突发事件下大学生心理危机状况及干预研究[J]. 心理学探新, 2023, 43(2): 184-192.
	XUE Yan, ZHU Daqing, FENG Cheng, et al. Study on the college students' psychological crisis under major emergencies and its intervention[J]. Psychological Exploration, 2023, 43(2): 184-192.
[13]	袁加锦, 汪宇, 鞠恩霞, 等. 情绪加工的性别差异及神经机制[J]. 心理科学进展, 2010, 18(12): 1899-1908.
	YUAN Jiajin, WANG Yu, JU Enxia, et al. Gender differences in emotional processing and its neural mechanisms[J]. Advances in Psychological Science, 2010, 18(12): 1899-1908.
[14]	卢家楣. 对气质的情绪特性的探讨[J]. 心理科学, 1995, 18(1): 59-61,58.
[15]	邓社军, 虞宇浩, 张俊林, 等. 基于行人恐慌情绪解析的改进社会力模型[J]. 中国安全科学学报, 2024, 34(2): 45-52. doi: 10.16265/j.cnki.issn1003-3033.2024.02.1277
	DENG Shejun, YU Yuhao, ZHANG Junlin, et al. Improved social force model based on analysis of pedestrian panic emotion[J]. China Safety Science Journal, 2024, 34(2): 45-52. doi: 10.16265/j.cnki.issn1003-3033.2024.02.1277
[16]	SANTARCANGELO J, ZHANG Xiaoping. Dynamic time-aligment k-means kernel clustering for time sequence clustering[C]. IEEE International Conference on Image Processing, 2015: 2532-2536.
[17]	蔡阿燕, 杨洁敏, 许爽, 等. 表达抑制调节负性情绪的男性优势:来自事件相关电位的证据[J]. 心理学报, 2016, 48(5): 482-494.
	CAI Ayan, YANG Jiemin, XU Shuang, et al. The male advantage in regulating negative emotion by expressive suppression: an event-related potential study[J]. Acta Psychologica Sinica, 2016, 48(5): 482-494. doi: 10.3724/SP.J.1041.2016.00482

名称	偏度	峰值	p
效价	-0.785	1.181	<0.01
唤醒度	-0.482	6.113	<0.01

名称	事故类型(平均值±标准差)					F	p
名称	地震	建筑火灾	森林火灾	生产事故	高处坠落	—	—
效价	0.09±0.10	0.15±0.11	0.07±0.10	0.06±0.09	0.10±0.10	2.521	0.044
唤醒度	0.29±0.05	0.30±0.05	0.28±0.06	0.30±0.04	0.29±0.07	0.307	0.873

事故类型	名称	试验前(均值±标准差)	试验后(均值±标准差)	统计量t	p	95%置信区间
地震	效价	-0.05±0.06	-0.09±0.10	2.037	0.050	[0.000,0.0716]
地震	唤醒度	0.33±0.07	0.29±0.05	5.283	<0.001	[0.028,0.063]
建筑火灾	效价	-0.08±0.09	-0.15±0.11	2.675	0.015	[0.015,0.125]
建筑火灾	唤醒度	0.35±0.08	0.30±0.05	3.541	0.002	[0.020,0.077]
森林火灾	效价	-0.04±0.08	-0.07±0.10	1.683	0.106	[-0.007,0.065]
森林火灾	唤醒度	0.33±0.07	0.28±0.06	3.286	0.003	[0.017,0.074]
生产事故	效价	-0.02±0.08	-0.06±0.09	2.469	0.020	[0.008,0.085]
生产事故	唤醒度	0.35±0.08	0.30±0.04	4.915	<0.001	[0.032,0.077]
高处坠落	效价	-0.08±0.07	-0.10±0.10	1.766	0.087	[-0.004,0.054]
高处坠落	唤醒度	0.33±0.09	0.29±0.07	3.594	0.001	[0.016,0.058]

情绪类别	偏度	峰值	p
愉快	3.842	15.841	<0.001
悲伤	2.535	8.878	<0.001
愤怒	3.374	16.179	<0.001
惊奇	3.266	12.889	<0.001
惧怕	5.058	29.655	<0.001
厌恶	3.108	9.761	<0.001

性别	情绪类别	愉快	悲伤	愤怒	惊奇	惧怕	厌恶
	女(n=46)中位数	0.002	0.067	0.009	0.005	0.002	0.001
	男(n=91)中位数	0.001	0.034	0.019	0.003	0.001	0.001
	p	0.133	0.022^*	0.037^*	0.188	0.006^**	0.779
气质类型	多血质(n=61)中位数	0.001	0.054	0.022	0.003	0.002	0.001
	抑郁质(n=17)中位数	0.004	0.047	0.007	0.004	0.001	0.001
	黏液质(n=15)中位数	0.002	0.029	0.015	0.001	0.002	0.002
	胆汁质(n=44)中位数	0.001	0.029	0.021	0.005	0.001	0.001
	p	0.369	0.029^*	0.048^*	0.565	0.910	0.850
事故类型	地震(n=34)中位数	0.002	0.026	0.020	0.005	0.002	0.001
	建筑火灾(n=20)中位数	0.002	0.070	0.022	0.003	0.004	0.002
	森林火灾(n=23)中位数	0.001	0.029	0.008	0.001	0.001	0.001
	生产事故(n=28)中位数	0.000	0.051	0.019	0.002	0.002	0.001
	高处坠落(n=32)中位数	0.004	0.041	0.015	0.006	0.001	0.002
	p	0.078	0.078	0.712	0.131	0.017^*	0.052