基于人脸分析技术探究惊悸不安人群的面部特征
Exploring Facial Features of Individuals with Anxiety Using Facial Analysis Technology
汪素梅 1程文龙 1盖路路 2齐向华3
作者信息
- 1. 山东中医药大学,山东 济南 250355
- 2. 山东大学,山东 济南 250100
- 3. 山东中医药大学附属医院,山东 济南 250355
- 折叠
摘要
目的:应用人脸关键点检测技术,探索惊悸不安人群面部特征.方法:基于主动外观模型作为人脸关键点检测技术,实现人脸68个关键点的自动识别;依据中医望面诊经验以及《人体测量手册》的相关论述,采用指数描述法提取5项人脸形态单元数据,实现对望面诊信息的数字化精确表达.在山东省中医院、齐鲁医院门诊采集符合《惊悸不安状态评定量表》147例为观察组,在山东中医药大学采集正常人145例为对照组.运用SPSS26.0统计软件对观察组与对照组的5项人脸形态单元数据进行组间差异比较,找到具有统计学差异的人脸形态单元数据.结果:对5项人脸形态单元数据进行正态性检验,R1与R4人脸形态单元数据符合正态性,R2、R3、R5人脸形态单元数据不符合正态分布;对R1与R4采用独立样本t检验,R1与R4人脸形态单元数据差异具有统计学意义;对R2、R3、R5采用秩和检验,R3人脸形态单元数据差异具有统计学意义,R2、R5人脸形态单元数据差异无统计学意义.结论:惊悸不安人群与正常人群具有人脸形态单元数据的差异,可以通过主动外观模型实现惊悸不安人群与正常人群面部特征的提取.
Abstract
Objective:To explore the facial features of individuals with anxiety using facial key point detection technology.Methods:The study utilized an active appearance model for facial key point detection to automatically identify 68 facial key points.Based on traditional Chinese medicine(TCM)facial diagnosis experience and relevant discussions in the"Handbook of Anthropometry,"the study employed an exponential description method to extract five facial morphology data points,achieving a precise digital representation of facial diagnostic information.The observation group consisted of 147 cases from the outpatient clinics of Shandong Provincial Hospital of Traditional Chinese Medicine and Qilu Hospital,who met the criteria of the"State of Anxiety Assessment Scale."The control group included 145 normal individuals from Shandong University of Traditional Chinese Medicine.SPSS 26.0 software was used to compare the five facial morphology data points between the observation and control groups,identifying statistically significant differences.Results:Normality tests of the five facial morphology data points showed that R1 and R4 data followed a normal distribution,while R2,R3,and R5 did not.Independent sample t-tests for R1 and R4 indicated statistically significant differences.Rank sum tests for R2,R3,and R5 revealed statistically significant differences for R3,but not for R2 and R5.Conclusion:There are differences in facial morphology data between individuals with anxiety and normal individuals.The active appearance model can effectively extract facial features of individuals with anxiety and normal individuals.
关键词
人脸分析技术/面诊/心理紊乱状态/主动外观模型Key words
Facial analysis technology/Facial diagnosis/Psychological disorder/Active appearance model引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据