面部视觉特征在抑郁症诊断中的研究进展  

Research Progress of Facial Visual Features in Depression Diagnosis

作　　者：杜铭浩 刘爽 刘潇雅 张文荃 明东 DU Ming-hao;LIU Shuang;LIU Xiao-ya;ZHANG Wen-quan;MING Dong(Academy of Medical Engineering and Translational Medicine,Tianjin University,Tianjin 300072,China;Department of Biomedical Engineering,College of Precision Instruments and Optoelectronics Engineering,Tianjin University,Tianjin 300072,China)

机构地区：[1]天津大学医学工程与转化医学研究院,天津300072 [2]天津大学精密仪器与光电子工程学院生物医学工程与科学仪器系,天津300072

出　　处：《小型微型计算机系统》2022年第3期483-489,共7页Journal of Chinese Computer Systems

基　　金：天津市自然科学基金项目(19JCYBJC29200)资助。

摘　　要：基于面部视觉特征的抑郁症诊断方法借助计算机视觉技术,通过分析被试的面部肌肉和眼球相关运动特征来辅助抑郁症的早期检测.与目前临床上通常采用的医生访谈方式相比,基于面部视觉特征的抑郁症诊断法具有被试无须与外人外物交流接触且客观高效、普及性强与成本低的显著优点,可极大缓解医生患者比例不足、误诊率偏高现状,拥有广阔应用前景.本文从抑郁症患者的面部行为特点入手,综合介绍了目前常用的诱发实验范式、现有面部视觉特征公开数据库及基于面部视觉特征的抑郁症诊断的最新研究成果,最后简要讨论了存在问题与发展动向.The depression diagnosis method based on facial visual features uses computer vision technology to assist the early detection of depression by analyzing the facial muscle and eye movement features of the subjects.Compared with the doctor interview methods currently used in clinical,the depression diagnosis method based on facial visual features has the advantages of being objective and efficient,strong universality,low cost and non-contact.It can alleviate the current situation of the inadequate proportion of doctors and the high misdiagnosis rate,and has a broad application prospect.Starting from the facial behavior characteristics of patients with depression,this study reviews the commonly used inductive experiment paradigms,existing databases and the latest research results,and ends with the discussion of the existing problems and development trends.

关 键 词：抑郁症诊断 非接触式 面部视觉特征 眼动 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]