抑郁症肝气郁症患者面部表情及眼神的识别

发布时间：2018-06-03 09:24

本文选题：抑郁症 + 肝气郁证　；参考：《中国中医基础医学杂志》2017年02期

【摘要】：目的:建立基于计算机的抑郁症肝气郁证患者面部表情及眼神模式识别系统,初步探索抑郁面部表情及眼神与抑郁情绪内心体验关系,寻求直接、准确地测评抑郁情绪的方法。方法:构建抑郁症肝气郁证患者和正常人的基本表情库,采集基本表情和视频,人工标记面部特征点,获得针对特定人的主动外观模型。使用主动外观模型对视频进行处理,跟踪每帧的面部特征点位置。结果:抑郁症肝气郁证患者的偏度值主要集中在0.5~7之间,中位数3.4左右,正常人在1.2`15之间,中位数6.3左右。结论:抑郁症肝气郁证患者的瞳孔波动较弱(目光呆滞),瞳孔内眼角距的绝对值的偏度在3.9以下,有较大概率患有抑郁症肝气郁证,是抑郁面部表情及眼神的主要特征。
[Abstract]:Objective: to establish a computer based facial expression and eye pattern recognition system for depressive patients with liver qi depression, to explore the relationship between depressive facial expressions and eyes and their inner experience, and to find out a direct and accurate method for evaluating depressive emotions. Methods: the basic facial expression database of patients with depression with liver qi depression and normal persons was constructed. The facial feature points were manually labeled to obtain the active appearance model for specific people. The active appearance model is used to process the video and track the position of facial feature points in each frame. Results: the deviations of depression patients with stagnation of liver qi were mainly between 0.5 and 7 with the median of 3.4, and the normal subjects between 1.2'15 and 6.3 respectively. Conclusion: the pupil fluctuation of depression patients with liver qi depression is weak (the absolute value of eye angle distance within pupil is less than 3.9, which is the main feature of depressive facial expression and eyes).
【作者单位】：山东中医药大学中医学院;山东省精神卫生中心;齐鲁工业大学;
【基金】：国家自然科学基金青年基金项目(81102738)-抑郁症肝气郁证抑郁眼神识别及其关键指标探索;国家自然科学基金面上项目(81573829)-抑郁症肝气郁证患者抑郁眼神、面部表情和姿态多模态识别及其脑区功能定位
【分类号】：R749.4

【相似文献】