基于中医情志理论的悲喜情绪识别模式研究

发布时间：2018-07-26 12:03

【摘要】：目的:情绪识别作为情绪应用研究中的一个重要方向,在诸多应用领域有着重要的影响和需求,但情绪的主观体验具有很强的主观性,且人与人之间的情绪反应存在变异性,本研究立足于中医情志理论,结合现代情绪理论及研究成果,利用现代化的仪器设备,研究悲、喜情绪下被试的眼动变化、生理反应、面部表情及头部朝向是否存在差异、存在何种差异,以期得出一个较客观的悲、喜情绪综合识别模式。方法:采用现象学还原分析方法首先对中医情志概念进行文献研究。采用实验的方法对悲、喜情绪下的眼动变化、生理变化、面部表情变化及头部朝向变化进行研究。此次实验采取被试内设计,被试在实验前填写大五人格问卷中国简式版及正性负性情绪问卷,实验后填写情绪自评问卷,使用SPSS20.0对实验所获得的数据进行处理。本次实验,招募某医学院校视力正常或矫正视力正常大学生被试,排除正在患病或有慢性病史以及正在服用影响血液循环和中枢神经活动药物的个体。选取中国情绪影像材料系统悲视频—《我的兄弟姐妹》片段(时长00:01:52)、喜视频《唐伯虎点秋香》片段(时长00:01:52)诱发被试情绪,另从中国情绪图片系统中选取信息量接近的中性情绪图片10张,使用Tobii TX300 Eye Tracker眼动仪记录并采集被试在悲、喜情绪下投射在中性情绪图片上的眼动数据;使用荷兰Spirit公司的8通道生物反馈仪Nexus-10即时收集被试皮肤电导、血容量脉冲数据;同时,利用眼动仪显示屏幕上自带的摄像头实时拍摄下被试在观看情绪诱发材料时的面部表情视频,采用FaceReader 7面部表情分析系统对视频进行分析,获取被试面部表情数据及以鼻头三维坐标为代表的头部朝向数据。结果:文献研究的结果显示,中医情志概念具有意识意向性特征,眼动可作为意识指向性特征的一个外显指标。实验研究结果:1眼动结果对悲、喜情绪下的被试眼动变化进行研究,结果显示:(1)对中性情绪图片不进行分区处理,喜情绪下到首个注视点所用时间用时长(0.36±0.76)s,悲情绪下到首个注视点所用时间用时短(0.10±0.26),配对样本t检验显示二者差异显著[t=-2.862,P0.05];喜情绪下总注视时间用时短(9.15±2.32)s,悲情绪下总注视时间用时长(9.71±1.40),二者差异显著[t=2.041,P0.05];喜情绪下总注视次数少(28.52±8.91),悲情绪下总注视次数多(33.56±8.09),经检验两者差异显著;喜情绪下到首个注视点所用时间、首个注视点所用时间、总注视时间、总注视次数四个指标变异性大,悲情绪下这四个指标变异性小。(2)对中性情绪图片进行分区处理,则悲情绪下,右上区域对比其他三个区域,首个注视点所用时间最短(0.87±0.52)s,总注视时间最短(1.82±1.34)s,总注视次数最少(6.90±5.28),差异显著;喜情绪下,左下区域相较其他三个区域差异显著,首个注视点所用时间最短(0.88±0.76)s,总注视时间最短(1.66±1.67)s,总注视次数最少(5.25±5.50)。2生理反应结果对悲、喜情绪下的被试血容量脉冲(BVP)、皮肤电导(SC)进行研究发现:(1)采用时间序列分析,在情绪视频阶段及中性情绪图片阶段,BVP在悲、喜情绪下均呈现周期性变化,悲情绪下BVP变化周期短,变化幅度大,喜情绪下周期长,周期内变化幅度较小;对悲、喜情绪下被试BVP进行均值士标准差比较,悲情绪状态下血容量脉冲绝对值大、变异性大(-24.1037±12.0033),喜情绪状态下血容量脉冲绝对值小、变异性小(-22.6677 ±11.8012),配对t检验显示二者差异显著[t=5.210,P0.05];对悲、喜情绪下被试SC进行均值±标准差比较,喜情绪状态下皮肤电导大、皮肤电导变异性大(6.5352±4.5178),悲情绪状态下皮肤电导小、皮肤电导变异性小(6.0592±4.2335),配对t检验显示二者差异显著[t=3.229,P0.05]。(2)中性情绪图片阶段,采用时间序列分析显示悲、喜情绪下SC呈现明显周期性变化,喜情绪下SC周期短,变化幅度小,悲情绪下周期长,变化幅度大;对悲、喜情绪下被试BVP进行均值±标准差比较,悲情绪下BVP(-22.8410±11.9696)与喜情绪下BVP(-22.2760±11.7974)二者差异不明显;对悲、喜情绪下被试SC进行均值±标准差比较,喜情绪下SC值大、SC变异性大(6.4278±4.7352),悲情绪状态下SC小、SC变异性小(6.0147±4.3564),配对t检验显示二者差异显著[t=3.701,P0.05];(3)不同情绪主观感受强度对被试BVP、SC变化无影响(P值均大于0.05);(4)在本次实验中发现,悲、喜情绪下BVP与神经质这一人格特质呈显著正相关关系(r值均为0.406,P0.05)。3面部表情及头部朝向结果对悲、喜情绪下被试的面部表情及头部朝向三维坐标进行分析,结果发现:(1)悲、喜情绪下,中性情绪表情均占比最大,均大于50%;愤怒、恐惧、厌恶表情占比接近,悲伤表情占比较小,悲情绪下(0.1559±0.1252),喜情绪下(0.1207±0.1044),高兴表情占比最小,喜情绪下(0.0641±0.1037),悲情绪下(0.0168±0.0261)。(2)对悲、喜情绪下悲伤表情进行配对t检验,显示二者差异显著[t=-3.298,P0.05]、悲、喜情绪下高兴表情的配对t检验显示差异显著[t=3.822,P0.05],悲、喜情绪下中性表情比较显示二者差异明显[t=-2.394,P0.05]。(3)对被试面部20个动作单元运动持续时间及运动强度进行对比,发现在悲情绪下,被试突出表现为第15、第17、第23、第24动作单元的运动,在喜情绪下被试突出表现为第6、第10、第12、第25动作单元的运动。(4)不同强度的主观情绪感受与面部表情变化无相关性(P值均大于0.05)。(5)悲情绪下,唤醒度(面部肌肉运动强度)与BVP呈显著正相关关系(r值为0.349,P0.05)。(6)FaceReader 7面部表情分析系统采用鼻头三维坐标点反应被试头部朝向,该系统中以Y轴代表前后,X轴代表左右,Z轴代表上下。实验结果显示,悲、喜情绪下被试头部朝向存在差异,喜情绪下的鼻头三维坐标点主要落在Y+X-Z+、Y-X-Z+两个象限中的被试个数多,而悲情绪下的鼻头三维坐标点落在象限Y-X+Z+及Y+X+Z+区域内的被试个数明显要多,对比喜情绪差异显著。结论:中医情志概念与现代心理学的情绪概念二者在旨趣、内涵及外延上同中有异,情志概念强调情绪的意向性特征,眼动变化可较明显的反应这一意识指向性特征。实验研究则反映出了悲、喜情绪下被试在眼动变化、生理反应、面部表情及头部朝向上均具有不同的变化趋势。1眼动变化上,在注视时间、注视次数,悲、喜情绪下均表现出了不同的变化趋势。《素问.举痛论》言"悲则心系急""悲哀愁忧则心动,心动则五脏六腑皆摇",而"喜则气和志达,荣卫通利,故气缓矣","五脏六腑皆摇"则心不定,对外界刺激犹疑,注视点即显出纠缠凝滞的特征,表现在眼睛运动上,则见总注视次数多,总注视时间长,注视指标的变异性小;"荣卫通利"则对刺激的反应更显流利顺畅,表现在眼睛运动上,可见对刺激反应速度加快,总注视次数较少、总注视时间较短,注视指标的变异性大。到首个注视点所用时长悲情绪下用时短、喜情绪下用时长则反应出了悲情绪下被试更早关注到细节。在注视位置上,悲、喜情绪下的显著差异表现在对某个区域注视时间及注视次数的显著减少上,悲情绪下被试很少关注到右上区域,而喜情绪下被试很少关注到左下区域。2生理变化上,悲、喜血容量脉冲表现出了明显的周期性变化,且变化反应出了缓、急的不同,悲情绪下血容量脉冲周期小,变化大,表现出急的态势,喜情绪下血容量脉冲周期较大,变化较小,表现出缓的变化趋势;皮肤电导在情绪恢复阶段出现了明显的周期性变化并且也呈现出缓、急不同的态势来,喜情绪下周期小、变化幅度小,悲情绪下周期大,变化大,皮肤电导近似呈直线下降趋势。被试情绪主观感受强度与生理反应的关系在本次实验中未见有显著相关关系,说明情绪的主观感受与情绪的生理变化之间,没有明显一致性的变化规律,不能从情绪的主观感受强弱推导出生理反应的强弱,从另一方面,也可能说明个体的情绪主观感受具有掩饰性,提示今后对情绪进行客观识别的重要性。3面部表情上,无论悲、喜情绪下,中性表情占主导,这在侧面反应了亚洲被试在面部情绪表达上的含蓄,受"表露规则"限制较多。悲情绪下悲伤表情占比值与其他表情相比较、喜情绪下高兴表情占比值与其他表情相比较均未见有明显的差异变化,但将悲情绪与喜情绪下的悲伤表情占比值、悲情绪与喜情绪下的高兴表情占比值进行比较,则差异明显,说明在悲伤、高兴表情与其他类别情绪面部表情占比不显的情况下,可以从悲、喜情绪下面部表情的变化中分析被试情绪状态。对面部运动单元进行细分比较,则发现悲、喜情绪下被试面部动作单元在运动变化上存在明显差别,悲情绪下突出嘴角下撇、绷紧嘴唇、紧压嘴唇、下巴抬起这些动作单元的运动,喜情绪下突出表现为面颊提起、上嘴唇提起、拉动嘴角、微张嘴巴这些面部动作单元运动,悲、喜情绪下被试面部动作单元主要反映在嘴唇、下巴区域,喜情绪下主要表现为向上向外的肌肉运动趋势,而悲情绪下则整体表现为向下、收紧的肌肉运动趋势,面部动作单元的运动差异可以较好的区分被试的情绪状态。被试情绪主观感受强度与面部表情变化之间没有显著的相关关系,悲情绪下不同主观感受强度与悲伤表情、喜情绪下不同主观感受强度与高兴表情之间没有一致的变化规律,不能从情绪的主观感受强弱推导出面部表情的变化强弱。4头部朝向上,在要求被试在实验过程中尽量保持身体与头部位置不动的情况下,悲、喜情绪下被试头部位置变化细微但出现了明显差异,悲情绪下被试头部朝向倾向于往后左上区域移动,喜情绪下被试头部位置倾向于往前右上区域移动。头部位置的上、下变化在本次实验中未见有明显差异。意义与价值:1学术意义对中医情志概念的文献研究及对悲、喜情绪下的被试眼动变化及头部朝向变化进行的实验研究,验证了中医关于情绪意向性的观点。2应用价值中医情志理论认为情绪具有意向性特征,且提出"目为心之使"的观点,本次实验通过测量悲、喜情绪下被试眼动变化及头部朝向变化,发现悲、喜情绪下被试眼动变化及头部朝向变化存在下显著差异。从中医的情志理论及本次实验结果来看,眼动指标及头部朝向指标可作为有效的客观指标融合进情绪的综合识别模式中。再者,综合眼动、头部朝向、面部动作单元、自主神经系统指标的情志识别指标体系,为自动情志识别系统开发提供理论基础及数据支持,可促进情绪识别在医学、公共安全、心理学研究、人机互动、远程教育、用户体验评价、消费者等领域的转化应用。
[Abstract]:Objective: as an important direction in the research of emotional application, emotion recognition has important influence and demand in many application fields, but the subjective experience of emotion has strong subjectivity and the emotional reaction between people exists variability. This study is based on the theory of Chinese traditional emotion and combines modern emotion theory and research results. Using modern instruments and equipment, we studied the changes in eye movement, physiological reaction, facial expression and head orientation in the sad and happy mood, and what difference existed, in order to get a more objective pattern of sad and happy emotion recognition. Method: first, the literature study of the concept of traditional Chinese medicine was carried out by the method of phenomenology and the original analysis. The experimental method was used to study the changes of eye movement, physiological change, facial expression change and head orientation under the sad and happy mood. The experiment was designed in the experiment. Before the experiment, the subjects filled out the five personality questionnaire in China and the positive negative emotion questionnaire. After the test, the emotional self-assessment questionnaire was filled in and the SPSS20.0 was used to test the experiment. The data were processed. This experiment recruited a college student who had normal vision or corrected visual acuity in a medical college to exclude individuals who were suffering from illness or chronic medical history, and were taking drugs affecting blood circulation and central nervous activity. Time length 00:01:52), happy video "Tang Bo tiger point autumn fragrance >" fragment (time length 00:01:52) induced subjects' emotion, and from the Chinese mood picture system, 10 neutral mood pictures were selected, and Tobii TX300 Eye Tracker eye movement instrument was used to record and collect the eye movement data of the subjects on the neutral mood pictures in sad and happy mood; The 8 channel biofeedback instrument Nexus-10 of Holland Spirit company collected the test skin conductance, blood volume pulse data, and the facial expression video when watching the emotional evoked material on the screen with the camera on the screen of the eye moving instrument, and the video was analyzed by the FaceReader 7 facial expression analysis system. The data of facial expression and head orientation represented by the three-dimensional coordinates of the nose were taken. Results: the results of literature study showed that the concept of traditional Chinese medicine has the characteristics of consciousness intentionality, eye movement can be used as an explicit indicator of the directional characteristics of consciousness. Experimental results: 1 eye movement results on the changes of eye movement under the sad and happy mood. The results were as follows: (1) there was no partition treatment for neutral mood images, and the time used for the first fixation point was longer (0.36 + 0.76) s, and the time used for the first fixation point was shorter (0.10 + 0.26). The paired sample t test showed that the differences were significantly [t=-2.862, P0.05], and the total fixation time under the happy mood was short (9.15 + 2). .32) s, the total fixation time was long (9.71 + 1.40), the two difference was significant [t=2.041, P0.05], the total fixation number was less (28.52 + 8.91) under the happy mood (28.52 + 8.91), and the total fixation time was more (33.56 + 8.09). The time of the first fixation point, the time of the first fixation, the total fixation time, the total note were observed. The variation of the four indexes was large, and the four indexes were small in the mood of grief. (2) the neutral mood images were divided into four regions. The first fixation point was the shortest (0.87 + 0.52) s, the total fixation time was shortest (1.82 + 1.34) s, and the total fixation time was least (6.90 + 5.28), and the difference was significant. In the happy mood, the difference in the left lower region was significant compared with the other three regions, the shortest time (0.88 + 0.76) s, the shortest total fixation time (1.66 + 1.67) s, the least total fixation (5.25 + 5.50).2 physiological response to the sad, the BVP, and the skin conductance (SC) study found: (1) using the time sequence. In the column analysis, in the stage of emotional video and neutral mood picture stage, BVP has periodic changes in sad and happy mood. The BVP change cycle is short, the change range is large, the period of happy mood is long, and the amplitude of the change is small in the happy mood. For the sad and happy mood, the standard deviation of the mean scholar is compared with the sad mood, and the blood volume pulse under the sad mood state is absolute. The value was large, the variability was large (-24.1037 + 12.0033), the absolute value of blood volume pulse was small and the variability was small (-22.6677 + 11.8012). The paired t test showed that the difference between two and [t=5.210 was significantly [t=5.210, P0.05]. The mean mean deviation of SC under the sad and happy mood was compared, the skin conductance was large under the happy mood and the variability of skin conductance was 6.5352 + 4. .5178), the skin conductance was small and the variability of skin conductance was small (6.0592 + 4.2335). The paired t test showed that the difference between the two and the two was significant. The P0.05]. (2) neutral mood picture stage, the time series analysis showed that the SC showed a distinct periodic change under the happy mood, the SC cycle under the happy mood was short, the change range was small, and the period of sadness under the mood. The difference between the mean mean standard deviation of BVP and BVP (-22.8410 + 11.9696) and BVP (-22.2760 + 11.7974) under the mood of sadness was not obvious in sadness and happy mood, and the mean mean standard deviation of SC under sad and happy mood was compared, SC value was large under happy mood, SC variability (6.4278 + 4.7352), and S under sad mood. C was small, SC was small (6.0147 + 4.3564), and paired t test showed that the difference between two and two was significant [t=3.701, P0.05]; (3) the subjective feeling intensity of different emotions had no effect on BVP and SC (P value was greater than 0.05). (4) in this experiment, it was found that there was a significant positive correlation between BVP and neuroticism in this experiment (R values were 0.406, P0.05).3. The facial expressions and head orientation results were analyzed in the facial expression and head orientation of the subjects under the happy mood. The results were as follows: (1) under the sad, happy mood, the neutral emotional expression was the largest, both greater than 50%; anger, fear, disgust expression was close, sad expression was smaller, sad mood (0.1559 + 0.1252), happy mood (0.1559 + 0.1252). 0.1207 + 0.1044), the proportion of happy expression was the smallest, happy mood (0.0641 + 0.1037), sad mood (0.0168 + 0.0261). (2) the sad, sad expression of sad expression of paired t test, showing that two differences significantly, P0.05], sad, happy mood paired t test showed significant differences in [t=3.822, P0.05], sad, mood under the neutral table. The comparison showed that the differences between the two groups were obviously [t=-2.394, and P0.05]. (3) compared the motion duration and movement intensity of the 20 action units of the tested face. It was found that under the sadness, the subjects highlighted the movement of fifteenth, seventeenth, twenty-third, twenty-fourth action units, and the subjects displayed the movement of sixth, tenth, twelfth, and twenty-fifth action units under the happy mood. (4) there was no correlation between the subjective emotional feelings of different intensity and facial expression changes (P value was greater than 0.05). (5) in the mood of grief, the arousal degree (facial muscle movement intensity) had a significant positive correlation with BVP (r value 0.349, P0.05). (6) the FaceReader 7 facial expression analysis system adopted the head orientation of the three-dimensional coordinate point of the nose, and the system was Y Before and after the axis is represented, the X axis stands for the left and right and the Z axis stands up and down. The experimental results show that there is a difference in the head orientation of the subjects under the sad and happy mood. The three dimensional coordinates of the nose in the happy mood are mainly in the Y+X-Z+, and the number of the subjects in the two Y-X-Z+ quadrants, while the three dimensional coordinates of the nose in the sad mood fall in the quadrant Y-X+Z+ and the Y+X+Z+ region. There are obvious differences in the number of emotions. Conclusion: the concept of emotion of Chinese medicine and the concept of emotion in modern psychology are two different in the intention, connotation and extension, the concept of emotion emphasizes the intentional character of emotion, and the change of eye movement can reflect the characteristic characteristic of the consciousness. The experimental study reflects the sad and happy mood. The changes in eye movement, physiological response, facial expression and head orientation have different changes in the.1 eye movement. There are different trends in the fixation time, the number of gaze, the sad and the happy mood. Zhi Da, Rong Wei and Tongli, so the Qi is slow "," the five viscera and six organs are all rocking "the heart is not fixed, the irritation of the outside world is uncertain, the focus of the eye shows the characteristics of entanglement and stagnation. In the eye movement, the total gaze is more frequent, the total fixation time is long, and the variation of the fixation index is small." It can be seen that the speed of the stimulus response is faster, the total fixation time is less, the total fixation time is shorter, and the variability of the fixation index is large. In a significant reduction in the duration of fixation and the number of fixations in a certain region, the subjects rarely pay attention to the upper right area in the mood of sadness, while the subjects in the happy mood rarely pay attention to the.2 physiological changes in the lower left region. The cycle is small, the change is big, the situation is urgent, the blood volume pulse cycle is larger, the change is smaller, the change trend is slow, the skin conductance appears in the stage of emotional recovery, there are obvious periodic changes in the emotional recovery stage and also show a slow and different situation, the period of happy mood is small, the change range is small, the period of sadness is big, and the period of sadness is big and changed. There is no significant correlation between the subjective feeling intensity and physiological response in the experiment, which shows that there is no significant correlation between the subjective feelings of the subjects and the physiological changes of emotions, and there is no obvious change in consistency between emotion and emotion, and the physiological response can not be derived from the subjective feeling of emotion. On the other hand, on the other hand, it may also indicate that the subjective feelings of the individual are disguised. The importance of the objective recognition of emotions in the future.3 facial expression, regardless of sorrow, in the mood of happiness, is dominant in the neutral expression, which reflects the implication of the Asian subjects on the emotional expression of the face in the side, and is limited by the "disclosure rules". There is no obvious difference in the ratio of happy expression to other expressions, but the ratio of sadness to sad expression, and the ratio of sad emotion to happy expression under happy mood, the difference is obvious, indicating that sadness is high. The emotional state of the subjects can be analyzed from the changes of sadness and emotion under the circumstances of other categories of emotional facial expressions. Tightener lips, tight lips, and chin lifting the movement of these action units. In the happy mood, the expression is highlighted on the cheek, the upper lip is lifted, the mouth is lifted, the mouth is pulled, and the mouth is moved, and the face action units under the happy mood are mainly reflected in the lips, the chin area, and the main expression is the upward outward muscles under the happy mood. The trend of meat movement, while in grief, the overall performance is downward, tightened muscle movement trend, and the movement difference of facial movement unit can distinguish the emotional state of the subjects better. There is no significant correlation between the subjective feeling intensity of the subjects and the facial expression change, and the different subjective feeling intensity and sad expression in the mood of grief, There is no consistent change between the subjective feeling intensity and the happy expression, and the change of the facial expression can not be derived from the subjective feeling of the emotion, and the head orientation of the.4 can not be derived from the subjective feeling of the emotion. However, there was a significant difference. The head orientation of the subjects tended to move to the left upper area, and the head position of the subjects was in a happy mood.
【学位授予单位】：广州中医药大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：R229

【参考文献】