WO2015149606A1 - 一种基于人体皮肤电阻变化的情绪检测方法及系统 - Google Patents

一种基于人体皮肤电阻变化的情绪检测方法及系统 Download PDF

Info

Publication number
WO2015149606A1
WO2015149606A1 PCT/CN2015/073954 CN2015073954W WO2015149606A1 WO 2015149606 A1 WO2015149606 A1 WO 2015149606A1 CN 2015073954 W CN2015073954 W CN 2015073954W WO 2015149606 A1 WO2015149606 A1 WO 2015149606A1
Authority
WO
WIPO (PCT)
Prior art keywords
resistance
value
emotional state
slope
window
Prior art date
Application number
PCT/CN2015/073954
Other languages
English (en)
French (fr)
Inventor
姚健欣
张昊
Original Assignee
姚健欣
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 姚健欣 filed Critical 姚健欣
Priority to SG11201600041PA priority Critical patent/SG11201600041PA/en
Publication of WO2015149606A1 publication Critical patent/WO2015149606A1/zh

Links

Images

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/05Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves 
    • A61B5/053Measuring electrical impedance or conductance of a portion of the body
    • A61B5/0531Measuring skin impedance
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/16Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
    • A61B5/165Evaluating the state of mind, e.g. depression, anxiety

Definitions

  • the invention relates to the field of human body detection, in particular to an emotion detection method based on a change in skin skin resistance.
  • the present invention discloses an emotion detecting method based on the change of human skin resistance, utilizing the skin resistance during the emotional fluctuation process.
  • the phenomenon of change is used to qualitatively and quantitatively analyze the degree of changes in human emotions and to make use of them.
  • an emotion detecting method based on a change in resistance of a human skin comprising the steps of:
  • Step 1 Parameter initialization
  • Step 2 Collect human body resistance data
  • Step 4 feedback the emotional state value to the user
  • the parameter initialization includes: setting an initial value of the emotional state value, a minimum value and a maximum value of the emotional state; setting a first resistance threshold lowThresh and a second resistance threshold highThresh, when the collected human skin resistance value is less than lowThresh, the human body is in an excited state When the collected human skin resistance value is greater than highThresh, the human body is in a calm state;
  • the data analysis includes data pre-processing and threshold analysis, slope analysis, slope and duration fit analysis, slope, duration, and threshold fit analysis.
  • the threshold analysis specifically includes the following steps:
  • the average value of the resistance of the current window is the average value of the resistance data collected in the window of length stepSize; MaxHistory and MinHistory are the maximum and minimum values of the average values of the resistances in the maxHistoryLen windows in front of the current window, respectively, maxHistoryLen
  • the value range is (3,20), ⁇ 1 , ⁇ 1 respectively represent the detection resistance drop and rise change sensitivity coefficient, ⁇ 1 value range is (0,1), ⁇ 1 value range is (1,5), ZZZ XXX indicates the step change of the emotional state, the value range is (1, 5), and ZZZ> XXX.
  • the initial state of the emotional state value is set to 1, the minimum value of the emotional state is set to 1, and the maximum value of the emotional state is set to 20.
  • the resistance threshold is set to 50K ⁇ according to finger sampling data, and highThresh is set to 800K ⁇ .
  • the length of the window is set to 80, the maxHistoryLen is set to 5, the ZZZ is set to 2, the XXX is set to 1, and ⁇ 1 and ⁇ 1 are set to 0.8 and 1.2, respectively.
  • slope analysis specifically includes the following steps:
  • MeanUpSlopes and MeanDownSlopes respectively represent the average value of the rising slope and the falling slope absolute value detected in the previous N times, the rising slope is the slope with the slope greater than 0, and the falling slope is the slope with the slope less than 0, ⁇ , ⁇ is the slope analysis
  • the parameter has a value range of (0, 1) and ⁇ > ⁇ .
  • DDD indicates the change step of the emotional state, the value range is (1, 5), and DDD>ZZZ>XXX.
  • ⁇ 2 and ⁇ 2 are set to 0.98 and 1.02, respectively, ⁇ is set to 0.85, ⁇ is set to 0.5, DDD is set to 3, and N is set to 5.
  • slope and duration matching analysis specifically includes the following steps:
  • the excitation duration DownDuration increases the time corresponding to one window
  • RecoverTime indicates the recovery time
  • ExcitTime indicates the excitement time
  • the range of values is (0,200s).
  • the RecoverTime is set to 8s and the ExcitTime is set to 5s.
  • slope, duration, and threshold fit analysis specifically includes the following steps:
  • the emotional state value is maintained; if the DownDuration exceeds the time length ExcitTime, and the current resistance value is less than ⁇ 1 *MaxHistory, the emotional state value is increased by ZZZ, otherwise the emotional state value is increased by XXX.
  • the resistance threshold is based on finger sampling data, lowThresh is set to 50K ⁇ , and highThresh is set to 800K ⁇ .
  • an emotion detecting system based on human skin resistance change for implementing the above method, which comprises sequentially connecting medical pole pieces, bridge resistance/conductance measuring circuit, amplifying circuit, A/D Conversion circuit, CPU and human-computer interface.
  • the invention combines the threshold analysis of the human body resistance, the slope analysis of the resistance change, and the analysis of the excitement and calm duration.
  • an adaptive algorithm is adopted, which can more clearly divide the calm and excitement state.
  • the tester's state of excitement is divided into different gears, and the degree of excitement is reflected by the difference in emotional state values.
  • the acquisition, calculation, and feedback research mode it can be used in future wearable devices.
  • Figure 1 is a basic framework of the present invention
  • Figure 2 is an overall flow chart of the present invention
  • Figure 5 is a schematic diagram showing changes in duration of the present invention.
  • Figure 6 is a flow chart showing the analysis of the slope and duration of the present invention.
  • Figure 8 is a schematic diagram of window division according to the present invention.
  • Figure 9 is a schematic diagram of the system of the present invention.
  • Figure 10 shows the human body deep breathing frequency of 2 seconds, time 20s, detected resistance / emotional state values - Time curve
  • Figure 11 is a graph showing the resistance/emotion state value-time curve of the human body with a deep breathing frequency of 2 seconds and a time of 60 s;
  • Figure 12 is a diagram showing the resistance/emotion state value-time curve of the human body deep breathing frequency of 2 seconds, time 60s;
  • Figure 13 is a graph showing the resistance/emotion state value-time curve of the human body with a deep breathing frequency of 10 seconds and a time of 60 seconds.
  • the main idea framework of the present invention is as shown in FIG. 1 , collecting the skin resistance of the user, and then performing analysis and processing after receiving the resistance, and then feeding back the analyzed emotional state value to the user through the light flashing frequency or the specific numerical value display.
  • the data analysis process includes preprocessing and parameter analysis.
  • the parameter analysis includes analysis of the combined use of resistance threshold analysis, resistance change slope analysis, and threshold, slope, and excitement calm duration, as shown in Figure 2.
  • each window, stepSize is set to 80, that is, 80 resistor values are collected, wherein the resistance sampling frequency is 50 Hz, that is, the resistance distribution within 1.6 seconds.
  • MaxHistory MinHistory
  • ⁇ 1 0.8 and 1.2, respectively.
  • MaxHistory and MinHistory are the maximum and minimum values of the average value of the resistance data in the window maxHistoryLen (set to 4) in front of the current window. If the current window resistance is greater than ⁇ 1 *MinHistory, the tester's resistance changes from small to large, the tester recovers from excitement to calm, and his emotional state decreases by XXX. Note that during the emotional state reduction, the emotional state minimum MINSTATE (set to 1).
  • the current window's resistance mean is less than ⁇ 1 *MaxHistory, the resistance changes from large to small, the tester changes from calm to excited, and the emotional state value increases by XXX (set to 1). Note that during the increase process, the maximum emotional state is set to MAXSTATE (set to 20).
  • Threshold adaptation means that each time the resistance is changed, the average value of the resistance in the front window is used, and when the window moves forward in sequence, the mean value retained in the front also moves forward.
  • Adaptive means that each time the object is compared, it is compared with the mean of the previous period, not fixed.
  • the current slope value is greater than 0 and greater than ⁇ *MeanUpSlopes, less than ⁇ *MeanUpSlopes, say The tester is slower from excitement to calmness, then the current emotional state is reduced by ZZZ (set to 2). If the current slope is less than 0, its absolute value is greater than ⁇ *MeanDownSlopes and less than ⁇ *MeanDownSlopes, indicating that the tester is calm. The excitement changes slowly, then the current emotional state increases ZZZ ( ⁇ is set to 0.5 here).
  • the current mood is reduced by XXX (set to 1); the slope is less than 0, and the absolute value is less than ⁇ *MeanDownSlopes, then the current emotional state is increased by XXX.
  • the previously collected slope value and threshold analysis data are used in determining the excitement and calm duration. If the slope is greater than 0 and the current window resistance average is greater than ⁇ 2 *MinHistory, the recovery duration UpDuration increases the time required to acquire a window of stepsize resistors. Conversely, if the slope is less than 0 and the current mean resistance in the window is less than ⁇ 2 *MaxHistory, then the excitement duration DownDuration increases stepSize. The initial values of UpDuration and DownDuration are 0.
  • DDD, ZZZ, and XXX can take values of 3, 2, and 1, respectively. Indicates that the magnitude of the increase is different.
  • the three parameters of slope, threshold and duration are used to analyze and analyze the emotional state of the tester. As shown in Figure 7. First determine whether the current slope value is positive. If it is greater than 0, it indicates a transition from excitement to calm. When the recovery duration exceeds RecoverTime (set to 8s), if the current resistance is greater than ⁇ 1 *MinHistory, the emotional state is reduced by ZZZ, otherwise the emotional state is decreased by XXX. If the slope is less than 0, it means changing from calm to excited state. When the excitement duration exceeds ExcitTime (set to 5s), if the current window average resistance is less than ⁇ 1 *MaxHistory, the emotional state value increases ZZZ, otherwise it increases XXX.
  • FIG. 9 is a schematic diagram of an emotion detection system based on human skin resistance change, comprising a medical pole piece, a bridge resistance/conductance measurement circuit, an amplification circuit, an A/D conversion circuit, a CPU, and a human-computer interaction interface. It is used to collect human skin resistance (conductance) data, and then detect emotional changes based on changes in human skin resistance (conductance). The method of the invention is implemented.
  • Figure 10 - Figure 13 is the experimental resistance/emotion state value-time curve.
  • inputData is the user's skin resistance value after pre-processing
  • EstimateState is the threshold, slope and duration combined with the analysis method. Status value.
  • Figure 10 shows the deep breathing frequency of the human body in 2 seconds, the time is 20s, and the detected resistance/emotional state value-time curve
  • Figure 11 shows the deep breathing frequency of the human body is 2 seconds once, the time is 60s, and the detected resistance/emotion State value - time curve
  • comparison shows that the longer the deep breathing time, the longer the duration of the excitement, the higher the detected emotional state value.
  • Figure 12 shows the deep breathing frequency of the human body in 2 seconds, the time is 60s, the detected resistance/emotional state value-time curve;
  • Figure 13 shows the deep breathing frequency of the human body is 10 seconds once, the time is 60s, the detected resistance/emotion State value - time curve; comparison found that the faster the deep breathing frequency, the higher the emotional state value.
  • the experimental data shows that the method of the present invention can more clearly divide the state of calm and excitement, and express the different levels of excitement of the human body.
  • the invention combines the threshold analysis of the human body resistance, the slope analysis of the resistance change, and the analysis of the excitement and calm duration.
  • an adaptive algorithm is adopted, which can more clearly divide the calm and excitement state.
  • the tester's state of excitement is divided into different gears, and the degree of excitement is reflected by the difference in emotional state values.
  • the acquisition, calculation, and feedback research mode it can be used in future wearable devices.

Abstract

一种基于人体皮肤电阻变化的情绪检测方法及系统,上述方法包括以下步骤:参数初始化、采集人体电阻数据、数据分析、将情绪状态值反馈给用户;所述数据分析包括预处理及阈值分析,斜率分析,斜率和持续时间配合分析,斜率、持续时间和阈值配合分析。上述系统包括依次连接的医疗极片、桥式电阻/电导测量电路、放大电路、A/D转换电路、CPU和人机交互界面。上述方法和系统采用自适应的算法,利用电阻的大小、斜率以及状态持续时间变化来反映情绪,能克服个体差异问题,更加精细检测测试者兴奋程度。

Description

一种基于人体皮肤电阻变化的情绪检测方法及系统
本申请要求了2014年4月1日提交的、申请号为201410128494.1、发明名称为“一种基于人体皮肤电阻变化的情绪检测方法及系统”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本发明涉及人体检测领域,具体涉及一种基于人体皮肤电阻变化的情绪检测方法。
背景技术
生物在感受到外界刺激以及情绪激动时,其皮肤电阻由于两点之间的电位差增大,导致其皮肤两点间的电阻变小,这种现象称之为生理电反射。人们对此现象产生的研究并不是很深入,目前较为权威的解释是人体皮肤电阻发生变化是与汗腺活动有关。情绪激动时,汗水分泌较多,而汗水中存在大量电解质,导致人体皮肤表面电阻发生变化。
许多研究者对于情绪和电阻之间的关系也曾有过分析,但是却致力于为何情绪波动会影响人体皮肤电阻的变化,他们仅仅以电阻大小的变化来区分测试者是处在平静还是兴奋状态。这样不仅不能够反映测试者兴奋程度,而且对于不同的测试者,其个体差异性较大,存在许多不确定因素(比如有的人激动时爱出汗,其电阻变化很明显,有的人则不然;不同体重的人,在同样的测试条件下,也会有不同的电阻变化),其电阻变化范围往往差距较大,因此不能直接以电阻大小的变化来确定其处在平静或兴奋状态,本发明旨在补充这一方面的缺陷。
发明内容
为了克服以往研究中只对情绪状态进行平静和兴奋两个状态的划分,且没有考虑个体差异的问题,本发明公开了一种基于人体皮肤电阻变化的情绪检测方法,利用情绪波动过程中皮肤电阻变化这一现象来定性定量地分析出人体情绪变化的程度,从而加以利用。
根据本发明的第一方面,提供一种基于人体皮肤电阻变化的情绪检测方法,包括以下步骤:
步骤一:参数初始化;
步骤二:采集人体电阻数据;
步骤三:数据分析;
步骤四:将情绪状态值反馈给用户;
所述参数初始化包括:设置情绪状态值初始值、情绪状态的最小值和最大值;设置第一电阻阈值lowThresh和第二电阻阈值highThresh,当所采集的人体皮肤电阻值小于lowThresh时,人体处于兴奋状态;当所采集的人体皮肤电阻值大于highThresh时,人体处于平静状态;
所述数据分析包括数据预处理及阈值分析、斜率分析、斜率和持续时间配合分析、斜率、持续时间和阈值配合分析。
进一步地,所述阈值分析具体包括以下步骤:
若当前窗口的电阻平均值小于lowThresh,则情绪状态值增加ZZZ;否则
若当前窗口的电阻平均值大于highThresh,则情绪状态值减小ZZZ;否则
若当前窗口的电阻平均值大于β1*MinHistory,则其情绪状态减小XXX;否则
若当前窗口的电阻平均值小于α1*MaxHistory,则情绪状态值增加XXX;否则情绪状态值保持不变;
其中当前窗口的电阻平均值是指长度为stepSize的窗口内采集到的电阻数据的平均值;MaxHistory和MinHistory分别为当前窗口前面的maxHistoryLen个窗口内电阻的平均值的最大值和最小值,maxHistoryLen取值范围为(3,20),α1,β1分别表示检测电阻下降和上升变化灵敏度系数,α1取值范围为(0,1),β1取值范围为(1,5),ZZZ、XXX表示情绪状态变化步长,取值范围为为(1,5),且ZZZ>XXX。
进一步地,所述情绪状态值初始值设置为1,情绪状态的最小值设置为1,情绪状态的最大值设置为20,所述电阻阈值根据手指采样数据,lowThresh设置为50KΩ,highThresh设置为800KΩ,窗口的长度stepSize设置为80,maxHistoryLen设置为5,ZZZ设置为2,XXX设置为1,α1,β1分别设置为0.8和1.2。
进一步地,所述斜率分析具体包括以下步骤:
1)检测电阻变化,如果有连续SlopesLen个窗口电阻的平均值小于α2*MaxHistory或大于β2*MinHistory时,其中α2、β2分别表示检测电阻下降和上升变化灵敏度系数,α2取值范围为(0,1),β2取值范围为(1,5),计算并记录这SlopesLen个窗口的斜率:
Slopes=(aveValue2–aveValue1)/(SlopesLen*stepSize),其中aveValue1表示第一个窗口的平均值,aveValue2表示最后一个窗口的平均值;
2)若Slopes大于0,且Slopes>ξ*MeanUpSlopes,则当前情绪状态减小DDD;否则若η*MeanUpSlopes<Slopes≤ξ*MeanUpSlopes,则当前情绪状态减小ZZZ;否则当前情绪减小XXX;
若Slopes小于0,且|Slopes|>ξ*MeanDownSlopes,则当前情绪状态增加DDD;否则若η*MeanDownSlopes<|Slopes|≤ξ*MeanDownSlopes,则当前情绪状态增加ZZZ;否则当前情绪状态增加XXX;
其中,MeanUpSlopes、MeanDownSlopes分别表示前面N次检测到的上升斜率和下降斜率绝对值的平均值,上升斜率是指斜率大于0的斜率,下降斜率是指斜率小于0的斜率,ξ,η为斜率分析参数,取值范围为(0,1),且ξ>η。DDD表示情绪状态变化步长,取值范围为(1,5),且DDD>ZZZ>XXX。
进一步地,所述α2、β2分别设置为0.98和1.02,ξ设置为0.85,η设置为0.5,DDD设置为3,N设置为5。
进一步地,所述斜率和持续时间配合分析具体包括以下步骤:
获取持续时间参数:
如果Slopes大于0,当前窗口内电阻平均值大于β2*MinHistory,则对恢复持续时间UpDuration增加采集一个窗口电阻需要的时间;
如果Slopes小于0,且当前窗口内电阻均值小于α2*MaxHistory,则兴奋持续时间DownDuration增加一个窗口对应的时间;
根据持续时间参数更新情绪状态值:
若Slopes大于0,且UpDuration大于RecoverTime,则情绪状态值减小XXX;
若Slopes小于0,且DownDuration大于ExcitTime,则情绪状态值增加XXX;
其中RecoverTime表示恢复时间,ExcitTime表示兴奋时间,取值范围均为(0,200s)
进一步地,所述RecoverTime设置为8s,ExcitTime设置为5s。
进一步地,所述斜率、持续时间和阈值配合分析具体包括以下步骤:
若Slopes大于0,且UpDuration小于时间长度RecoverTime,则情绪状态值保持;否则若UpDuration大于时间长度RecoverTime,且当前电阻值大于β1*MinHistory,则情绪状态值减小ZZZ,否则情绪状态值减小XXX;
若Slopes小于0,且DownDuration小于时间长度ExcitTime,则情绪状态值保持;若DownDuration超过时间长度ExcitTime,且当前电阻值小于α1*MaxHistory,则情绪状态值增加ZZZ,否则情绪状态值增加XXX。
所述电阻阈值根据手指采样数据,lowThresh设置为50KΩ,highThresh设置为800KΩ。
因为电导是电阻的倒数,所以以上分析方法同样适用于基于电导变化来分析情绪变化。
根据本发明的第二方面,提供一种基于人体皮肤电阻变化的情绪检测系统,用于实施上述方法,其包括依次连接的医疗极片、桥式电阻/电导测量电路、放大电路、A/D转换电路、CPU和人机交互界面。
本发明结合了人体电阻的阈值分析,电阻变化斜率分析以及兴奋、平静持续时间分析,对于个性问题,采用自适应的算法,能更加精细地划分平静和兴奋状态。利用情绪状态这一变量把测试者的兴奋状态分成了不同档位,通过情绪状态值的不同来反应多级兴奋程度。并且利用采集、计算、反馈式的研究模式,可以在以后的可穿戴设备等中加以利用。
附图说明
图1为本发明基本框架;
图2为本发明整体流程图;
图3为本发明阈值分析流程图;
图4为本发明斜率分析流程图;
图5为本发明持续时间变化示意图;
图6为本发明斜率和持续时间配合分析流程图;
图7为本发明斜率、持续时间和阈值配合分析流程图;
图8为本发明窗口划分示意图;
图9为本发明的系统原理图;
图10为人体深呼吸频率为2秒一次,时间为20s,检测到的电阻/情绪状态值— 时间曲线图;
图11为人体深呼吸频率为2秒一次,时间为60s,检测到的电阻/情绪状态值—时间曲线图;
图12为人体深呼吸频率为2秒一次,时间为60s,检测到的电阻/情绪状态值—时间曲线图;
图13为人体深呼吸频率为10秒一次,时间为60s,检测到的电阻/情绪状态值—时间曲线图。
具体实施方式
以下结合附图对本发明进行进一步说明,
本发明主要思路框架如图1所示,采集用户的皮肤电阻,等接收到电阻后进行分析处理,然后将分析得到的情绪状态值通过灯光闪烁频率或者具体数值显示反馈给用户。其中的数据分析过程包括预处理和参数分析。其中参数分析包括通过电阻阈值分析、电阻变化斜率分析以及阈值、斜率、兴奋平静持续时间三个参数的配合使用分析,见附图2。
1、预处理。
由于皮肤电阻采集仪器有规定的量程,其采集的电阻值在超出量程范围后,便会出现电阻为0的值,因此需要对采集的电阻数据进行预处理,将电阻为0的值用前一时刻的电阻代替;其次由于测试者的电阻波动较大,需要选用一定的窗口stepSize进行平均以减小波动便于分析。如图7所示。每个窗口的长度stepSize设置为80,即采集80个电阻值,其中电阻采样频率为50Hz,即1.6秒内的电阻分布。
2、参数分析
1)、阈值分析,具体流程见附图3。
①共性分析。从所有采样到的手指电阻数据可以看到,所有数据中电阻小于50KΩ时,90%左右都是处于兴奋状态;同样大于800KΩ时,80%处于平静状态,因此将lowThresh设置为50KΩ,highThresh设置为800KΩ。
检测测试者的情绪状态时,首先进行共性分析。当其当前电阻值小于50KΩ时,其情绪状态(CurrES)增加ZZZ(设为2),当其当前电阻值大于800KΩ时,其情绪状态(CurrES)减小ZZZ。
lowThresh、highThresh能很好的反映所有测试者共有的性质。
②阈值参数学习。在这里另外设置了四个变量:MaxHistory、MinHistory、α1,β1。这里α1,β1分别设置为0.8和1.2。MaxHistory、MinHistory为当前窗口前面maxHistoryLen(设为4)个窗口内电阻数据平均值的最大值和最小值。如果当前窗口的电阻均值大于β1*MinHistory,说明测试者电阻从小到大变化,测试者从兴奋向平静恢复,其情绪状态减小XXX。在情绪状态减小过程中注意,情绪状态最小值MINSTATE(设为1)。如果当前窗口的电阻均值小于α1*MaxHistory,说明电阻从大到小变化,测试者从平静到兴奋转变,其情绪状态值增加XXX(设为1)。在增加过程须注意,情绪状态最大值设置为MAXSTATE(设为20)。
采用阈值自适应的算法,能够更加准确地划分测试者所处的状态。阈值自适应是指每次在判断电阻变化时,都用前面窗口内的电阻均值,窗口依次向前移动的时候,前面保留的均值也是向前移动。自适应就是说每次比较的对象不同,都是和前一段时间内的均值比较,而不是固定的。
2)、斜率分析,具体流程见附图4。
①斜率采集。
首先检测电阻变化,如果有连续SlopesLen(设为5)个窗口(窗口大小为80)电阻小于α2*MaxHistory或大于β2*MinHistory时,α2、β2分别设置为0.98和1.02。记录此SlopesLen个窗口的电阻均值,并计算斜率Slopes=(aveValue2–aveValue1)/(SlopesLen*stepSize),其中aveValue2表示SlopesLen个窗口测得的最后一个电阻值,aveValue1表示SlopesLen个窗口测得的第一个电阻值;
②不同斜率分级。
如果当前斜率值大于0,且大于ξ*MeanUpSlopes,说明测试者由兴奋到平静恢复地较快,那么当前情绪状态减小DDD(设为3);如果当前斜率小于0,其绝对值大于ξ*MeanDownSlopes,说明测试者由平静到兴奋转变的很快,那么当前情绪状态增加DDD。MeanUpSlopes、MeanDownSlopes分别表示检测到的上升和下降斜率绝对值的平均值。(
Figure PCTCN2015073954-appb-000001
设置为0.85)
如果当前斜率值大于0,且大于η*MeanUpSlopes,小于ξ*MeanUpSlopes,说 明测试者由兴奋到平静恢复地较慢,那么当前情绪状态减小ZZZ(设为2),如果当前斜率小于0,其绝对值大于η*MeanDownSlopes、小于ξ*MeanDownSlopes,说明测试者由平静到兴奋转变的较慢,那么当前情绪状态增加ZZZ(η在这里设置为0.5)。
如果检测到斜率大于0,且小于η*MeanUpSlopes,则当前情绪减小XXX(设为1);斜率小于0,且绝对值小于η*MeanDownSlopes,则那么当前情绪状态增加XXX。
增加过程须注意,不能超过最大状态MAXSTATE(1);减小过程须注意,不能低于最小状态MINSTATE(20)。
3)、阈值、斜率、持续时间配合分析,具体见图5、图6、图7。
①获取持续时间参数
如图5所示,在判定兴奋、平静持续时间时需要用到先前采集到的斜率值及阈值分析数据。如果斜率大于0,且当前窗口内电阻均值大于β2*MinHistory,则恢复持续时间UpDuration增加采集一个窗口stepsize个电阻需要的时间。反之,若斜率小于0,且当前窗口内电阻均值小于α2*MaxHistory,则兴奋持续时间DownDuration增加stepSize。UpDuration和DownDuration的初始值为0。
②斜率、持续时间参数配合分析
如图6所示。若斜率大于0,且恢复持续时间UpDuration超过时间长度RecoverTime(如8s),则情绪状态值减小XXX。若斜率小于0,且DownDuration超过某个时间长度ExcitTime(如5s),则情绪状态值增加XXX。
DDD、ZZZ、XXX分别可以取值3、2、1。表示增加的幅度不同。
③阈值、斜率、持续时间参数配合分析
为了能更加准确的分析测试者的兴奋程度,用斜率、阈值、持续时间三个参数配合学习分析测试者的情绪状态。如图7所示。首先判断当前斜率值是否为正,如果大于0,说明从兴奋到平静转变。当恢复持续时间超过RecoverTime(设为8s)时,若当前电阻大于β1*MinHistory,则情绪状态减小ZZZ,否则情绪状态减小XXX。如果斜率小于0,说明从平静转变为兴奋状态,当兴奋持续时间超过ExcitTime(设为5s),若当前窗口内电阻均值小于α1*MaxHistory,则情绪状态值增加ZZZ,否则增加XXX。
图9为本发明一种基于人体皮肤电阻变化的情绪检测系统原理图,包括依次连接的医疗极片、桥式电阻/电导测量电路、放大电路、A/D转换电路、CPU和人机交互界面,用来采集人体皮肤电阻(电导)数据,进而基于人体皮肤电阻(电导)变化检测情绪变化。实现本发明的方法。
图10——图13为实验得到的电阻/情绪状态值——时间曲线图,图中inputData是经过预处理后用户的皮肤电阻值,EstimateState即为阈值、斜率和持续时间配合分析方法到的情绪状态值。
实施例1:
图10为人体深呼吸频率为2秒一次,时间为20s,检测到的电阻/情绪状态值——时间曲线图;图11为人体深呼吸频率为2秒一次,时间为60s,检测到的电阻/情绪状态值——时间曲线图;对比发现深呼吸时间越久,表示兴奋持续时间越长,检测到的情绪状态值越高。
实施例2:
图12为人体深呼吸频率为2秒一次,时间为60s,检测到的电阻/情绪状态值——时间曲线图;图13为人体深呼吸频率为10秒一次,时间为60s,检测到的电阻/情绪状态值——时间曲线图;对比发现,深呼吸频率越快,情绪状态值越高。
实验数据表明,通过本发明的方法可以能更加精细地划分平静和兴奋状态,表现人体不同的兴奋程度。
本发明结合了人体电阻的阈值分析,电阻变化斜率分析以及兴奋、平静持续时间分析,对于个性问题,采用自适应的算法,能更加精细地划分平静和兴奋状态。利用情绪状态这一变量把测试者的兴奋状态分成了不同档位,通过情绪状态值的不同来反应多级兴奋程度。并且利用采集、计算、反馈式的研究模式,可以在以后的可穿戴设备等中加以利用。

Claims (10)

  1. 一种基于人体皮肤电阻变化的情绪检测方法,其特征在于,包括以下步骤:
    步骤一:参数初始化;
    步骤二:采集人体电阻数据;
    步骤三:数据分析;
    步骤四:将情绪状态值反馈给用户;
    所述参数初始化包括:设置情绪状态值初始值、情绪状态的最小值和最大值;设置第一电阻阈值lowThresh和第二电阻阈值highThresh,当所采集的人体皮肤电阻值小于lowThresh时,人体处于兴奋状态;当所采集的人体皮肤电阻值大于highThresh时,人体处于平静状态;
    所述数据分析包括数据预处理及阈值分析、斜率分析、斜率和持续时间配合分析、斜率、持续时间和阈值配合分析。
  2. 根据权利要求1所述的基于人体皮肤电阻变化的情绪检测方法,其特征在于,所述阈值分析具体包括以下步骤:
    若当前窗口的电阻平均值小于lowThresh,则情绪状态值增加ZZZ;否则
    若当前窗口的电阻平均值大于highThresh,则情绪状态值减小ZZZ;否则
    若当前窗口的电阻平均值大于β1*MinHistory,则其情绪状态减小XXX;否则
    若当前窗口的电阻平均值小于α1*MaxHistory,则情绪状态值增加XXX;否则情绪状态值保持不变;
    其中当前窗口的电阻平均值是指长度为stepSize的窗口内采集到的电阻数据的平均值;MaxHistory和MinHistory分别为当前窗口前面的maxHistoryLen个窗口内电阻的平均值的最大值和最小值,maxHistoryLen取值范围为(3,20),α1,β1分别表示检测电阻下降和上升变化灵敏度系数,α1取值范围为(0,1),β1取值范围为(1,5),ZZZ、XXX表示情绪状态变化步长,取值范围为为(1,5),且ZZZ>XXX。
  3. 根据权利要求2所述的基于人体皮肤电阻变化的情绪检测方法,其特征在于,所述情绪状态值初始值设置为1,情绪状态的最小值设置为1,情绪状态的最大值设置为20,所述电阻阈值根据手指采样数据,lowThresh设置为50KΩ,highThresh设置为800KΩ,窗口的长度stepSize设置为80,maxHistoryLen设置为5,ZZZ设置为2,XXX设置为1,α1,β1分别设置为0.8和1.2。
  4. 根据权利要求2所述的基于人体皮肤电阻变化的情绪检测方法,其特征在于,所述斜率分析具体包括以下步骤:
    1)实时检测电阻变化,如果有连续SlopesLen个窗口电阻的平均值小于α2*MaxHistory或大于β2*MinHistory时,其中α2、β2分别表示检测电阻下降和上升变化灵敏度系数,α2取值范围为(0,1),β2取值范围为(1,5),计算并记录这SlopesLen个窗口的斜率:
    Slopes=(aveValue2–aveValue1)/(SlopesLen*stepSize),其中aveValue1表示第一个窗口的平均值,aveValue2表示最后一个窗口的平均值;
    2)若Slopes大于0,且Slopes>ξ*MeanUpSlopes,则当前情绪状态减小DDD;否则若η*MeanUpSlopes<Slopes≤ξ*MeanUpSlopes,则当前情绪状态减小ZZZ;否则当前情绪减小XXX;
    若Slopes小于0,且|Slopes|>ξ*MeanDownSlopes,则当前情绪状态增加DDD;否则若η*MeanDownSlopes<|Slopes|≤ξ*MeanDownSlopes,则当前情绪状态增加ZZZ;否则当前情绪状态增加XXX;
    其中,MeanUpSlopes、MeanDownSlopes分别表示前面N次检测到的上升斜率和下降斜率绝对值的平均值,其中,上升斜率是指斜率大于0的斜率,下降斜率是指斜率小于0的斜率,ξ,η为斜率分析参数,取值范围为(0,1),且ξ>η。DDD表示情绪状态变化步长,取值范围为(1,5),且DDD>ZZZ>XXX。
  5. 根据权利要求4所述的基于人体皮肤电阻变化的情绪检测方法,其特征在于,所述α2、β2分别设置为0.98和1.02,ξ设置为0.85,η设置为0.5,DDD设置为3,N设置为5。
  6. 根据权利要求4所述的基于人体皮肤电阻变化的情绪检测方法,其特征在于,所述斜率和持续时间配合分析具体包括以下步骤:
    获取持续时间参数:
    如果Slopes大于0,当前窗口内电阻平均值大于β2*MinHistory,则对恢复持续时间UpDuration增加采集一个窗口电阻需要的时间;
    如果Slopes小于0,且当前窗口内电阻均值小于α2*MaxHistory,则兴奋持续时间DownDuration增加一个窗口对应的时间;
    根据持续时间参数更新情绪状态值:
    若Slopes大于0,且UpDuration大于RecoverTime,则情绪状态值减小XXX;
    若Slopes小于0,且DownDuration大于ExcitTime,则情绪状态值增加XXX;
    其中RecoverTime表示恢复时间,ExcitTime表示兴奋时间,取值范围均为(0,200s)。
  7. 根据权利要求6所述的基于人体皮肤电阻变化的情绪检测方法,其特征在于,所述RecoverTime设置为8s,ExcitTime设置为5s。
  8. 根据权利要求6所述的基于人体皮肤电阻变化的情绪检测方法,其特征在于,所述斜率、持续时间和阈值配合分析具体包括以下步骤:
    若Slopes大于0,且UpDuration小于时间长度RecoverTime,则情绪状态值保持;否则若UpDuration大于时间长度RecoverTime,且当前电阻值大于β1*MinHistory,则情绪状态值减小ZZZ,否则情绪状态值减小XXX;
    若Slopes小于0,且DownDuration小于时间长度ExcitTime,则情绪状态值保持;若DownDuration超过时间长度ExcitTime,且当前电阻值小于α1*MaxHistory,则情绪状态值增加ZZZ,否则情绪状态值增加XXX。
  9. 根据权利要求1-8中任一项所述的基于人体皮肤电阻变化的情绪检测方法,其特征在于,所述电阻为电导的倒数,通过人体皮肤电导变化检测情绪。
  10. 一种基于人体皮肤电阻变化的情绪检测系统,用于实施如权利要求1-9中任一项所述的方法,其特征在于,包括依次连接的医疗极片、桥式电阻/电导测量电路、放大电路、A/D转换电路、CPU和人机交互界面。
PCT/CN2015/073954 2014-04-01 2015-03-10 一种基于人体皮肤电阻变化的情绪检测方法及系统 WO2015149606A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
SG11201600041PA SG11201600041PA (en) 2014-04-01 2015-03-10 Method and system for analyzing human emotional state based on human skin resistance measurement

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410128494.1A CN103892830B (zh) 2014-04-01 2014-04-01 一种基于人体皮肤电阻变化的情绪检测方法及系统
CN201410128494.1 2014-04-01

Publications (1)

Publication Number Publication Date
WO2015149606A1 true WO2015149606A1 (zh) 2015-10-08

Family

ID=50984646

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/073954 WO2015149606A1 (zh) 2014-04-01 2015-03-10 一种基于人体皮肤电阻变化的情绪检测方法及系统

Country Status (3)

Country Link
CN (1) CN103892830B (zh)
SG (1) SG11201600041PA (zh)
WO (1) WO2015149606A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021080451A1 (ru) * 2019-10-23 2021-04-29 Акционерное общество "Нейротренд" Способ анализа эмоционального восприятия аудиовизуального контента у группы пользователей

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103892830B (zh) * 2014-04-01 2016-01-20 新加坡情桥私人有限公司 一种基于人体皮肤电阻变化的情绪检测方法及系统
CN104545911B (zh) * 2015-01-29 2017-02-22 姚健欣 一种检测人体生理状态转变的方法及系统
CN106974652A (zh) * 2016-01-19 2017-07-25 大连民族大学 身心平衡测试仪及其测试方法
CN106580322A (zh) * 2016-12-15 2017-04-26 芯海科技(深圳)股份有限公司 一种具有高重复性的人体电阻抗测量方法
CN109394212A (zh) * 2017-08-16 2019-03-01 厦门智汇权科技有限公司 检测人体表面电气信号的方法及应用其的穿戴装置
CN107854221B (zh) * 2017-11-30 2023-05-26 北京蓝天腾麟医疗科技有限公司 一种具有多感应点的可联网智能尿布及其监测方法
CN110101398A (zh) * 2018-11-29 2019-08-09 华南理工大学 一种检测情绪的方法和系统
CN111248910B (zh) * 2020-01-20 2022-07-01 北京津发科技股份有限公司 一种基于多通道的皮肤电阻测量方法
CN112057090B (zh) * 2020-09-04 2021-10-29 浙江大学 基于体表极低频电势差特征的情绪判断穿戴式设备和方法

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN2569745Y (zh) * 2002-09-19 2003-09-03 昌祐精密工业股份有限公司 情绪检测器
WO2006106408A2 (en) * 2005-04-04 2006-10-12 Icon Orbital Technologies Corporation System and method for estimation of human activity and mood states based on rapid eye movements and galvanic skin response
CN101366627A (zh) * 2008-09-26 2009-02-18 鲁东大学 一种能提示血压测量结果可信度的电子血压计
CN101610716A (zh) * 2007-02-13 2009-12-23 皇家飞利浦电子股份有限公司 用于测量用户的唤醒的计算机程序产品、设备和方法
CN102512169A (zh) * 2011-12-31 2012-06-27 重庆邮电大学 一种皮肤电信号测试顶偏自适应控制方法和装置
CN102920444A (zh) * 2012-11-20 2013-02-13 浙江大学 便携式同时检测皮肤温度、皮肤电阻和脉搏的装置及方法
CN102986200A (zh) * 2010-07-01 2013-03-20 诺基亚公司 对用户的情绪状态改变进行响应
US20130317318A1 (en) * 2012-05-25 2013-11-28 Qualcomm Incorporated Methods and devices for acquiring electrodermal activity
CN103892830A (zh) * 2014-04-01 2014-07-02 新加坡情桥私人有限公司 一种基于人体皮肤电阻变化的情绪检测方法及系统

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101969850B (zh) * 2008-03-14 2016-01-20 皇家飞利浦电子股份有限公司 用于维持对象的状态的方法和系统
WO2009147625A1 (en) * 2008-06-06 2009-12-10 Koninklijke Philips Electronics N.V. Method of obtaining a desired state in a subject
TW201404359A (zh) * 2012-07-18 2014-02-01 Univ Nat Taiwan 情緒亢奮度量測系統及方法

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN2569745Y (zh) * 2002-09-19 2003-09-03 昌祐精密工业股份有限公司 情绪检测器
WO2006106408A2 (en) * 2005-04-04 2006-10-12 Icon Orbital Technologies Corporation System and method for estimation of human activity and mood states based on rapid eye movements and galvanic skin response
CN101610716A (zh) * 2007-02-13 2009-12-23 皇家飞利浦电子股份有限公司 用于测量用户的唤醒的计算机程序产品、设备和方法
CN101366627A (zh) * 2008-09-26 2009-02-18 鲁东大学 一种能提示血压测量结果可信度的电子血压计
CN102986200A (zh) * 2010-07-01 2013-03-20 诺基亚公司 对用户的情绪状态改变进行响应
CN102512169A (zh) * 2011-12-31 2012-06-27 重庆邮电大学 一种皮肤电信号测试顶偏自适应控制方法和装置
US20130317318A1 (en) * 2012-05-25 2013-11-28 Qualcomm Incorporated Methods and devices for acquiring electrodermal activity
CN102920444A (zh) * 2012-11-20 2013-02-13 浙江大学 便携式同时检测皮肤温度、皮肤电阻和脉搏的装置及方法
CN103892830A (zh) * 2014-04-01 2014-07-02 新加坡情桥私人有限公司 一种基于人体皮肤电阻变化的情绪检测方法及系统

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021080451A1 (ru) * 2019-10-23 2021-04-29 Акционерное общество "Нейротренд" Способ анализа эмоционального восприятия аудиовизуального контента у группы пользователей

Also Published As

Publication number Publication date
SG11201600041PA (en) 2016-02-26
CN103892830A (zh) 2014-07-02
CN103892830B (zh) 2016-01-20

Similar Documents

Publication Publication Date Title
WO2015149606A1 (zh) 一种基于人体皮肤电阻变化的情绪检测方法及系统
EP3358485A1 (en) General noninvasive blood glucose prediction method based on timing analysis
JP6190466B2 (ja) 生体信号測定器及び接触状態推定方法
CN107028603B (zh) 使用脉搏触诊信号来检测人体内的糖尿病的装置和方法
KR101025510B1 (ko) 감성인식장치의 개인별 최적화시스템 및 그 최적화 방법
CN107874750B (zh) 脉率变异性和睡眠质量融合的心理压力监测方法及装置
RU2018134296A (ru) Основанный на вероятности коэффициент усиления контроллера
ES2396844A1 (es) Sistema y método para la estimación simultánea y no invasiva de la glucosa en la sangre, nivel de glucocorticoides y presión arterial
CN111513730B (zh) 基于多通道生理数据的心理压力预测方法和系统
JP2016063995A (ja) 筋組織系及び神経系の活動量及び活動効率の評価システム及び評価方法並びにそのプログラム
JP2009232990A (ja) 状態評価装置、及び、状態評価プログラム
Das et al. Arduino-based noise robust online heart-rate detection
CN104487871B (zh) 生物体感知器以及省电模式设定方法
CN109394248B (zh) 驾驶疲劳检测方法及系统
CN104757938A (zh) 基于近红外光谱技术的皮肤油脂含量检测方法
CN111613338B (zh) 棘慢复合波检测模型构建方法、系统
Rachakonda et al. Stress-Lysis: An IoMT-enabled device for automatic stress level detection from physical activities
Jian-Jun et al. Decrease in Hurst exponent of human gait with aging and neurodegenerative diseases
CN106361327B (zh) 睡眠状态分析中清醒状态检测方法和系统
CN106344008B (zh) 睡眠状态分析中清醒状态检测方法和系统
Milici et al. System for highlighting the emotional states, used in assessing the teaching methods
WO2016119400A1 (zh) 一种检测人体生理状态转变的方法及系统
CN113331824B (zh) 一种基于激光测量的眨眼判断方法及装置
CN106510705B (zh) 一种人体体质年龄评估方法
CN107095668B (zh) 一种基于时域分析的自适应脑电波信号异常的检测方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15772695

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 09.03.2017)

122 Ep: pct application non-entry in european phase

Ref document number: 15772695

Country of ref document: EP

Kind code of ref document: A1