CN109472290A - Mood swing model analysis method based on finite state machine - Google Patents

Mood swing model analysis method based on finite state machine Download PDF

Info

Publication number
CN109472290A
CN109472290A CN201811181916.6A CN201811181916A CN109472290A CN 109472290 A CN109472290 A CN 109472290A CN 201811181916 A CN201811181916 A CN 201811181916A CN 109472290 A CN109472290 A CN 109472290A
Authority
CN
China
Prior art keywords
mood
state machine
finite state
emotional
finite
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
CN201811181916.6A
Other languages
Chinese (zh)
Other versions
CN109472290B (en
Inventor
林巧民
潘敏
赵慧娟
丁楠
徐康
叶宁
王汝传
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Post and Telecommunication University
Original Assignee
Nanjing Post and Telecommunication University
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 Nanjing Post and Telecommunication University filed Critical Nanjing Post and Telecommunication University
Priority to CN201811181916.6A priority Critical patent/CN109472290B/en
Publication of CN109472290A publication Critical patent/CN109472290A/en
Application granted granted Critical
Publication of CN109472290B publication Critical patent/CN109472290B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • G06F18/241Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/21Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214Generating training patterns; Bootstrap methods, e.g. bagging or boosting

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The mood swing model analysis method based on finite state machine that present invention discloses a kind of, comprising: S1, acquire physiological data using sign acquisition equipment, carry out feature extraction;S2, feature is subjected to mood Classification and Identification with support vector machine classifier, obtains the mood in each period;S3, description affective state transfer process, construct simultaneously estimated state sequence;S4, mental emotion questionnaire survey is carried out, affective state sequence sets is calculated;S5, the questionnaire survey of PAD Affect Scale is carried out, result is obtained into mood sample after the training of PAD category of model, marks emotional state, exercise supervision learning training;S6, pass through emotional state sequence sets, modal analysis results and supervised learning training result, prediction of the completion to mood swing, mood quality.This method can be matched with existing method, realize subjective and objective both sides fusion, breached the relationship between single emotional physiological signal and emotion and only considered the study limitation of objective factor.

Description

Mood swing model analysis method based on finite state machine
Technical field
The present invention relates to a kind of analysis methods of mood swing model, are based on finite state machine in particular to one kind Mood swing model analysis and emotional prediction method, belong to the affection computation field in artificial intelligence.
Background technique
PAD three-dimensional emotion model is a kind of more common model using dimensional space mark affective state, in audio-video There is relatively broad application in the fields such as speech synthesis, affection computation.PAD three-dimensional emotion model is equal to 1974 by Mehrabian Year is put forward for the first time comprising three pleasant degree, activity and dominance dimensions.PAD three-dimensional emotion model mainly has following characteristics: In PAD model, each emotion all uniquely corresponds to a PAD spatial coordinate location.After PAD parameter normalization, emotion can To be identified with unique three-dimensional coordinate, the evaluation with high confidence level.In PAD model, pass through one group of standard emotion scale The determination of PAD parameter coordinate is completed, the independence between each dimension of PAD easier can distinguish the text for being located at different emotions dimension This emotion.
In existing research, realization mainly includes the following three types method: 1, unsupervised learning to the identification of emotional state Method.The training sample of no concept label (classification) is learnt, to find the structured knowledge of training sample concentration, depth Learning network (DLN) is exactly typical unsupervised learning.Using layered structure similar with neural network, become by layer-by-layer feature It changes, the character representation by sample in former space transforms to a new feature space, to make to classify or predict to be more easier.With people The method of work rule construct feature is compared, and deep learning, come learning characteristic, can more portray the abundant interior of data using big data In information.2, supervised learning method.The training sample of (classification) is marked to learn to concept, as far as possible to training sample (classification) prediction is marked in data outside collection.Common supervised learning method have support vector machines, neural network, decision tree, Bayesian network, K- neighbour and hidden Markov model etc..3, semi-supervised learning method.This learning method realizes supervision The combination of study and unsupervised learning considers how originally to be trained and divide using a small amount of mark sample and a large amount of non-standard specimen Class.Specifically, affection computation is the computation model established by constructing emotional state, it carries out behavior and physiological signal Analysis, and mood interaction is established between man-machine based on the emotional state measured.Its most important link is that " mood is known Not (Emotion Recognition) " is i.e. by the behavior of user and physiological reaction come the corresponding emotional state of predictive estimation.
In general, studied currently based on the Emotion identification of physiological signal it is very much, but by various factors, as stimulation takes, The influence for inducing classification, acquisition equipment, feature extracting method, different dimensionality reductions and sorting algorithm of mood etc., between each research Recognition accuracy otherness is very big, is difficult to be compared for use.
Finite state machine (FSM) is a kind of calculating process and certain class of languages and the meter that takes out to study finite state Model is calculated, a finite state machine includes following components: a finite state collection, for the different shapes in description system State;One input set, the different input information received for indicating system;One node transition rule collection, for stating system System is transferred to the rule of another state when receiving different inputs from a state.Wherein finite state machine mainly includes two Class is limited receiver and finite transducer respectively.Finite state machine is that have a kind of discrete mathematics for outputting and inputting system Model, and the corresponding relationship between mood and environmental stimuli is also discrete, and the transfer of affective state is also limited From a kind of specific state to the transformation of another state in affective state space.Affective state based on FSM shifts individual emotion Activation it is not only related with extraneous input stimulus, it is also related with the personality of current human and psychology.This intrinsic property with it is limited The function description of state machine is identical.
In conclusion how to propose a kind of mood swing model based on finite state machine point on the basis of existing technology Analysis and emotional prediction method effectively realize the analysis to mood swing model, complete emotional prediction, also just become skill in the industry The new research direction of art personnel.
Summary of the invention
In view of the prior art there are drawbacks described above, the invention proposes a kind of mood swing model based on finite state machine Analysis method includes the following steps:
S1, the physiological data that tested personnel is acquired using sign acquisition equipment, physiological data obtained is uploaded in real time, And carry out feature extraction;
S2, the feature extracted is subjected to mood Classification and Identification with support vector machine classifier, obtained in each period Mood;
S3, affective state transfer process is described by finite state machine, construct status switch, and estimate to status switch Meter;
S4, mental emotion questionnaire survey is carried out to tested personnel, obtains the mental emotion age of tested personnel, passes through calculating Obtain affective state sequence sets;
S5, the questionnaire survey of PAD Affect Scale is carried out to tested personnel, will fill in the obtained result of questionnaire by PAD mould Mood sample is obtained after type classification based training, and the emotional state in mood sample is labeled, and then mood sample is carried out Supervised learning training;
S6, by the analysis of emotional state sequence sets and mood swing model as a result, and combining the supervision of mood sample Training result is practised, the prediction to following mood swing and mood quality is completed.
Preferably, the sign acquisition equipment includes skin electric transducer and heart rate sensor.
Preferably, affective state transfer process described in S3 include emotional state initial until emotional state termination.
Preferably, status switch is constructed described in S3, and status switch is estimated, is specifically comprised the following steps:
S31, the finite state machine triple for introducing I, S and M, wherein
I indicates input environmental stimuli (i1,i2,i3,i4,…,in) finite nonempty set,
S indicates the emotional state (s of individual1,s2,s3,s4,…,sn) finite nonempty set;
S32, by emotional state be divided into it is glad, angry, surprised, detest, frightened, tranquil and seven kinds sad, and by every kind of feelings Not-ready status is sub-divided into slight, general and very three kinds according to degree;
S33, it selects environmental stimuli time span and itself psychological quality as input set, finite state is shifted and is advised Collection introduces intensity factor K, and the size of the K is related to environmental stimuli intensity and itself psychological quality, the value model of the K It is trapped among in 0 to 1.
Preferably, affective state sequence sets are obtained by calculation described in S4, specifically comprise the following steps:
S41, foundation calculation formula, feeling quotrient=(mental emotion age/actual age) * 100% calculate feeling quotrient yt
S42, by feeling quotrient ytAnd the obtained mental emotion age is applied in combinatorial formula, obtains mood swing model
Wherein, c is a constant, c > 0, αpIndicate the mood at pth moment to current emotional effect, utFor white noise Sound process, if utIt allows for personal self the counterattack adaptability to changes for treating various fluctuations, be a variable related to time, That is utIt changes with time,WithIt is linear related;
Parameters and conditional variance in S43, solution mood swing model, then describe to obtain by finite state machine Another group of affective state sequence summarizes to obtain affective state sequence sets.
Preferably, the training of supervised learning described in S5 includes the following steps: using the artificial neuron in supervised learning method Network learns mood sample, then to training sample set outside data be marked, prediction of classifying.
Compared with prior art, advantages of the present invention is mainly reflected in the following aspects:
The mood swing model based on finite state machine in the present invention can be matched with existing Emotion identification method, Subjective and objective both sides fusion is realized, to breach the relationship between single emotional physiological signal and emotion and only consider visitor The study limitation of sight factor.The present invention, which can according to need, formulates personal mood volatility model, to prevent and avoid bad mood institute Bring side effect.Mood can also be carried out more careful classification by the finite state machine arrived used in the present invention, be realized Subtle mood is more accurately described.
Meanwhile the solution of the present invention will acquire the affective state sequence and mood swing analysis that equipment obtains by sign Affective state sequence out is effectively combined, to more accurately observe the fluctuation situation of mood and realize Prediction to the following mood swing.
In addition, the present invention also provides reference for other relevant issues in same domain, can be opened up on this basis Extension is stretched, and applies to have very wide in same domain in the technical solution of other mood swing model analysis and emotional prediction Application prospect.
Just attached drawing in conjunction with the embodiments below, the embodiment of the present invention is described in further detail, so that of the invention Technical solution is more readily understood, grasps.
Detailed description of the invention
Fig. 1 is transference process schematic in the present invention;
Fig. 2 is flow diagram of the invention.
Specific embodiment
As shown in FIG. 1 to FIG. 2, the mood swing model analysis method based on finite state machine that present invention discloses a kind of, Include the following steps:
S1, the physiological data that tested personnel is acquired using sign acquisition equipment, physiological data obtained is uploaded in real time, And carry out feature extraction.The sign acquisition equipment includes skin electric transducer and heart rate sensor.
S2, the feature extracted is subjected to mood Classification and Identification with support vector machines (SVM) classifier, obtains each time Mood in section.
S3, affective state transfer process is described by finite state machine, construct status switch, and estimate to status switch Meter.The affective state transfer process include emotional state initial until emotional state termination.
The building status switch, and status switch is estimated, specifically comprise the following steps:
S31, the finite state machine triple for introducing I, S and M, wherein
I indicates input environmental stimuli (i1,i2,i3,i4,…,in) finite nonempty set,
S indicates the emotional state (s of individual1,s2,s3,s4,…,sn) finite nonempty set,
M is expressed as a fuzzy subset on product space I*S, the referred to as fuzzy relation from I to S.
S32, by emotional state be divided into it is glad, angry, surprised, detest, frightened, tranquil and seven kinds sad, and by every kind of feelings Not-ready status is sub-divided into slight, general and very three kinds according to degree.
Such as the states such as anger, happiness are subdivided into slight angry, general angry and as mad as a wet hen.Then S is represented by (slight angry, general angry, as mad as a wet hen, slight glad, general glad, very delight ... ...) totally 21 states.People's The variation of emotional state, and environmental stimuli, the psychology of itself and personality, and the time span of stimulation etc. have very big pass System.
S33, it selects environmental stimuli time span and itself psychological quality as input set, finite state is shifted and is advised Collection introduces intensity factor K, and the size of the K is related to environmental stimuli intensity and itself psychological quality, the value model of the K It is trapped among in 0 to 1.
It is happiness in the case where having obtained specific mood by SVM classifier, such as through the mood after feature extraction, then It is slight glad when 0 < K≤0.33.It is general glad as 0.33 < K≤0.66.It is very delight when 0.66 < K≤1.
S4, mental emotion questionnaire survey is carried out to tested personnel, obtains the mental emotion age of tested personnel, passes through calculating Obtain affective state sequence sets.
It is described that affective state sequence sets are obtained by calculation, specifically comprise the following steps:
S41, foundation calculation formula, feeling quotrient=(mental emotion age/actual age) * 100%, feeling quotrient is Emotional Intelligence Referred to as, the fluctuation with mood is fluctuated to be positively correlated.Calculate feeling quotrient yt
S42, by feeling quotrient ytAnd the obtained mental emotion age is applied in combinatorial formula, obtains mood swing model
Wherein, c is a constant, c > 0, αpIndicate the mood at pth moment to current emotional effect, utFor white noise Sound process, if utIt allows for personal self the counterattack adaptability to changes for treating various fluctuations, be a variable related to time, That is utIt changes with time,WithIt is linear related.
Parameters and conditional variance in S43, solution mood swing model, then describe to obtain by finite state machine Another group of affective state sequence summarizes to obtain affective state sequence sets.
The solution procedure of mood swing model described further below,
yt=c+ α1yt-12yt-2+…+αpyt-p+ut, (1)
Wherein, c is a constant, c > 0, αpIndicate the mood at pth moment to current emotional effect.In equation utFor white-noise process, if this variable allows for personal self the counterattack adaptability to changes for treating various fluctuations, i.e., it is a kind of self Adjust control ability.utIt is a variable related to time, i.e. utIt changes with time.One method is exactly table It is brightWithIt is linearly related, it may be assumed that
It can to sum up obtain,
Because of utIt is random, andIt is unlikely to be negative, ξ > 0,0≤β12+…+βm<1.When these conditions meet When,Unconditional expectation may be expressed as:
This explanationUnconditional expectation assuming that under conditions of be constant, and can be obtained by estimation. (5) formula shows mood swing meeting for the previous period so that mood swing today is still very big, unless there are other reverse impacts It offsets, similarly, for the previous period even-tempered and good-humoured is but also the mood Anomaly quiescence of today, unless occurring big impact suddenly So thatSuddenly increase.In estimation, we be can be considered (5) formula in another form, that is, be assumed:
Wherein { vtIt is the sequence for being independently distributed and having zero-mean and unit variance.If met:
So just have
This is exactly (5) formula, and this variation shows if utMeet (6) and (7) formula to build just now then (5) formula is just set up Vertical conditional expectation is constant.
In estimation, in order to which simplification (1) formula is rewritten into following form:
Wherein, x 'tIt is the vector for indicating all explanatory variables, it is clear that it includes the lagged value of y, utStill full The condition of foot (2) formula.M observation for taking away the beginning is condition (t=-m+1 ,-m+2 ..., 0).With t=1, the sight of 2 ... ..., T Measured value, which is done, to be estimated.Enable YtThe vector for representing the observation obtained in time t, then have
Yt=(yt,yt-1,…,y0,…,y-m+1,x′t,…,x′1,x′0,…x′m+1)′。
Assuming that vt~i.i.d N (0,1), i.e. vtIndependent same distribution process is submitted to, then yt|t-1~N (x 't,β,ht), because This
Wherein, δ=(ξ, β1,…,βm)′,
All parameters to be estimated are formed into vector θ: θ=(β ', δ ') ', because of m logarithm conditional likelihood before this It is:
(12) formula differentiates to θ, can take first by (10) formula to θ derivation here, then the mode summed carries out, so It obtains:
It can be by asking minimax bounding method to find out parameter θ, therefore the parameter in mood swing model with (14) formula It can find out, conditional variance can also be found out from (10) formula.
S5, the questionnaire survey of PAD Affect Scale is carried out to tested personnel, will fill in the obtained result of questionnaire by PAD mould Mood sample is obtained after type classification based training, and the emotional state in mood sample is labeled, and then mood sample is carried out Supervised learning training.Mood sample is learnt using artificial neural network common in supervised learning method, it is then right Data outside training sample set are marked, prediction of classifying.
S6, by the analysis of emotional state sequence sets and mood swing model as a result, and combining the supervision of mood sample Training result is practised, the prediction to following mood swing and mood quality is completed.Pass through prolonged data accumulation and data Analysis is compared, and the mood that the emotional state sequence sets and mood swing model obtained by sign acquisition device measuring calculate is poor Different smaller, two ways is effectively combined the mood of the fluctuation situation that can more accurately observe mood and prediction future Fluctuation.
The mood swing model based on finite state machine in the present invention can be matched with existing Emotion identification method, Subjective and objective both sides fusion is realized, to breach the relationship between single emotional physiological signal and emotion and only consider visitor The study limitation of sight factor.The present invention, which can according to need, formulates personal mood volatility model, to prevent and avoid bad mood institute Bring side effect.Mood can also be carried out more careful classification by the finite state machine arrived used in the present invention, be realized Subtle mood is more accurately described.
Meanwhile the solution of the present invention will acquire the affective state that equipment (skin pricktest and heart rate sensor) obtains by sign Sequence and mood swing analyze the affective state sequence come and are effectively combined, to more accurately observe The fluctuation situation of mood simultaneously realizes the prediction to the following mood swing.
In addition, the present invention also provides reference for other relevant issues in same domain, can be opened up on this basis Extension is stretched, and applies to have very wide in same domain in the technical solution of other mood swing model analysis and emotional prediction Application prospect.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie In the case where without departing substantially from spirit and essential characteristics of the invention, the present invention can be realized in other specific forms.Therefore, no matter From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims Variation is included within the present invention, and any reference signs in the claims should not be construed as limiting the involved claims.
In addition, it should be understood that although this specification is described in terms of embodiments, but not each embodiment is only wrapped Containing an independent technical solution, this description of the specification is merely for the sake of clarity, and those skilled in the art should It considers the specification as a whole, the technical solutions in the various embodiments may also be suitably combined, forms those skilled in the art The other embodiments being understood that.

Claims (6)

1. a kind of mood swing model analysis method based on finite state machine, characterized by the following steps:
S1, the physiological data that tested personnel is acquired using sign acquisition equipment, physiological data obtained is uploaded in real time, is gone forward side by side Row feature extraction;
S2, the feature extracted is subjected to mood Classification and Identification with support vector machine classifier, obtains the feelings in each period Thread;
S3, affective state transfer process is described by finite state machine, construct status switch, and estimate status switch;
S4, mental emotion questionnaire survey is carried out to tested personnel, obtains the mental emotion age of tested personnel, is obtained by calculation Affective state sequence sets;
S5, the questionnaire survey of PAD Affect Scale is carried out to tested personnel, will fill in the obtained result of questionnaire by PAD model point Mood sample is obtained after class training, and the emotional state in mood sample is labeled, is then exercised supervision to mood sample Learning training;
S6, by the analysis of emotional state sequence sets and mood swing model as a result, and the supervised learning of mood sample is combined to instruct Practice as a result, completing the prediction to following mood swing and mood quality.
2. the mood swing model analysis method according to claim 1 based on finite state machine, it is characterised in that: described It includes skin electric transducer and heart rate sensor that sign, which acquires equipment,.
3. the mood swing model analysis method according to claim 1 based on finite state machine, it is characterised in that: in S3 The affective state transfer process include emotional state initial until emotional state termination.
4. the mood swing model analysis method according to claim 1 based on finite state machine, which is characterized in that in S3 The building status switch, and status switch is estimated, specifically comprise the following steps:
S31, the finite state machine triple for introducing I, S and M, wherein
I expression input environmental stimuli () finite nonempty set,
The individual emotional state of S expression () finite nonempty set,
M is expressed as a fuzzy subset on product space I*S, the referred to as fuzzy relation from I to S;
S32, by emotional state be divided into it is glad, angry, surprised, detest, frightened, tranquil and seven kinds sad, and by every kind of mood shape State is sub-divided into slight, general and very three kinds according to degree;
S33, select environmental stimuli time span and itself psychological quality as input set, for finite state transition rule Collection introduces intensity factor K, and the size of the K is related to environmental stimuli intensity and itself psychological quality, the value model of the K It is trapped among in 0 to 1.
5. the mood swing model analysis method according to claim 1 based on finite state machine, which is characterized in that in S4 It is described that affective state sequence sets are obtained by calculation, specifically comprise the following steps:
S41, foundation calculation formula, feeling quotrient=(mental emotion age/actual age) * 100% calculate feeling quotrient
S42, by feeling quotrientAnd the obtained mental emotion age is applied in combinatorial formula, obtains mood swing model
,
Wherein, c is a constant, c0,Indicate the mood at pth moment to current emotional effect,For white noise Process, ifIt allows for personal self the counterattack adaptability to changes for treating various fluctuations, be a variable related to time, i.e.,It changes with time,With ……,, linear related;
Parameters and conditional variance in S43, solution mood swing model, then describe to obtain another by finite state machine Group affective state sequence, summarizes to obtain affective state sequence sets.
6. the mood swing model analysis method according to claim 1 based on finite state machine, which is characterized in that in S5 Supervised learning training, include the following steps: using the artificial neural network in supervised learning method come to mood sample into Row study, then to training sample set outside data be marked, prediction of classifying.
CN201811181916.6A 2018-10-11 2018-10-11 Emotion fluctuation model analysis method based on finite-state machine Active CN109472290B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811181916.6A CN109472290B (en) 2018-10-11 2018-10-11 Emotion fluctuation model analysis method based on finite-state machine

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811181916.6A CN109472290B (en) 2018-10-11 2018-10-11 Emotion fluctuation model analysis method based on finite-state machine

Publications (2)

Publication Number Publication Date
CN109472290A true CN109472290A (en) 2019-03-15
CN109472290B CN109472290B (en) 2022-10-14

Family

ID=65663369

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811181916.6A Active CN109472290B (en) 2018-10-11 2018-10-11 Emotion fluctuation model analysis method based on finite-state machine

Country Status (1)

Country Link
CN (1) CN109472290B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110147432A (en) * 2019-05-07 2019-08-20 大连理工大学 A kind of Decision Search engine implementing method based on finite-state automata
CN111723867A (en) * 2020-06-22 2020-09-29 山东大学 Intelligent evaluation system and method for online game fascination degree
CN112263252A (en) * 2020-09-28 2021-01-26 贵州大学 PAD (PAD application aided differentiation) emotion dimension prediction method based on HRV (high resolution video) features and three-layer SVR (singular value representation)
CN112618911A (en) * 2020-12-31 2021-04-09 四川音乐学院 Music feedback adjusting system based on signal processing
CN114863636A (en) * 2022-03-25 2022-08-05 吉林云帆智能工程有限公司 Emotion recognition algorithm for rail vehicle driver

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106419936A (en) * 2016-09-06 2017-02-22 深圳欧德蒙科技有限公司 Emotion classifying method and device based on pulse wave time series analysis
CN107766898A (en) * 2017-12-08 2018-03-06 南京邮电大学盐城大数据研究院有限公司 The three classification mood probabilistic determination methods based on SVM
CN107944473A (en) * 2017-11-06 2018-04-20 南京邮电大学 A kind of physiological signal emotion identification method based on the subjective and objective fusion of multi-categorizer

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106419936A (en) * 2016-09-06 2017-02-22 深圳欧德蒙科技有限公司 Emotion classifying method and device based on pulse wave time series analysis
CN107944473A (en) * 2017-11-06 2018-04-20 南京邮电大学 A kind of physiological signal emotion identification method based on the subjective and objective fusion of multi-categorizer
CN107766898A (en) * 2017-12-08 2018-03-06 南京邮电大学盐城大数据研究院有限公司 The three classification mood probabilistic determination methods based on SVM

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110147432A (en) * 2019-05-07 2019-08-20 大连理工大学 A kind of Decision Search engine implementing method based on finite-state automata
CN110147432B (en) * 2019-05-07 2023-04-07 大连理工大学 Decision search engine implementation method based on finite state automaton
CN111723867A (en) * 2020-06-22 2020-09-29 山东大学 Intelligent evaluation system and method for online game fascination degree
CN112263252A (en) * 2020-09-28 2021-01-26 贵州大学 PAD (PAD application aided differentiation) emotion dimension prediction method based on HRV (high resolution video) features and three-layer SVR (singular value representation)
CN112263252B (en) * 2020-09-28 2024-05-03 贵州大学 PAD emotion dimension prediction method based on HRV features and three-layer SVR
CN112618911A (en) * 2020-12-31 2021-04-09 四川音乐学院 Music feedback adjusting system based on signal processing
CN112618911B (en) * 2020-12-31 2023-02-03 四川音乐学院 Music feedback adjusting system based on signal processing
CN114863636A (en) * 2022-03-25 2022-08-05 吉林云帆智能工程有限公司 Emotion recognition algorithm for rail vehicle driver

Also Published As

Publication number Publication date
CN109472290B (en) 2022-10-14

Similar Documents

Publication Publication Date Title
CN109472290A (en) Mood swing model analysis method based on finite state machine
Lotfian et al. Curriculum learning for speech emotion recognition from crowdsourced labels
Aracena et al. Neural networks for emotion recognition based on eye tracking data
JP6977901B2 (en) Learning material recommendation method, learning material recommendation device and learning material recommendation program
JP5520886B2 (en) Behavior model learning apparatus, method, and program
CN107153812A (en) A kind of exercising support method and system based on machine vision
O’Toole et al. A perceptual learning theory of the information in faces
US10610109B2 (en) Emotion representative image to derive health rating
KR101855168B1 (en) Emotion classification method based on deep learning and method thereof
CN111728590A (en) Individual cognitive ability prediction method and system based on dynamic function connection
CN113408852B (en) Meta-cognition ability evaluation model based on online learning behavior and deep neural network
CN110298303B (en) Crowd identification method based on long-time memory network glance path learning
Borges et al. Classifying confusion: autodetection of communicative misunderstandings using facial action units
CN115422973A (en) Electroencephalogram emotion recognition method of space-time network based on attention
CN110147822B (en) Emotion index calculation method based on face action unit detection
Bernstein et al. Using deep learning for alcohol consumption recognition
Zheng et al. A data-driven emotion model for english learners based on machine learning
CN112464774A (en) Emotion identification method in video learning based on eye movement measurement
Dapogny et al. On Automatically Assessing Children's Facial Expressions Quality: A Study, Database, and Protocol
Li et al. ELM-based classification of ADHD patients using a novel local feature extraction method
Ahmadieh et al. Visual image reconstruction based on EEG signals using a generative adversarial and deep fuzzy neural network
Liliana et al. The Fuzzy Emotion Recognition Framework Using Semantic-Linguistic Facial Features
CN112396114A (en) Evaluation system, evaluation method and related product
González-Meneses et al. Automatic Recognition of Learning-Centered Emotions
Anbusegaran Unobtrusive Assessment Of Student Engagement Levels In Online Classroom Environment Using Emotion Analysis

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant