CN108039181A - The emotion information analysis method and device of a kind of voice signal - Google Patents
The emotion information analysis method and device of a kind of voice signal Download PDFInfo
- Publication number
- CN108039181A CN108039181A CN201711065483.3A CN201711065483A CN108039181A CN 108039181 A CN108039181 A CN 108039181A CN 201711065483 A CN201711065483 A CN 201711065483A CN 108039181 A CN108039181 A CN 108039181A
- Authority
- CN
- China
- Prior art keywords
- expressed
- information
- emotion
- emotion information
- voice signal
- 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
Links
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
- G10L25/63—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination for estimating an emotional state
Abstract
An embodiment of the present invention provides the emotion information analysis method and device of a kind of voice signal, when its method is included in the emotion information expressed by the voice signal that analysis user sends, text message and speech parameter information in voice signal are extracted；Text emotion is carried out to text information to analyze to obtain the emotion information expressed by text information, and speech emotional is carried out to the speech parameter information and analyzes to obtain the emotion information expressed by the speech parameter；The emotion information expressed by emotion information and the speech parameter information according to expressed by text information obtains the expressed emotion information of the voice signal.The embodiment of the present invention can improve the accuracy of the emotion information expressed by definite voice signal.
Description
Technical field
The present invention relates to field of computer technology, more particularly to the emotion information analysis method and dress of a kind of voice signal
Put.
Background technology
People can express various emotion informations when speaking, such as, glad, angry, shock, it is sad and in
Property etc..
With the rapid development of technology, intelligent sound interactive terminal is widely used, and more and more enterprises utilize
Intelligent sound interactive terminal services to provide a user, in order to improve service quality during service is provided a user, intelligence
Energy interactive voice terminal generally requires the voice signal emotion to be expressed that analysis user sends.
Wherein, in the prior art, the voice signal that intelligent sound interactive terminal can be sent according to user analyzes the sound
Emotion information expressed by sound signal, such as size, intonation and word speed of sound when being spoken by user etc. determine use
Emotion information expressed by family.For example, user is very angry at this time, " your this way is said with loud, quick and high intonation
Make the people very indignant " with the angry emotion information of expression, sound when intelligent sound interactive terminal says this word according to user is big
It is very angry at this time that small, word speed and intonation analyze user.
However, it is found by the inventors that if user is very angry at this time, but user but says " you with the more tranquil tone
This way makes the people very indignant ", the not up to angry standard of sound size, intonation and word speed when being spoken due to user,
Therefore, the emotion information expressed by the words will not be determined as anger by intelligent sound interactive terminal, and is likely in being determined as
Property, so as to definite mistake occur, cause to determine that the accuracy of emotion information expressed by voice signal that user sends is relatively low.
The content of the invention
The embodiment of the present invention the technical problem to be solved is that：Determine the emotion letter expressed by the voice signal that user sends
The accuracy of breath is relatively low.
In order to improve the accuracy of the emotion information expressed by the voice signal that definite user sends, the embodiment of the present invention carries
A kind of sentiment analysis method and apparatus of voice signal are supplied.
In a first aspect, an embodiment of the present invention provides a kind of sentiment analysis method of voice signal, the described method includes：
Extract the text message and speech parameter information in voice signal；
Text emotion is carried out to the text message to analyze to obtain the emotion information expressed by the text message；
The emotion that speech emotional is analyzed to obtain expressed by the speech parameter information is carried out to the speech parameter information to believe
Breath；
The emotion information expressed by emotion information and the speech parameter information according to expressed by the text message obtains
Take the emotion information expressed by the voice signal.
Wherein, the emotion for analyzing to obtain expressed by the text message to text message progress text emotion is believed
Breath, including：
Text emotion analysis is carried out to the text message using LSTM algorithms, is obtained each expressed by the text message
The probable value of a emotion information.
Wherein, it is described that speech parameter information progress speech emotional is analyzed to obtain expressed by the speech parameter information
Emotion information, including：
Speech emotional analysis is carried out to the speech parameter using CNN algorithms, is obtained each expressed by the speech parameter
The probable value of a emotion information.
Wherein, the emotion information according to expressed by the text message and the feelings expressed by the speech parameter information
Feel the emotion information expressed by voice signal described in acquisition of information, including：
For each emotion information, the probable value of the emotion information according to expressed by the text message with it is described
The probable value of the emotion information expressed by speech parameter information, calculates the emotion information expressed by the voice signal
Combined chance value；
The highest emotion information of combined chance value is determined as to the expressed emotion information of the voice signal.
Wherein, the probable value of the emotion information according to expressed by the text message is believed with the speech parameter
The probable value of the expressed emotion information of breath, including：
Calculate between the probable value of the emotion information expressed by the text message and pre-set text emotion coefficient
First product；
Calculate the probable value of the emotion information expressed by the speech parameter information and default speech emotional coefficient it
Between the second product；
Calculate the 3rd product between first product and the default matrix-vector of the emotion information；
Calculate the 4th product between second product and the default matrix-vector of the emotion information；
The synthesis of the emotion according to expressed by the 3rd product, the 4th product obtain the voice signal is general
Rate value.
Second aspect, an embodiment of the present invention provides a kind of emotion information analytical equipment of voice signal, described device bag
Include：
Extraction module, for extracting text message and speech parameter information in voice signal；
First analysis module, analyzes to obtain expressed by the text message for carrying out text emotion to the text message
Emotion information；
Second analysis module, analyzes to obtain the speech parameter letter for carrying out the speech parameter information speech emotional
The expressed emotion information of breath；
Acquisition module, for expressed by the emotion information according to expressed by the text message and the speech parameter information
Emotion information obtain emotion information expressed by the voice signal.
Wherein, first analysis module is specifically used for：Text emotion is carried out to the text message using LSTM algorithms
Analysis, obtains the probable value of each emotion information expressed by the text message.
Wherein, second analysis module is specifically used for：Speech emotional point is carried out to the speech parameter using CNN algorithms
Analysis, obtains the probable value of each emotion information expressed by the speech parameter.
Wherein, the acquisition module includes：
Computing unit, for for each emotion information, the emotion information according to expressed by the text message
Probable value and the speech parameter information expressed by the emotion information probable value, calculate expressed by the voice signal
The emotion information combined chance value；
Determination unit, for the highest emotion information of combined chance value to be determined as to the expressed feelings of the voice signal
Feel information.
Wherein, the computing unit includes：
First computation subunit, for calculating the probable value of the emotion information expressed by the text message with presetting
The first product between text emotion coefficient；
Second computation subunit, for calculate the probable value of the emotion information expressed by the speech parameter information with
The second product between default speech emotional coefficient；
3rd computation subunit, for calculating between first product and the default matrix-vector of the emotion information
3rd product；
4th computation subunit, for calculating between second product and the default matrix-vector of the emotion information
4th product；
Subelement is obtained, expressed by obtaining the voice signal according to the 3rd product, the 4th product
The combined chance value of the emotion information.
Compared with prior art, the embodiment of the present invention includes advantages below：
In embodiments of the present invention, in the emotion information expressed by the voice signal that analysis user sends, sound is extracted
Text message and speech parameter information in signal；Text emotion is carried out to text information to analyze to obtain text information institute table
The emotion information reached, and the emotion that speech emotional is analyzed to obtain expressed by the speech parameter is carried out to the speech parameter information and is believed
Breath；The emotion information expressed by emotion information and the speech parameter information according to expressed by text information obtains sound letter
Number expressed emotion information.
When determining the expressed emotion information of the voice signal, the prior art is only according to big in the voice signal
Small, intonation and word speed determine the expressed emotion information of the voice signal, and the embodiment of the present invention is believed according to the sound
Text message and speech parameter information in number determine the expressed emotion information of the voice signal.
Compared with the prior art, the embodiment of the present invention is in addition to according to speech parameter information, in combination with text
Information, more fully hereinafter analyzes the emotion information expressed by the voice signal, therefore can avoid the occurrence of the prior art
In erroneous judgement situation, therefore, the embodiment of the present invention can improve the accuracy of the emotion information expressed by definite voice signal.
Brief description of the drawings
Fig. 1 is a kind of step flow chart of the emotion information analysis method embodiment of voice signal of the present invention；
Fig. 2 is a kind of structure diagram of the emotion information analytical equipment embodiment of voice signal of the present invention.
Embodiment
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, it is below in conjunction with the accompanying drawings and specific real
Applying mode, the present invention is described in further detail.
Reference Fig. 1, shows a kind of step flow of the emotion information analysis method embodiment of voice signal of the present invention
Figure, specifically may include steps of：
In step S101, text message and speech parameter information in voice signal are extracted；
In embodiments of the present invention, can be carried using DNN (Deep Neural Network, deep neural network) algorithm
Take the text message and speech parameter information in the voice signal, alternatively, using LSTM (Long Short-Term Memory,
Shot and long term memory network) algorithm and CTC (Connectionist temporal classification, the classification of connection sequential)
Text message and speech parameter information in the model extraction voice signal.
Wherein, text message includes the content expressed by voice signal, for example, user says in short：" you so allow me
It is very angry ", " you so allow me very angry " this eight words can be the text message of the voice signal.
Speech parameter information includes speech speed, signal-to-noise ratio, voice size, tone, average pitch, the fundamental tone of voice signal
Scope and fundamental tone change etc..
In embodiments of the present invention, after voice signal is sent by the face and nose of user, in some frequencies
Signal strength can reduce, for example, the signal strength of high frequency treatment reduces, and less than the signal strength at low frequency, can so cause sound
Sound signal distortion, and then the accuracy of the emotion information expressed by voice signal can be reduced.Therefore, believe to improve definite sound
The accuracy of emotion information expressed by number is, it is necessary to detect signal strength of the voice signal in each frequency, when detecting
Signal strength can then strengthen signal strength on these frequencies when relatively low in some frequencies.
In an alternative embodiment of the invention, it is necessary to which voice signal is split as multiple short voice signals, soldier couple according to the time
Multiple short voice signals carry out short signal intensive analysis, in short-term zero passage analysis, in short-term average signal strength analysis, correlation respectively
Property analysis and average signal strength difference analysis, to determine the voiceless sound and voiced sound etc. in voice signal, in order to afterwards
Extract the speech parameter information of the voice signal.
Secondly, the environment where when user speaks is also typically present noise, and usual noise is existing always, and sound is believed
Number it is not existing always, therefore, it is necessary to detect whether there are voice signal, is detecting whether there are during voice signal, can be with
The starting point and ending point of voice signal is detected using the methods of double threshold diagnostic method, and then determines the voice signal, is kept away
Exempt from that excessive noise is mixed in the voice signal while is handled, it is possible to reduce the data volume of processing and time, secondly may be used also
The influence brought to avoid the sentiment analysis result of the noise on analysis voice signal, to improve the emotion of voice signal parsing knot
The accuracy of fruit.
In step s 102, the emotion that text emotion is analyzed to obtain expressed by text information is carried out to text information to believe
Breath；
In embodiments of the present invention, sentiment analysis can be carried out to text information text using LSTM algorithms, obtains text
The probable value of each emotion information expressed by this information, and the emotion information expressed by as text information.
Certainly, when carrying out text emotion analysis to text information, the embodiment of the present invention can also use other texts
Sentiment analysis method, the embodiment of the present invention do not limit the text emotion to being utilized during the progress text emotion analysis of text information
Analysis method.
In embodiments of the present invention, a variety of emotions can be locally located in advance in technical staff, for example, glad, angry, shake
It is frightened, sad, worried and neutral etc..In this way, after by analyzing text message, can obtain expressed by text information
Angry probable value, the glad probable value expressed by text information, the probable value of shock expressed by text information,
Worried probable value expressed by sad probable value, text information and text information institute expressed by text information
The neutral probable value of expression.
In step s 103, speech emotional is carried out to the speech parameter information to analyze to obtain the feelings expressed by the speech parameter
Feel information；
In embodiments of the present invention, calculated using CNN (Convolutional Neural Network, convolutional neural networks)
Method carries out speech emotional analysis to the speech parameter, obtains the probable value of each emotion information expressed by the speech parameter, and
As the emotion information expressed by the speech parameter information.
For example, obtain the angry probable value expressed by the speech parameter, the glad probability expressed by the speech parameter
Value, the probable value of shock expressed by the speech parameter, the sad probable value expressed by the speech parameter, the speech parameter institute
Neutral probable value expressed by the worried probable value of expression and the speech parameter.
Certainly, when carrying out speech emotional analysis to the speech parameter, the embodiment of the present invention can also use other voices
Sentiment analysis method, the embodiment of the present invention do not limit the speech emotional to being utilized during voice messaging progress text emotion analysis
Analysis method.
In step S104, the feelings expressed by emotion information and the speech parameter information according to expressed by text information
Emotion information expressed by the sense acquisition of information voice signal.
In embodiments of the present invention, can be according to this article for any one emotion in pre-set multiple emotions
The probable value of the emotion information expressed by this information and the probable value of the emotion information expressed by the speech parameter information, meter
Calculate the combined chance value of the emotion information expressed by the voice signal；For in pre-set multiple emotions other are each
A emotion, equally performs aforesaid operations, and the synthesis that can so respectively obtain each emotion expressed by the voice signal is general
The highest emotion information of combined chance value, is then determined as the expressed emotion information of the voice signal by rate value.
Wherein, the probable value of the emotion according to expressed by text information and the feelings expressed by the speech parameter information
The probable value of sense, the specific steps for calculating the combined chance value of the emotion expressed by the voice signal can be by following flow
Realize, including：
Calculate first between the probable value of the emotion information expressed by text information and pre-set text emotion coefficient
Product；Calculate second between the probable value of the emotion information expressed by the speech parameter information and default speech emotional coefficient
Product；Calculate the 3rd product between the first product and the default matrix-vector of the emotion information；Calculate the second product and the feelings
Feel the 4th product between the default matrix-vector of information；According to expressed by the 3rd product, the 4th product obtain the voice signal
The emotion information combined chance value.For example, by the 3rd product and the 4th product input tanh functions, sound letter is obtained
The combined chance value of the emotion information expressed by number.
Wherein, in embodiments of the present invention, presetting speech emotional coefficient can be identical with pre-set text emotion coefficient, also may be used
With difference.
Technical staff in advance can count the voice signal of substantial amounts of expression user feeling, count text message
The weight to show emotion is each able to speech parameter, if the weight that text message can show emotion is joined more than voice
The weight that number information can show emotion, then can set pre-set text emotion coefficient to be more than default speech emotional coefficient；If
The weight that text message can show emotion is less than the weight that speech parameter information can show emotion, then can set default text
This emotion coefficient is less than default speech emotional coefficient；If the weight that text message can show emotion is equal to speech parameter information
The weight that can be showed emotion, then can set pre-set text emotion coefficient to be equal to default speech emotional coefficient.Afterwards, will set
Good pre-set text emotion coefficient and default speech emotional coefficient is respectively stored in local so that in step S104 can directly from
It is local to obtain pre-set text emotion coefficient and default speech emotional coefficient, then calculate the emotion letter expressed by text information
The first product between the probable value and pre-set text emotion coefficient of breath；Calculate the emotion letter expressed by the speech parameter information
The second product between the probable value of breath and default speech emotional coefficient；Calculate the default matrix of the second product and the emotion information
The 4th product between vector；The emotion information according to expressed by the 3rd product, the 4th product obtain the voice signal it is comprehensive
Close probable value.For example, by the 3rd product and the 4th product input tanh functions, the emotion expressed by the voice signal is obtained
The combined chance value of information.
In embodiments of the present invention, in the emotion information expressed by the voice signal that analysis user sends, sound is extracted
Text message and speech parameter information in signal；Text emotion is carried out to text information to analyze to obtain text information institute table
The emotion information reached, and the emotion that speech emotional is analyzed to obtain expressed by the speech parameter is carried out to the speech parameter information and is believed
Breath；The emotion information expressed by emotion information and the speech parameter information according to expressed by text information obtains sound letter
Number expressed emotion information.
When determining the expressed emotion information of the voice signal, the prior art is only according to big in the voice signal
Small, intonation and word speed determine the expressed emotion information of the voice signal, and the embodiment of the present invention is believed according to the sound
Text message and speech parameter information in number determine the expressed emotion information of the voice signal.
Compared with the prior art, the embodiment of the present invention is in addition to according to speech parameter information, in combination with text
Information, more fully hereinafter analyzes the emotion information expressed by the voice signal, therefore can avoid the occurrence of the prior art
In erroneous judgement situation, therefore, the embodiment of the present invention can improve the accuracy of the emotion information expressed by definite voice signal.
It should be noted that for embodiment of the method, in order to be briefly described, therefore it is all expressed as to a series of action group
Close, but those skilled in the art should know, the embodiment of the present invention and from the limitation of described sequence of movement, because according to
According to the embodiment of the present invention, some steps can use other orders or be carried out at the same time.Secondly, those skilled in the art also should
Know, embodiment described in this description belongs to preferred embodiment, and the involved action not necessarily present invention is implemented
Necessary to example.
With reference to Fig. 2, a kind of structure diagram of the emotion information analytical equipment embodiment of voice signal of the present invention, tool are shown
Body can include following module：
Extraction module 11, for extracting text message and speech parameter information in voice signal；
First analysis module 12, analyzes to obtain text message institute table for carrying out text emotion to the text message
The emotion information reached；
Second analysis module 13, analyzes to obtain the speech parameter for carrying out speech emotional to the speech parameter information
Emotion information expressed by information；
Acquisition module 14, for the emotion information according to expressed by the text message and speech parameter information institute table
The emotion information reached obtains the emotion information expressed by the voice signal.
Wherein, first analysis module 12 is specifically used for：Using shot and long term memory network LSTM algorithms to the text
Information carries out text emotion analysis, obtains the probable value of each emotion information expressed by the text message.
Wherein, second analysis module 13 is specifically used for：Using convolutional neural networks CNN algorithms to the speech parameter
Speech emotional analysis is carried out, obtains the probable value of each emotion information expressed by the speech parameter.
Wherein, the acquisition module 14 includes：
Computing unit, for for each emotion information, the emotion information according to expressed by the text message
Probable value and the speech parameter information expressed by the emotion information probable value, calculate expressed by the voice signal
The emotion information combined chance value；
Determination unit, for the highest emotion information of combined chance value to be determined as to the expressed feelings of the voice signal
Feel information.
Wherein, the computing unit includes：
First computation subunit, for calculating the probable value of the emotion information expressed by the text message with presetting
The first product between text emotion coefficient；
Second computation subunit, for calculate the probable value of the emotion information expressed by the speech parameter information with
The second product between default speech emotional coefficient；
3rd computation subunit, for calculating between first product and the default matrix-vector of the emotion information
3rd product；
4th computation subunit, for calculating between second product and the default matrix-vector of the emotion information
4th product；
Subelement is obtained, expressed by obtaining the voice signal according to the 3rd product, the 4th product
The combined chance value of the emotion information.
In embodiments of the present invention, in the emotion information expressed by the voice signal that analysis user sends, sound is extracted
Text message and speech parameter information in signal；Text emotion is carried out to text information to analyze to obtain text information institute table
The emotion information reached, and the emotion that speech emotional is analyzed to obtain expressed by the speech parameter is carried out to the speech parameter information and is believed
Breath；The emotion information expressed by emotion information and the speech parameter information according to expressed by text information obtains sound letter
Number expressed emotion information.
When determining the expressed emotion information of the voice signal, the prior art is only according to big in the voice signal
Small, intonation and word speed determine the expressed emotion information of the voice signal, and the embodiment of the present invention is believed according to the sound
Text message and speech parameter information in number determine the expressed emotion information of the voice signal.
Compared with the prior art, the embodiment of the present invention is in addition to according to speech parameter information, in combination with text
Information, more fully hereinafter analyzes the emotion information expressed by voice signal, therefore can avoid the occurrence of in the prior art
Erroneous judgement situation, therefore, the embodiment of the present invention can improve the accuracy of the emotion information expressed by definite voice signal.
For device embodiment, since it is substantially similar to embodiment of the method, so description is fairly simple, it is related
Part illustrates referring to the part of embodiment of the method.
Each embodiment in this specification is described by the way of progressive, what each embodiment stressed be with
The difference of other embodiment, between each embodiment identical similar part mutually referring to.
It should be understood by those skilled in the art that, the embodiment of the embodiment of the present invention can be provided as method, apparatus or calculate
Machine program product.Therefore, the embodiment of the present invention can use complete hardware embodiment, complete software embodiment or combine software and
The form of the embodiment of hardware aspect.Moreover, the embodiment of the present invention can use one or more wherein include computer can
With in the computer-usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) of program code
The form of the computer program product of implementation.
The embodiment of the present invention be with reference to according to the method for the embodiment of the present invention, terminal device (system) and computer program
The flowchart and/or the block diagram of product describes.It should be understood that it can realize flowchart and/or the block diagram by computer program instructions
In each flow and/or block and flowchart and/or the block diagram in flow and/or square frame combination.These can be provided
Computer program instructions are set to all-purpose computer, special purpose computer, Embedded Processor or other programmable data processing terminals
Standby processor is to produce a machine so that is held by the processor of computer or other programmable data processing terminal equipments
Capable instruction is produced and is used for realization in one flow of flow chart or multiple flows and/or one square frame of block diagram or multiple square frames
The device for the function of specifying.
These computer program instructions, which may also be stored in, can guide computer or other programmable data processing terminal equipments
In the computer-readable memory to work in a specific way so that the instruction being stored in the computer-readable memory produces bag
The manufacture of command device is included, which realizes in one flow of flow chart or multiple flows and/or one side of block diagram
The function of being specified in frame or multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing terminal equipments so that
Series of operation steps is performed on computer or other programmable terminal equipments to produce computer implemented processing, so that
The instruction performed on computer or other programmable terminal equipments is provided and is used for realization in one flow of flow chart or multiple flows
And/or specified in one square frame of block diagram or multiple square frames function the step of.
Although having been described for the preferred embodiment of the embodiment of the present invention, those skilled in the art once know base
This creative concept, then can make these embodiments other change and modification.So appended claims are intended to be construed to
Including preferred embodiment and fall into all change and modification of range of embodiment of the invention.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or order.Moreover, term " comprising ", "comprising" or its any other variant meaning
Covering non-exclusive inclusion, so that process, method, article or terminal device including a series of elements are not only wrapped
Those key elements are included, but also including other elements that are not explicitly listed, or further include as this process, method, article
Or the key element that terminal device is intrinsic.In the absence of more restrictions, wanted by what sentence "including a ..." limited
Element, it is not excluded that also there are other identical element in the process including the key element, method, article or terminal device.
Above to the emotion information analysis method and device of a kind of voice signal provided by the present invention, detailed Jie has been carried out
Continue, specific case used herein is set forth the principle of the present invention and embodiment, and the explanation of above example is only
It is the method and its core concept for being used to help understand the present invention；Meanwhile for those of ordinary skill in the art, according to this hair
Bright thought, there will be changes in specific embodiments and applications, in conclusion this specification content should not manage
Solve as limitation of the present invention.
Claims (10)
- A kind of 1. emotion information analysis method of voice signal, it is characterised in that the described method includes：Extract the text message and speech parameter information in voice signal；Text emotion is carried out to the text message to analyze to obtain the emotion information expressed by the text message；Speech emotional is carried out to the speech parameter information to analyze to obtain the emotion information expressed by the speech parameter information；Emotion information expressed by emotion information and the speech parameter information according to expressed by the text message obtains institute State the emotion information expressed by voice signal.
- 2. according to the method described in claim 1, it is characterized in that, described analyze text message progress text emotion To the emotion information expressed by the text message, including：Text emotion analysis is carried out to the text message using shot and long term memory network LSTM algorithms, obtains the text message The probable value of expressed each emotion information.
- 3. according to the method described in claim 2, it is characterized in that, described carry out speech emotional point to the speech parameter information Analysis obtains the emotion information expressed by the speech parameter information, including：Speech emotional analysis is carried out to the speech parameter using convolutional neural networks CNN algorithms, obtains the speech parameter institute The probable value of each emotion information of expression.
- 4. the according to the method described in claim 3, it is characterized in that, emotion information according to expressed by the text message The emotion information expressed by the voice signal is obtained with the emotion information expressed by the speech parameter information, including：For each emotion information, probable value and the voice of the emotion information according to expressed by the text message The probable value of the emotion information expressed by parameter information, calculates the comprehensive of the emotion information expressed by the voice signal Close probable value；The highest emotion information of combined chance value is determined as to the expressed emotion information of the voice signal.
- 5. the according to the method described in claim 4, it is characterized in that, emotion according to expressed by the text message The probable value of information and the probable value of the emotion information expressed by the speech parameter information, including：Calculate first between the probable value of the emotion information expressed by the text message and pre-set text emotion coefficient Product；Calculate between the probable value of the emotion information expressed by the speech parameter information and default speech emotional coefficient Second product；Calculate the 3rd product between first product and the default matrix-vector of the emotion information；Calculate the 4th product between second product and the default matrix-vector of the emotion information；The synthesis of the emotion information according to expressed by the 3rd product, the 4th product obtain the voice signal is general Rate value.
- 6. the emotion information analytical equipment of a kind of voice signal, it is characterised in that described device includes：Extraction module, for extracting text message and speech parameter information in voice signal；First analysis module, analyzes to obtain the feelings expressed by the text message for carrying out text emotion to the text message Feel information；Second analysis module, analyzes to obtain the speech parameter information institute for carrying out speech emotional to the speech parameter information The emotion information of expression；Acquisition module, for the emotion information according to expressed by the text message and the feelings expressed by the speech parameter information Feel the emotion information expressed by voice signal described in acquisition of information.
- 7. device according to claim 6, it is characterised in that first analysis module is specifically used for：Utilize shot and long term Memory network LSTM algorithms carry out text emotion analysis to the text message, obtain each feelings expressed by the text message Feel the probable value of information.
- 8. device according to claim 7, it is characterised in that second analysis module is specifically used for：Utilize convolution god Speech emotional analysis is carried out to the speech parameter through network C NN algorithms, obtains each emotion expressed by the speech parameter The probable value of information.
- 9. device according to claim 8, it is characterised in that the acquisition module includes：Computing unit, for for each emotion information, the emotion information according to expressed by the text message it is general Rate value and the probable value of the emotion information expressed by the speech parameter information, calculate the institute expressed by the voice signal State the combined chance value of emotion information；Determination unit, the expressed emotion for the highest emotion information of combined chance value to be determined as to the voice signal are believed Breath.
- 10. device according to claim 9, it is characterised in that the computing unit includes：First computation subunit, for calculating the probable value and pre-set text of the emotion information expressed by the text message The first product between emotion coefficient；Second computation subunit, for calculating the probable value of the emotion information expressed by the speech parameter information with presetting The second product between speech emotional coefficient；3rd computation subunit, for calculating the 3rd between first product and the default matrix-vector of the emotion information Product；4th computation subunit, for calculating the 4th between second product and the default matrix-vector of the emotion information Product；Subelement is obtained, for according to expressed by the 3rd product, the 4th product acquisition voice signal The combined chance value of emotion information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711065483.3A CN108039181B (en) | 2017-11-02 | 2017-11-02 | Method and device for analyzing emotion information of sound signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711065483.3A CN108039181B (en) | 2017-11-02 | 2017-11-02 | Method and device for analyzing emotion information of sound signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108039181A true CN108039181A (en) | 2018-05-15 |
| CN108039181B CN108039181B (en) | 2021-02-12 |
Family
ID=62092727
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711065483.3A Active CN108039181B (en) | 2017-11-02 | 2017-11-02 | Method and device for analyzing emotion information of sound signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108039181B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109192225A (en) * | 2018-09-28 | 2019-01-11 | 清华大学 | The method and device of speech emotion recognition and mark |
| CN109243492A (en) * | 2018-10-28 | 2019-01-18 | 国家计算机网络与信息安全管理中心 | A kind of speech emotion recognition system and recognition methods |
| CN110390956A (en) * | 2019-08-15 | 2019-10-29 | 龙马智芯（珠海横琴）科技有限公司 | Emotion recognition network model, method and electronic equipment |
| CN110570879A (en) * | 2019-09-11 | 2019-12-13 | 深圳壹账通智能科技有限公司 | Intelligent conversation method and device based on emotion recognition and computer equipment |
| CN110675859A (en) * | 2019-09-05 | 2020-01-10 | 华南理工大学 | Multi-emotion recognition method, system, medium, and apparatus combining speech and text |
| WO2021139108A1 (en) * | 2020-01-10 | 2021-07-15 | 平安科技（深圳）有限公司 | Intelligent emotion recognition method and apparatus, electronic device, and storage medium |
Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100145695A1 (en) * | 2008-12-08 | 2010-06-10 | Electronics And Telecommunications Research Institute | Apparatus for context awareness and method using the same |
| CN102623009A (en) * | 2012-03-02 | 2012-08-01 | 安徽科大讯飞信息技术股份有限公司 | Abnormal emotion automatic detection and extraction method and system on basis of short-time analysis |
| CN102819744A (en) * | 2012-06-29 | 2012-12-12 | 北京理工大学 | Emotion recognition method with information of two channels fused |
| US20130268273A1 (en) * | 2012-04-10 | 2013-10-10 | Oscal Tzyh-Chiang Chen | Method of recognizing gender or age of a speaker according to speech emotion or arousal |
| CN103456314A (en) * | 2013-09-03 | 2013-12-18 | 广州创维平面显示科技有限公司 | Emotion recognition method and device |
| CN103810994A (en) * | 2013-09-05 | 2014-05-21 | 江苏大学 | Method and system for voice emotion inference on basis of emotion context |
| CN104200804A (en) * | 2014-09-19 | 2014-12-10 | 合肥工业大学 | Various-information coupling emotion recognition method for human-computer interaction |
| CN105334743A (en) * | 2015-11-18 | 2016-02-17 | 深圳创维－Rgb电子有限公司 | Intelligent home control method and system based on emotion recognition |
| CN105427869A (en) * | 2015-11-02 | 2016-03-23 | 北京大学 | Session emotion autoanalysis method based on depth learning |
| CN105976809A (en) * | 2016-05-25 | 2016-09-28 | 中国地质大学（武汉） | Voice-and-facial-expression-based identification method and system for dual-modal emotion fusion |
| CN106128479A (en) * | 2016-06-30 | 2016-11-16 | 福建星网视易信息系统有限公司 | A kind of performance emotion identification method and device |
| CN106297826A (en) * | 2016-08-18 | 2017-01-04 | 竹间智能科技（上海）有限公司 | Speech emotional identification system and method |
| CN106297783A (en) * | 2016-08-05 | 2017-01-04 | 易晓阳 | A kind of interactive voice identification intelligent terminal |
| CN106598948A (en) * | 2016-12-19 | 2017-04-26 | 杭州语忆科技有限公司 | Emotion recognition method based on long-term and short-term memory neural network and by combination with autocoder |
| CN106782615A (en) * | 2016-12-20 | 2017-05-31 | 科大讯飞股份有限公司 | Speech data emotion detection method and apparatus and system |
| CN106782602A (en) * | 2016-12-01 | 2017-05-31 | 南京邮电大学 | Speech-emotion recognition method based on length time memory network and convolutional neural networks |
| CN106847309A (en) * | 2017-01-09 | 2017-06-13 | 华南理工大学 | A kind of speech-emotion recognition method |
| CN107038154A (en) * | 2016-11-25 | 2017-08-11 | 阿里巴巴集团控股有限公司 | A kind of text emotion recognition methods and device |
| US20170278067A1 (en) * | 2016-03-25 | 2017-09-28 | International Business Machines Corporation | Monitoring activity to detect potential user actions |
| CN107247702A (en) * | 2017-05-05 | 2017-10-13 | 桂林电子科技大学 | A kind of text emotion analysis and processing method and system |
| CN107272607A (en) * | 2017-05-11 | 2017-10-20 | 上海斐讯数据通信技术有限公司 | A kind of intelligent home control system and method |
| CN107291696A (en) * | 2017-06-28 | 2017-10-24 | 达而观信息科技（上海）有限公司 | A kind of comment word sentiment analysis method and system based on deep learning |
-
2017
- 2017-11-02 CN CN201711065483.3A patent/CN108039181B/en active Active
Patent Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100145695A1 (en) * | 2008-12-08 | 2010-06-10 | Electronics And Telecommunications Research Institute | Apparatus for context awareness and method using the same |
| CN102623009A (en) * | 2012-03-02 | 2012-08-01 | 安徽科大讯飞信息技术股份有限公司 | Abnormal emotion automatic detection and extraction method and system on basis of short-time analysis |
| US20130268273A1 (en) * | 2012-04-10 | 2013-10-10 | Oscal Tzyh-Chiang Chen | Method of recognizing gender or age of a speaker according to speech emotion or arousal |
| CN102819744A (en) * | 2012-06-29 | 2012-12-12 | 北京理工大学 | Emotion recognition method with information of two channels fused |
| CN103456314A (en) * | 2013-09-03 | 2013-12-18 | 广州创维平面显示科技有限公司 | Emotion recognition method and device |
| CN103810994A (en) * | 2013-09-05 | 2014-05-21 | 江苏大学 | Method and system for voice emotion inference on basis of emotion context |
| CN104200804A (en) * | 2014-09-19 | 2014-12-10 | 合肥工业大学 | Various-information coupling emotion recognition method for human-computer interaction |
| CN105427869A (en) * | 2015-11-02 | 2016-03-23 | 北京大学 | Session emotion autoanalysis method based on depth learning |
| CN105334743A (en) * | 2015-11-18 | 2016-02-17 | 深圳创维－Rgb电子有限公司 | Intelligent home control method and system based on emotion recognition |
| US20170278067A1 (en) * | 2016-03-25 | 2017-09-28 | International Business Machines Corporation | Monitoring activity to detect potential user actions |
| CN105976809A (en) * | 2016-05-25 | 2016-09-28 | 中国地质大学（武汉） | Voice-and-facial-expression-based identification method and system for dual-modal emotion fusion |
| CN106128479A (en) * | 2016-06-30 | 2016-11-16 | 福建星网视易信息系统有限公司 | A kind of performance emotion identification method and device |
| CN106297783A (en) * | 2016-08-05 | 2017-01-04 | 易晓阳 | A kind of interactive voice identification intelligent terminal |
| CN106297826A (en) * | 2016-08-18 | 2017-01-04 | 竹间智能科技（上海）有限公司 | Speech emotional identification system and method |
| CN107038154A (en) * | 2016-11-25 | 2017-08-11 | 阿里巴巴集团控股有限公司 | A kind of text emotion recognition methods and device |
| CN106782602A (en) * | 2016-12-01 | 2017-05-31 | 南京邮电大学 | Speech-emotion recognition method based on length time memory network and convolutional neural networks |
| CN106598948A (en) * | 2016-12-19 | 2017-04-26 | 杭州语忆科技有限公司 | Emotion recognition method based on long-term and short-term memory neural network and by combination with autocoder |
| CN106782615A (en) * | 2016-12-20 | 2017-05-31 | 科大讯飞股份有限公司 | Speech data emotion detection method and apparatus and system |
| CN106847309A (en) * | 2017-01-09 | 2017-06-13 | 华南理工大学 | A kind of speech-emotion recognition method |
| CN107247702A (en) * | 2017-05-05 | 2017-10-13 | 桂林电子科技大学 | A kind of text emotion analysis and processing method and system |
| CN107272607A (en) * | 2017-05-11 | 2017-10-20 | 上海斐讯数据通信技术有限公司 | A kind of intelligent home control system and method |
| CN107291696A (en) * | 2017-06-28 | 2017-10-24 | 达而观信息科技（上海）有限公司 | A kind of comment word sentiment analysis method and system based on deep learning |
Non-Patent Citations (5)
| Title |
|---|
| 徐莹莹: "基于深度神经网络模型的句子级文本情感分类研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 曹宇慧: "基于深度学习的文本情感分析研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 蔡慧苹: "基于word embedding和CNN的情感分析类型", 《计算机应用研究》 * |
| 谢坷珍: "融合人脸表情和语音的双模态情感识别研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 陈雷: "面向股票评论的情感分析系统研究与实现", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109192225A (en) * | 2018-09-28 | 2019-01-11 | 清华大学 | The method and device of speech emotion recognition and mark |
| CN109243492A (en) * | 2018-10-28 | 2019-01-18 | 国家计算机网络与信息安全管理中心 | A kind of speech emotion recognition system and recognition methods |
| CN110390956A (en) * | 2019-08-15 | 2019-10-29 | 龙马智芯（珠海横琴）科技有限公司 | Emotion recognition network model, method and electronic equipment |
| CN110675859A (en) * | 2019-09-05 | 2020-01-10 | 华南理工大学 | Multi-emotion recognition method, system, medium, and apparatus combining speech and text |
| CN110675859B (en) * | 2019-09-05 | 2021-11-23 | 华南理工大学 | Multi-emotion recognition method, system, medium, and apparatus combining speech and text |
| CN110570879A (en) * | 2019-09-11 | 2019-12-13 | 深圳壹账通智能科技有限公司 | Intelligent conversation method and device based on emotion recognition and computer equipment |
| WO2021047180A1 (en) * | 2019-09-11 | 2021-03-18 | 深圳壹账通智能科技有限公司 | Emotion recognition-based smart chat method, device, and computer apparatus |
| WO2021139108A1 (en) * | 2020-01-10 | 2021-07-15 | 平安科技（深圳）有限公司 | Intelligent emotion recognition method and apparatus, electronic device, and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108039181B (en) | 2021-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108039181A (en) | The emotion information analysis method and device of a kind of voice signal | |
| CN108630193B (en) | Voice recognition method and device | |
| JP6755304B2 (en) | Information processing device | |
| WO2021128741A1 (en) | Voice emotion fluctuation analysis method and apparatus, and computer device and storage medium | |
| KR102413692B1 (en) | Apparatus and method for caculating acoustic score for speech recognition, speech recognition apparatus and method, and electronic device | |
| CN109545192B (en) | Method and apparatus for generating a model | |
| CN109545193B (en) | Method and apparatus for generating a model | |
| WO2020253128A1 (en) | Voice recognition-based communication service method, apparatus, computer device, and storage medium | |
| CN112581938B (en) | Speech breakpoint detection method, device and equipment based on artificial intelligence | |
| CN110570853A (en) | Intention recognition method and device based on voice data | |
| CN110992942B (en) | Voice recognition method and device for voice recognition | |
| CN107919137A (en) | The long-range measures and procedures for the examination and approval, device, equipment and readable storage medium storing program for executing | |
| US10971149B2 (en) | Voice interaction system for interaction with a user by voice, voice interaction method, and program | |
| CN112992191B (en) | Voice endpoint detection method and device, electronic equipment and readable storage medium | |
| CN114127849A (en) | Speech emotion recognition method and device | |
| CN109994126A (en) | Audio message segmentation method, device, storage medium and electronic equipment | |
| CN113096647B (en) | Voice model training method and device and electronic equipment | |
| CN109215647A (en) | Voice awakening method, electronic equipment and non-transient computer readable storage medium | |
| CN108877779B (en) | Method and device for detecting voice tail point | |
| CN111667834B (en) | Hearing-aid equipment and hearing-aid method | |
| CN112735385A (en) | Voice endpoint detection method and device, computer equipment and storage medium | |
| CN107910021A (en) | A kind of symbol insertion method and device | |
| US11551707B2 (en) | Speech processing method, information device, and computer program product | |
| CN111414748A (en) | Traffic data processing method and device | |
| CN110444194A (en) | A kind of speech detection method and device |
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 |