CN110070893A - A kind of system, method and apparatus carrying out sentiment analysis using vagitus - Google Patents
A kind of system, method and apparatus carrying out sentiment analysis using vagitus Download PDFInfo
- Publication number
- CN110070893A CN110070893A CN201910227535.5A CN201910227535A CN110070893A CN 110070893 A CN110070893 A CN 110070893A CN 201910227535 A CN201910227535 A CN 201910227535A CN 110070893 A CN110070893 A CN 110070893A
- Authority
- CN
- China
- Prior art keywords
- model
- sound
- vagitus
- training data
- detection
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
- G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks
-
- 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
The invention belongs to depth learning technology fields, and in particular to a kind of system, method and apparatus that sentiment analysis is carried out using vagitus；Crying detection module, training data sample, obtains detection model, using the sound to be detected of detection model detection input, judges whether the sound of input is vagitus；Crying analysis module, training data sample, obtains analysis model, does emotional semantic classification to the vagitus detected using the analysis model；Model modification module uploads the acoustic segment labeled as identification mistake and corresponding class label, while the disaggregated model that upload user is currently used；The voice data and original training data uploaded using user, the disaggregated model that fine tuning user uploads；Trained new model is downloaded, original model is replaced.Have the advantages that portable, recognition accuracy is high and applicability is wide.
Description
Technical field
The invention belongs to depth learning technology fields, and in particular to a kind of to be using what vagitus carried out sentiment analysis
System, method and apparatus.
Background technique
Consumption in terms of baby-monitoring now, which only resides within, employs nurse and hospital to nurse a baby.With the development of society, people
The problems such as power is expensive gradually emerges in large numbers, and undoubtedly increases the burden for the young parent that those need that nurse is engaged to nurse family.
In addition the busy time is more and more outside by present young parent, therefore lacks to oneself baby and look after.If giving old man to look after
Baby, old man are getting in years, and looking after baby may be neglected, it may appear that baby cry is not slept by timely nursing, at night
Baby kick quilt son nobody know, babies uncomfortable unmanned dawn phenomena such as.
ZL201310440063.4 discloses a kind of baby monitor that can identify vagitus and vagitus identification side
Method；Which disclose following technical solutions: a kind of baby monitor that can identify vagitus, including main control module, baby
Crying identification module and SMS transmission module, wherein main control module are receiving baby that vagitus identification module is sent
After the information cry and screamed, SMS transmission module is sent that information to；Vagitus identification module, is connected with main control module, in real time
The voice messaging in ambient enviroment is acquired, and voice messaging is handled, by other sound in the crying and environment of baby
It distinguishes, vagitus identification module includes that voice messaging acquisition module, speech signal analysis module and voice messaging determine mould
Block, wherein voice messaging acquisition module acquires the voice messaging in ambient enviroment in real time, and speech signal analysis module will collect
Voice messaging be first split into speech frame, frame length is 100 milliseconds, and it is 50 milliseconds that frame, which moves, is then made plus Hanning window processing, finally right
Every frame voice carries out Fast Fourier Transform (FFT), will treated transmission of speech information to voice messaging determination module, voice determines
Module calculates the ratio between 1k Hz to 3k Hz frequency band energy and a frame energy summation, and ratio is greater than to 0.4 speech frame
Labeled as crying frame, comprehensive continuous 20 frame, when having 10 frames in continuous 20 frame the above are when crying frame, judgement detects baby cried
Sound, and baby cry information is sent to main control module；SMS transmission module is connected with main control module, receives main control module
After the information that the baby of transmission is crying and screaming.
Bulky entity apparatus, it is not readily portable or need to connect server and analyze, depend on network.And it lacks
Few feedback to analysis result.
Summary of the invention
In view of this, the main purpose of the present invention is to provide it is a kind of using vagitus carry out sentiment analysis system,
Method and apparatus have the advantages that portable, recognition accuracy is high and applicability is wide.
In order to achieve the above objectives, the technical scheme of the present invention is realized as follows:
As shown in Figure 1, a kind of system for carrying out sentiment analysis using vagitus, the system comprises:
Crying detection module, training data sample, obtains detection model, uses the to be detected of detection model detection input
Sound judges whether the sound of input is vagitus；
Crying analysis module, training data sample, obtains analysis model, using the analysis model to the baby cried detected
Sound does emotional semantic classification；
Model modification module uploads the acoustic segment labeled as identification mistake and corresponding class label, while upload user
Currently used disaggregated model；The voice data and original training data uploaded using user, the classification that fine tuning user uploads
Model；Trained new model is downloaded, original model is replaced.
Further, the crying detection module includes: detection training module, and training data, training are collected in training part
A model that can detecte out vagitus out；Test module is detected, the sound to be detected of detection model detection input is used
Sound judges whether the sound of input is vagitus.
Further, the crying analysis module includes: analyzing and training module, and training data sample obtains analysis model,
Train the model that can analyze vagitus；Test module is analyzed, emotional semantic classification is done to the vagitus detected.
Further, the model modification module includes: data uploading module, uploads the acoustic segment labeled as identification mistake
And corresponding class label, while the disaggregated model that upload user is currently used；Model training module, the sound uploaded using user
Sound data and original training data, the disaggregated model that fine tuning user uploads；New model download module downloads trained new mould
Type replaces original model.
A method of sentiment analysis being carried out using vagitus, the method executes following steps:
Training data sample, obtains detection model, using the sound to be detected of detection model detection input, judges to input
Sound whether be vagitus；
Training data sample, obtains analysis model, does emotional semantic classification to the vagitus detected using the analysis model；
Upload the acoustic segment labeled as identification mistake and corresponding class label, while the classification that upload user is currently used
Model；The voice data and original training data uploaded using user, the disaggregated model that fine tuning user uploads；Downloading trains
New model, replace original model.
Further, the training data sample, obtains detection model, uses the to be detected of detection model detection input
Sound, judge input sound whether be vagitus method execute following steps:
Various ambient sounds are collected as training data, and manually add a tag along sort for every section of sound；
It is training set and test set by training data random division；
Every section of sound in training set is sampled, at random cutting and normalized so that the numerical value of each sampled point
In [- 1,1] range；
Training data is sent into neural network training model；
Acquire sound to be detected；
Collected sound to be detected is sampled, random cutting and normalized, so that the numerical value of each sampled point
In [- 1,1] range；
The collected multistage sound of multistage is sent into trained neural network in advance, prediction result is obtained, prediction is tied
Fruit is cooked ballot processing, and highest prediction result of winning the vote is final prediction result.
Further, the training data sample, obtains analysis model, using the analysis model to the baby cried detected
The method that sound does emotional semantic classification executes following steps:
Crying of the baby under different emotions state is collected, and puts on class label；
It is training set and test set by training data random division；
Every section of sound in training set is pre-processed, comprising: sample, have and overlappingly cut, sound is done at normalization
Reason, so that the numerical value of each sampled point is in [- 1,1] range；
Training data is sent into neural network training model；
The crying that vagitus detection part is detected is as input.
Sound is sent into trained neural network in advance, prediction result is obtained, ballot processing is done to prediction result, win the vote
Highest prediction result is final prediction result.
A kind of device carrying out sentiment analysis using vagitus, described device includes: a kind of computer of non-transitory
Readable storage medium storing program for executing, the storage medium store computations comprising: training data sample obtains detection model, uses this
Detection model detection input sound to be detected, judge input sound whether be vagitus code segment；Training data sample
This, is obtained analysis model, is done the code segment of emotional semantic classification to the vagitus detected using the analysis model；Upload is labeled as
Identify wrong acoustic segment and corresponding class label, while the disaggregated model that upload user is currently used；It is uploaded using user
Voice data and original training data, fine tuning user upload disaggregated model；Trained new model is downloaded, replacement is original
Model code segment.
A kind of system, method and apparatus being carried out sentiment analysis using vagitus of the invention are had following beneficial to effect
Fruit: can automatically detect the crying of baby, and analyze emotion reason corresponding to this kind of crying.It can be directly mounted at intelligent hand
On the intelligent terminals such as machine, tablet computer, without additional hardware；Conventional crying detection and analysis is all based on local (intelligence
Energy mobile phone, tablet computer etc. itself), do not depend on network；User can make feedback to analysis result, collect the language of analysis mistake
The regular upload server of segment, updates classifier, and the exclusive crying classifier of customization baby improves classification accuracy.
Detailed description of the invention
Fig. 1 is the method flow schematic diagram of the system of the invention that sentiment analysis is carried out using vagitus.
Fig. 2 is the structural representation of the crying detection module of the system of the invention that sentiment analysis is carried out using vagitus
Figure.
Fig. 3 is that the structure of the crying analysis detection module of the system of the invention that sentiment analysis is carried out using vagitus is shown
It is intended to.
Fig. 4 is the structural representation of the model modification module of the system of the invention that sentiment analysis is carried out using vagitus
Figure.
Fig. 5 is the method flow schematic diagram of the invention that sentiment analysis method is carried out using vagitus.
Specific embodiment
With reference to the accompanying drawing and the embodiment of the present invention is described in further detail method of the invention.
A kind of system carrying out sentiment analysis using vagitus, the system comprises:
Crying detection module, training data sample, obtains detection model, uses the to be detected of detection model detection input
Sound judges whether the sound of input is vagitus；
Crying analysis module, training data sample, obtains analysis model, using the analysis model to the baby cried detected
Sound does emotional semantic classification；
Model modification module uploads the acoustic segment labeled as identification mistake and corresponding class label, while upload user
Currently used disaggregated model；The voice data and original training data uploaded using user, the classification that fine tuning user uploads
Model；Trained new model is downloaded, original model is replaced.
Further, the crying detection module includes: detection training module, and training data, training are collected in training part
A model that can detecte out vagitus out；Test module is detected, the sound to be detected of detection model detection input is used
Sound judges whether the sound of input is vagitus.
Further, the crying analysis module includes: analyzing and training module, and training data sample obtains analysis model,
Train the model that can analyze vagitus；Test module is analyzed, emotional semantic classification is done to the vagitus detected.
Further, the model modification module includes: data uploading module, uploads the acoustic segment labeled as identification mistake
And corresponding class label, while the disaggregated model that upload user is currently used；Model training module, the sound uploaded using user
Sound data and original training data, the disaggregated model that fine tuning user uploads；New model download module downloads trained new mould
Type replaces original model.
The method of operation of crying detection module are as follows: collect various ambient sounds as training data, and be manually every section of sound
Sound add tag along sort (shared K class) (such as: vagitus, the patter of rain, sound of the wind, laugh, mew, barking, footsteps,
Enabling sound etc.).
It then is training set and test set by training data random division.
Every section of sound in training set is pre-processed, including sampling (sample frequency 16khz), random cutting (so that
The length of every section of sound is 25ms, that is, includes 16000 × 0.25=400 sampled point), normalized is done to sound, so that
The numerical value of each sampled point is in [- 1,1] range.Each section of sound can generate multiple isometric acoustic segments in this way.Finally
N number of training sample is generated, { xi, yi }, wherein xi is acoustic vector (length 400, value range are [- 1,1]), and yi is sound
Label (value range be [0, K])
Training data is sent into neural network training model.Network structure is two one-dimensional convolution module (each module packets
Containing an one-dimensional convolutional layer, one Relu layers, one pool layers), (each module includes a two dimension volume to three two-dimensional convolution layers
Lamination, one Relu layers, one pool layers), three full articulamentums.Loss function is cross entropy.Calculation formula is as follows:
The output of neural network is K dimensional vector [a0, a1 ... ak-1], brings softmax formula into and K dimensional vector is calculated
S=[s0, s1 ..., sk-1]；
By the label vector y [y0, y1 ..., yk-1] of the S being calculated and the sample, (wherein yi=1, i are the sample
Corresponding classification) bring cross entropy formula into, obtain loss L.
Wherein, cross entropy formula is as follows:
Sound is acquired in real time by intelligent terminal.
Collected sound is pre-processed, is sampled (sample frequency 16khz), random cutting (so that every section of sound
Length is 25ms, that is, includes 16000 × 0.25=400 sampled point), normalized (Data=Data*1.0/ is done to sound
Max (abs (Data))) so that the numerical value of each sampled point obtains M sections of sound in [- 1,1] range.
The collected multistage sound of multistage is sent into trained neural network in advance, M prediction result is obtained, to M
Prediction result does ballot processing, and highest prediction result of winning the vote is final prediction result.
The method of operation of crying analysis module is as follows: collect baby under different emotions state crying (it is hungry, it is sleepy,
Want to have the hiccups, pain, uncomfortable (diaper is wet, heat etc.)) and put on class label.
It is training set and test set by training data random division.
Every section of sound in training set is pre-processed, including sampling (sample frequency 16khz), has and overlappingly cuts
(it is cut once at interval of 10ms, so that the length of every section of sound is 25ms, that is, include 16000 × 0.25=400 sampled point),
Normalized is done to sound, so that the numerical value of each sampled point is in [- 1,1] range.Each section of sound in this way
Generate multiple isometric acoustic segments.N number of training sample is finally generated, { xi, yi }, wherein xi is that acoustic vector is (long
Degree is 400, and value range is [- 1,1]), yi is the label of sound (value range is [0,5])
Training data is sent into neural network training model.Network structure is two one-dimensional convolution module (each module packets
Containing an one-dimensional convolutional layer, one Relu layers, one pool layers), (each module includes two two dimension volumes to three two-dimensional convolution layers
Lamination, one Relu layers, one pool layers), three full articulamentums.Loss function is cross entropy.Calculation formula is as follows:
The output of neural network is K dimensional vector [a0, a1 ... ak-1], brings softmax formula into and K dimensional vector is calculated
S=[s0, s1 ..., sk-1]；
By the label vector y [y0, y1 ..., yk-1] of the S being calculated and the sample, (wherein yi=1, i are the sample
Corresponding classification) bring cross entropy formula into, obtain loss L.
It regard the crying (preprocessed obtained M section sound) that vagitus detection part detects as input.
M sound is sent into trained neural network in advance, M prediction result is obtained, voting booth is done to M prediction result
Reason, highest prediction result of winning the vote is final prediction result.
User feedback, user can make feedback to analysis result, click Yes or No button, and if it is No, user can be with
Select the classification that she praises.
A method of sentiment analysis being carried out using vagitus, the method executes following steps:
Training data sample, obtains detection model, using the sound to be detected of detection model detection input, judges to input
Sound whether be vagitus；
Training data sample, obtains analysis model, does emotional semantic classification to the vagitus detected using the analysis model；
Upload the acoustic segment labeled as identification mistake and corresponding class label, while the classification that upload user is currently used
Model；The voice data and original training data uploaded using user, the disaggregated model that fine tuning user uploads；Downloading trains
New model, replace original model.
Further, the training data sample, obtains detection model, uses the to be detected of detection model detection input
Sound, judge input sound whether be vagitus method execute following steps:
Various ambient sounds are collected as training data, and manually add a tag along sort for every section of sound；
It is training set and test set by training data random division；
Every section of sound in training set is sampled, at random cutting and normalized so that the numerical value of each sampled point
In [- 1,1] range；
Training data is sent into neural network training model；
Acquire sound to be detected；
Collected sound to be detected is sampled, random cutting and normalized, so that the numerical value of each sampled point
In [- 1,1] range；
The collected multistage sound of multistage is sent into trained neural network in advance, prediction result is obtained, prediction is tied
Fruit is cooked ballot processing, and highest prediction result of winning the vote is final prediction result.
Further, the training data sample, obtains analysis model, using the analysis model to the baby cried detected
The method that sound does emotional semantic classification executes following steps:
Crying of the baby under different emotions state is collected, and puts on class label；
It is training set and test set by training data random division；
Every section of sound in training set is pre-processed, comprising: sample, have and overlappingly cut, sound is done at normalization
Reason, so that the numerical value of each sampled point is in [- 1,1] range；
Training data is sent into neural network training model；
The crying that vagitus detection part is detected is as input.
Sound is sent into trained neural network in advance, prediction result is obtained, ballot processing is done to prediction result, win the vote
Highest prediction result is final prediction result.
A kind of device carrying out sentiment analysis using vagitus, described device includes: a kind of computer of non-transitory
Readable storage medium storing program for executing, the storage medium store computations comprising: training data sample obtains detection model, uses this
Detection model detection input sound to be detected, judge input sound whether be vagitus code segment；Training data sample
This, is obtained analysis model, is done the code segment of emotional semantic classification to the vagitus detected using the analysis model；Upload is labeled as
Identify wrong acoustic segment and corresponding class label, while the disaggregated model that upload user is currently used；It is uploaded using user
Voice data and original training data, fine tuning user upload disaggregated model；Trained new model is downloaded, replacement is original
Model code segment.
Person of ordinary skill in the field can be understood that, for convenience and simplicity of description, foregoing description
The specific work process of system and related explanation, can refer to corresponding processes in the foregoing method embodiment, details are not described herein.
It should be noted that system provided by the above embodiment, only illustrate with the division of above-mentioned each functional module
It is bright, in practical applications, it can according to need and complete above-mentioned function distribution by different functional modules, i.e., it will be of the invention
Module or step in embodiment are decomposed or are combined again, for example, the module of above-described embodiment can be merged into a module,
It can also be further split into multiple submodule, to complete all or part of the functions described above.The present invention is implemented
Module, the title of step involved in example, it is only for distinguish modules or step, be not intended as to of the invention improper
It limits.
Person of ordinary skill in the field can be understood that, for convenience and simplicity of description, foregoing description
The specific work process and related explanation of storage device, processing unit, can refer to corresponding processes in the foregoing method embodiment,
Details are not described herein.
Those skilled in the art should be able to recognize that, mould described in conjunction with the examples disclosed in the embodiments of the present disclosure
Block, method and step, can be realized with electronic hardware, computer software, or a combination of the two, software module, method and step pair
The program answered can be placed in random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electric erasable and can compile
Any other form of storage well known in journey ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field is situated between
In matter.In order to clearly demonstrate the interchangeability of electronic hardware and software, in the above description according to function generally
Describe each exemplary composition and step.These functions are executed actually with electronic hardware or software mode, depend on technology
The specific application and design constraint of scheme.Those skilled in the art can carry out using distinct methods each specific application
Realize described function, but such implementation should not be considered as beyond the scope of the present invention.
Term " first ", " second " etc. are to be used to distinguish similar objects, rather than be used to describe or indicate specific suitable
Sequence or precedence.
Term " includes " or any other like term are intended to cover non-exclusive inclusion, so that including a system
Process, method, article or equipment/device of column element not only includes those elements, but also including being not explicitly listed
Other elements, or further include the intrinsic element of these process, method, article or equipment/devices.
So far, it has been combined preferred embodiment shown in the drawings and describes technical solution of the present invention, still, this field
Technical staff is it is easily understood that protection scope of the present invention is expressly not limited to these specific embodiments.Without departing from this
Under the premise of the principle of invention, those skilled in the art can make equivalent change or replacement to the relevant technologies feature, these
Technical solution after change or replacement will fall within the scope of protection of the present invention.
The foregoing is only a preferred embodiment of the present invention, is not intended to limit the scope of the present invention.
Claims (8)
1. a kind of system for carrying out sentiment analysis using vagitus, which is characterized in that the system comprises:
Crying detection module, training data sample, obtains detection model, uses the sound to be detected of detection model detection input
Sound judges whether the sound of input is vagitus；
Crying analysis module, training data sample, obtains analysis model, is done using the analysis model to the vagitus detected
Emotional semantic classification；
Model modification module uploads the acoustic segment labeled as identification mistake and corresponding class label, while upload user is current
The disaggregated model used；The voice data and original training data uploaded using user, the disaggregated model that fine tuning user uploads；
Trained new model is downloaded, original model is replaced.
2. the system as described in claim 1 for carrying out sentiment analysis using vagitus, which is characterized in that the crying detection
Module includes: detection training module, and training data is collected in training part, trains the mould that can detecte out vagitus
Type；Test module is detected, using the sound to be detected of detection model detection input, judges whether the sound of input is baby cried
Sound.
3. the system as claimed in claim 2 for carrying out sentiment analysis using vagitus, which is characterized in that the crying analysis
Module includes: analyzing and training module, and training data sample obtains analysis model, and training one can analyze vagitus
Model；Test module is analyzed, emotional semantic classification is done to the vagitus detected.
4. the system as claimed in claim 3 for carrying out sentiment analysis using vagitus, which is characterized in that the model modification
Module includes: data uploading module, uploads the acoustic segment labeled as identification mistake and corresponding class label, while upload user
Currently used disaggregated model；Model training module, the voice data and original training data uploaded using user, microcall
The disaggregated model that family uploads；New model download module downloads trained new model, replaces original model.
5. a kind of method for carrying out sentiment analysis using vagitus, which is characterized in that the method executes following steps:
Training data sample, obtains detection model, using the sound to be detected of detection model detection input, judges the sound of input
Whether sound is vagitus；
Training data sample, obtains analysis model, does emotional semantic classification to the vagitus detected using the analysis model；
Upload the acoustic segment labeled as identification mistake and corresponding class label, while the classification mould that upload user is currently used
Type；The voice data and original training data uploaded using user, the disaggregated model that fine tuning user uploads；It downloads trained
New model replaces original model.
6. the method as claimed in claim 5 for carrying out sentiment analysis using vagitus, which is characterized in that the training data
Sample obtains detection model, using the sound to be detected of detection model detection input, judges whether the sound of input is baby
The method of crying executes following steps:
Various ambient sounds are collected as training data, and manually add a tag along sort for every section of sound；
It is training set and test set by training data random division；
Every section of sound in training set is sampled, at random cutting and normalized so that the numerical value of each sampled point [-
1,1] in range；
Training data is sent into neural network training model；
Acquire sound to be detected；
Collected sound to be detected is sampled, random cutting and normalized so that the numerical value of each sampled point [-
1,1] in range；
The collected multistage sound of multistage is sent into trained neural network in advance, prediction result is obtained, prediction result is done
Ballot processing, highest prediction result of winning the vote is final prediction result.
7. the method as claimed in claim 6 for carrying out sentiment analysis using vagitus, which is characterized in that the training data
Sample obtains analysis model, executes following step using the method that the analysis model does emotional semantic classification to the vagitus detected
It is rapid:
Crying of the baby under different emotions state is collected, and puts on class label；
It is training set and test set by training data random division；
Every section of sound in training set is pre-processed, comprising: sample, have and overlappingly cut, normalized is done to sound, is made
The numerical value of each sampled point is obtained in [- 1,1] range；
Training data is sent into neural network training model；
The crying that vagitus detection part is detected is as input；
Sound is sent into trained neural network in advance, obtains prediction result, ballot processing is done to prediction result, highest of winning the vote
Prediction result be final prediction result.
8. a kind of device for carrying out sentiment analysis using vagitus, which is characterized in that described device includes: a kind of non-transitory
Computer readable storage medium, which stores computations comprising: training data sample obtains detection mould
Type, using the detection model detection input sound to be detected, judge input sound whether be vagitus code segment；Instruction
Practice data sample, obtains analysis model, do the code segment of emotional semantic classification to the vagitus detected using the analysis model；On
Pass the acoustic segment labeled as identification mistake and corresponding class label, while the disaggregated model that upload user is currently used；It utilizes
The voice data and original training data that user uploads, the disaggregated model that fine tuning user uploads；Trained new model is downloaded,
Replace the code segment of original model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910227535.5A CN110070893A (en) | 2019-03-25 | 2019-03-25 | A kind of system, method and apparatus carrying out sentiment analysis using vagitus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910227535.5A CN110070893A (en) | 2019-03-25 | 2019-03-25 | A kind of system, method and apparatus carrying out sentiment analysis using vagitus |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110070893A true CN110070893A (en) | 2019-07-30 |
Family
ID=67366561
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910227535.5A Pending CN110070893A (en) | 2019-03-25 | 2019-03-25 | A kind of system, method and apparatus carrying out sentiment analysis using vagitus |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110070893A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111354375A (en) * | 2020-02-25 | 2020-06-30 | 咪咕文化科技有限公司 | Cry classification method, device, server and readable storage medium |
| CN111785300A (en) * | 2020-06-12 | 2020-10-16 | 北京快鱼电子股份公司 | Crying detection method and system based on deep neural network |
| CN113270115A (en) * | 2020-02-17 | 2021-08-17 | 广东美的制冷设备有限公司 | Infant monitoring device, infant monitoring method thereof, control device and storage medium |
| CN114463937A (en) * | 2022-03-07 | 2022-05-10 | 云知声智能科技股份有限公司 | Infant monitoring method and device, electronic equipment and storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103426438A (en) * | 2012-05-25 | 2013-12-04 | 洪荣昭 | Method and system for analyzing baby crying |
| CN104347066A (en) * | 2013-08-09 | 2015-02-11 | 盛乐信息技术（上海）有限公司 | Deep neural network-based baby cry identification method and system |
| CN106653001A (en) * | 2016-11-17 | 2017-05-10 | 沈晓明 | Baby crying identifying method and system |
| CN107122807A (en) * | 2017-05-24 | 2017-09-01 | 努比亚技术有限公司 | A kind of family's monitoring method, service end and computer-readable recording medium |
| CN107657963A (en) * | 2016-07-25 | 2018-02-02 | 韦创科技有限公司 | Sob identification system and sob discrimination method |
| CN107818779A (en) * | 2017-09-15 | 2018-03-20 | 北京理工大学 | A kind of infant's crying sound detection method, apparatus, equipment and medium |
| CN108305642A (en) * | 2017-06-30 | 2018-07-20 | 腾讯科技（深圳）有限公司 | The determination method and apparatus of emotion information |
| CN109065034A (en) * | 2018-09-25 | 2018-12-21 | 河南理工大学 | A kind of vagitus interpretation method based on sound characteristic identification |
| CN109243493A (en) * | 2018-10-30 | 2019-01-18 | 南京工程学院 | Based on the vagitus emotion identification method for improving long memory network in short-term |
| CN109509484A (en) * | 2018-12-25 | 2019-03-22 | 科大讯飞股份有限公司 | A kind of prediction technique and device of baby crying reason |
-
2019
- 2019-03-25 CN CN201910227535.5A patent/CN110070893A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103426438A (en) * | 2012-05-25 | 2013-12-04 | 洪荣昭 | Method and system for analyzing baby crying |
| CN104347066A (en) * | 2013-08-09 | 2015-02-11 | 盛乐信息技术（上海）有限公司 | Deep neural network-based baby cry identification method and system |
| CN107657963A (en) * | 2016-07-25 | 2018-02-02 | 韦创科技有限公司 | Sob identification system and sob discrimination method |
| CN106653001A (en) * | 2016-11-17 | 2017-05-10 | 沈晓明 | Baby crying identifying method and system |
| CN107122807A (en) * | 2017-05-24 | 2017-09-01 | 努比亚技术有限公司 | A kind of family's monitoring method, service end and computer-readable recording medium |
| CN108305642A (en) * | 2017-06-30 | 2018-07-20 | 腾讯科技（深圳）有限公司 | The determination method and apparatus of emotion information |
| CN107818779A (en) * | 2017-09-15 | 2018-03-20 | 北京理工大学 | A kind of infant's crying sound detection method, apparatus, equipment and medium |
| CN109065034A (en) * | 2018-09-25 | 2018-12-21 | 河南理工大学 | A kind of vagitus interpretation method based on sound characteristic identification |
| CN109243493A (en) * | 2018-10-30 | 2019-01-18 | 南京工程学院 | Based on the vagitus emotion identification method for improving long memory network in short-term |
| CN109509484A (en) * | 2018-12-25 | 2019-03-22 | 科大讯飞股份有限公司 | A kind of prediction technique and device of baby crying reason |
Non-Patent Citations (1)
| Title |
|---|
| 杨汝清等: "《智能控制工程》", 31 January 2001 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113270115A (en) * | 2020-02-17 | 2021-08-17 | 广东美的制冷设备有限公司 | Infant monitoring device, infant monitoring method thereof, control device and storage medium |
| CN113270115B (en) * | 2020-02-17 | 2023-04-11 | 广东美的制冷设备有限公司 | Infant monitoring device, infant monitoring method thereof, control device and storage medium |
| CN111354375A (en) * | 2020-02-25 | 2020-06-30 | 咪咕文化科技有限公司 | Cry classification method, device, server and readable storage medium |
| CN111785300A (en) * | 2020-06-12 | 2020-10-16 | 北京快鱼电子股份公司 | Crying detection method and system based on deep neural network |
| CN111785300B (en) * | 2020-06-12 | 2021-05-25 | 北京快鱼电子股份公司 | Crying detection method and system based on deep neural network |
| CN114463937A (en) * | 2022-03-07 | 2022-05-10 | 云知声智能科技股份有限公司 | Infant monitoring method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110070893A (en) | A kind of system, method and apparatus carrying out sentiment analysis using vagitus | |
| CN108053838B (en) | In conjunction with fraud recognition methods, device and the storage medium of audio analysis and video analysis | |
| CN108009521B (en) | Face image matching method, device, terminal and storage medium | |
| San-Segundo et al. | Feature extraction from smartphone inertial signals for human activity segmentation | |
| US11355138B2 (en) | Audio scene recognition using time series analysis | |
| CN110853617B (en) | Model training method, language identification method, device and equipment | |
| CN103294199B (en) | A kind of unvoiced information identifying system based on face's muscle signals | |
| CN110570873A (en) | voiceprint wake-up method and device, computer equipment and storage medium | |
| CN109783798A (en) | Method, apparatus, terminal and the storage medium of text information addition picture | |
| CN104036776A (en) | Speech emotion identification method applied to mobile terminal | |
| CN108717852A (en) | A kind of intelligent robot Semantic interaction system and method based on white light communication and the cognition of class brain | |
| CN110972112B (en) | Subway running direction determining method, device, terminal and storage medium | |
| CN109394258A (en) | A kind of classification method, device and the terminal device of lung's breath sound | |
| CN108109331A (en) | Monitoring method and monitoring system | |
| CN111667818A (en) | Method and device for training awakening model | |
| CN107085717A (en) | A kind of family's monitoring method, service end and computer-readable recording medium | |
| CN109658921A (en) | A kind of audio signal processing method, equipment and computer readable storage medium | |
| Turan et al. | Monitoring Infant's Emotional Cry in Domestic Environments Using the Capsule Network Architecture. | |
| CN104123930A (en) | Guttural identification method and device | |
| CN110136726A (en) | A kind of estimation method, device, system and the storage medium of voice gender | |
| CN109602421A (en) | Health monitor method, device and computer readable storage medium | |
| CN114078472A (en) | Training method and device for keyword calculation model with low false awakening rate | |
| Mahmoud et al. | Smart nursery for smart cities: Infant sound classification based on novel features and support vector classifier | |
| Nguyen et al. | A potential approach for emotion prediction using heart rate signals | |
| Eyobu et al. | A real-time sleeping position recognition system using IMU sensor motion data |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190730 |
|
| RJ01 | Rejection of invention patent application after publication |