CN105957531A - Speech content extracting method and speech content extracting device based on cloud platform - Google Patents
Speech content extracting method and speech content extracting device based on cloud platform Download PDFInfo
- Publication number
- CN105957531A CN105957531A CN201610260647.7A CN201610260647A CN105957531A CN 105957531 A CN105957531 A CN 105957531A CN 201610260647 A CN201610260647 A CN 201610260647A CN 105957531 A CN105957531 A CN 105957531A
- Authority
- CN
- China
- Prior art keywords
- speech
- audio frequency
- voice
- video
- module
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 239000000463 material Substances 0.000 claims abstract description 27
- 230000011218 segmentation Effects 0.000 claims abstract description 25
- 230000006978 adaptation Effects 0.000 claims description 43
- 239000000284 extract Substances 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 11
- 239000004568 cement Substances 0.000 claims description 7
- 230000000694 effects Effects 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 238000001514 detection method Methods 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 230000010354 integration Effects 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 11
- 238000004422 calculation algorithm Methods 0.000 description 9
- 230000007423 decrease Effects 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 230000007935 neutral effect Effects 0.000 description 3
- 241000282326 Felis catus Species 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 230000003203 everyday effect Effects 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 238000007477 logistic regression Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000002354 daily effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000029087 digestion Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/28—Constructional details of speech recognition systems
- G10L15/30—Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Electrically Operated Instructional Devices (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The invention discloses a speech content extracting method and a speech content extracting device based on a cloud platform. The speech content extracting method is characterized in that audio materials and video materials of a speech are acquired, and are cached in a PC for a pretreatment; the pretreated audio materials, the pretreated video materials, related data including speech slides, and related reading materials are transmitted to a server; the server is used for voice segmentation of the received audio materials, and is used for segmenting the audio materials according to speakers; automatic voice identification is used for converting the segmented audio materials into words, and adopts acoustic self-adaption and language model self-adaption; key words are extracted from texts after the voice identification, and content notes are generated. By adopting the speech content extracting method, the audio materials are converted into the text form, which can be read repeatedly, by the audio identification, and the language model self-adaption and the acoustic model self-adaption are used to improve the identification accuracy. Because of knowledge integration, time for reading redundant information is saved. The invention also discloses the speech content extracting device based on the cloud platform. The s speech content extracting device comprises a speech recording module, a material transmitting module, a voice segmenting module, a voice identifying module, and a key word and a content note extracting module.
Description
Technical field
The present invention relates to the technology of a kind of word processing field, a kind of speech content extracting method based on cloud platform
And device.
Background technology
In the information age, the development and progress of technology makes us can obtain every day from information all over the world, through the ages,
The quantity of these information exceeds well over people can listen to the scope of digestion.In order to help the more efficient acquisition information of people, voice signal
Process and natural language processing technique can effectively automatically process magnanimity information, and extract key message therein and content for people
Quickly reading.
In life, everyone listens to substantial amounts of information by the channel such as media, classroom every day, and these information retrievals are become
The textual form that can repeatedly read becomes most important, and it makes the people can quickly reading learning, language model adaptation and acoustics
Model self-adapting method improves the accuracy rate of speech recognition.And carry out Knowledge Integration, it is to avoid time flower is being read in redundancy.
Find that Chinese patent literature CN102292766B discloses one " for speech processes through the retrieval of existing technology
Method and apparatus ", the method and device relate to provide for the method for framework of the adaptive composite model of speech recognition, dress
Putting and computer program, phonetic feature preference pattern based on specific enunciator improves recognition accuracy.But the method does not relates to
And the accuracy rate for specialized vocabulary is improved for language model adaptation.
Retrieve discovery, Chinese patent literature CN102122506A further, disclose one " method of speech recognition ", should
System utilizes the text train language model that search engine retrieving is relevant, it is possible to increase phonetic recognization rate, reduces the work of artificial check and correction
Amount.But the method need to utilize external search engine, the longest, it is unfavorable for processing a large amount of voice.
Summary of the invention
The present invention is directed to deficiencies of the prior art, propose a kind of speech content extracting method based on cloud platform and device,
By speech recognition, audio identification is become the textual form that can repeatedly read, use language model adaptation and acoustics model adaptation to carry
High recognition accuracy.And carry out Knowledge Integration, it is to avoid time flower is being read in redundancy.
The present invention is achieved by the following technical solutions:
The present invention relates to a kind of speech content extracting method based on cloud platform, including:
Step 1) gather the audio frequency and video of speech, by the audio and video buffer that collects to PC, and carry out pretreatment；
Step 2) send pretreated audio frequency and video and related data include giving a lecture lantern slide, relevant reading material waits until server；
Step 3) server to receive audio frequency carry out phonetic segmentation, audio frequency by speaker split；
Step 4) carry out automatic speech recognition the audio conversion after segmentation is changed to word, speech recognition uses acoustics self adaptation and language
Model adaptation；
Step 5) from the text of speech recognition, extract keyword and generate content notes.
Described collection, is preferably used mike, images the audio frequency and video that first-class equipment collection is given a lecture, utilize wired or wireless network
It is cached in PC simultaneously；
Use PC that audio frequency carries out speech enhan-cement and remove noise, and be compressed audio frequency and video processing.
The mode of described phonetic segmentation is that server carries out voice activity detection to the audio frequency received, and carries out according to the pause of voice
Cutting；The described speaker that mode is every section of voice of identification splitting voice by speaker, splits audio frequency by speaker.
Described acoustics self adaptation includes the adaptation to playback environ-ment, noise types, speaker's type etc.；
Described language model adaptation includes the adaptation of specialized vocabulary in courseware and relevant reading material.
Described extraction includes: extract keyword relevant with speech content in the text of speech recognition, and according in text every
The notes relevant with speech are extracted to the degree of association of speech content.
The present invention relates to a kind of speech content extraction element realizing said method, including: it is used for gathering speech audio frequency and video, will adopt
Collect to audio and video buffer in the PC in classroom, and the speech carrying out pretreatment records module, for sending pretreated sound
Video and related data include that speech lantern slide, relevant reading material wait until the material sending module of server, for the sound received
Frequency carries out phonetic segmentation, audio frequency is split module by the voice that speaker is split, for carrying out automatic speech recognition the sound after segmentation
Frequency is converted to word, speech recognition use acoustics self adaptation and the sound identification module of language model adaptation and for server from
Word extracts keyword and generates keyword and the content notes extraction module of content notes.
Described speech is recorded module and is used for using mike, imaging the audio frequency and video that first-class equipment collection is given a lecture, and utilizes wired or nothing
Gauze network is cached in PC simultaneously, uses PC that audio frequency carries out speech enhan-cement and removes noise, and is compressed audio frequency and video processing.
Described phonetic segmentation, for the audio frequency received is carried out voice activity detection, carries out cutting according to the pause of voice；Described
Split voice for identifying the speaker of every section of voice by speaker, split audio frequency by speaker.
Described sound identification module is for using automatic speech recognition to obtain every text corresponding to audio frequency, and described acoustics is adaptive
It is applied to the adaptation to playback environ-ment, noise types, speaker's type etc.；Described language model adaptation is for speech magic lantern
The adaptation of specialized vocabulary in sheet and relevant reading material.
Described keyword and content notes extraction module in the text extracting speech recognition with key that speech content is relevant
Word, and extract and relevant notes of giving a lecture to the degree of association of speech content according in text every.
Technique effect
Compared with prior art, the present invention becomes audio identification by speech recognition the textual form that can repeatedly read, and uses language
Model adaptation and acoustics model adaptation improve recognition accuracy.And carry out Knowledge Integration, it is to avoid time flower is being read redundancy letter
On breath.
Accompanying drawing explanation
Fig. 1 is the inventive method flow chart；
Fig. 2 is apparatus of the present invention structural representations.
Detailed description of the invention
Embodiment 1
The present embodiment comprises the following steps:
101, gather the audio frequency and video of speech, by the audio and video buffer that collects to PC, and carry out pretreatment；
In present example, gather the audio frequency and video of speech, by the audio and video buffer that collects to PC, and carry out pretreatment
Including using mike, imaging the audio frequency and video that first-class equipment collection is given a lecture, utilize wired or wireless network to be cached in PC simultaneously；
Use PC that audio frequency carries out speech enhan-cement and remove noise, and be compressed audio frequency and video processing.
102, send pretreated audio frequency and video and related data includes that speech lantern slide, relevant reading material wait until server；
103, server carries out phonetic segmentation, audio frequency is split by speaker the audio frequency received；
In present example, the mode of described phonetic segmentation is that server carries out voice activity detection to the audio frequency received, and presses
Pause according to voice carries out cutting；The described speaker that mode is every section of voice of identification splitting voice by speaker, by speaker
Segmentation audio frequency.
104, carrying out automatic speech recognition and the audio conversion after segmentation is changed to word, speech recognition uses acoustics self adaptation and language
Model adaptation；
In present example, described acoustics self adaptation includes the adaptation to playback environ-ment, noise types, speaker's type etc.；
Described language model adaptation includes the adaptation of specialized vocabulary in speech lantern slide and relevant reading material.
105, from the text of speech recognition, extract keyword and generate content notes.
In present example, from the text of speech recognition, from the text of speech recognition, extract keyword and generate content notes
Including: extract keyword relevant with speech content in the text of speech recognition, and relevant to speech content according in text every
Degree extracts the notes relevant to speech.
Embodiment 2
As in figure 2 it is shown, the data serching device structural representation based on interactive mode input provided for the embodiment of the present invention, this dress
Put and include: module 21 material sending module 22 voice segmentation module 23 sound identification module 24 and keyword and content are recorded in speech
Notes extraction module 25.
Module 21 is recorded in described speech, is used for gathering speech audio frequency and video, by the PC of the audio and video buffer that collects to classroom
In, and carry out pretreatment；
Described speech is recorded module 21 and is used for using mike, imaging the audio frequency and video that first-class equipment collection is given a lecture, and utilizes wired
Or wireless network is cached in PC simultaneously, uses PC that audio frequency carries out speech enhan-cement and remove noise, and audio frequency and video are compressed place
Reason.
Such as using video camera to record a degree of depth learned lesson, teacher wears clip-on microphone, answer a question
The raw wireless microphone that uses, video that caching is recorded and audio frequency, in the PC in classroom, use filter method such as adaptive cancellation method to remove and carry on the back
Scape sound such as air conditioner noises, construction noise etc., compression Video & Audio makes file size be suitable for network transmission.
Described material sending module 22, is used for sending pretreated audio frequency and video and related data includes give a lecture lantern slide, phase
Close reading material and wait until server.
Specifically, audio frequency and video, degree of depth study lantern slide and degree of depth study reading material after transmission speech enhan-cement, compression wait until
Http server.
Described voice segmentation module 23, for carrying out phonetic segmentation, audio frequency being split by speaker to the audio frequency received.
In described voice segmentation module 23, phonetic segmentation is for carrying out voice activity detection, according to voice to the audio frequency received
Pause carries out cutting；Split voice for identifying the speaker of every section of voice by speaker, split audio frequency by speaker.
Specifically, detect the part being syncopated as voice according to short-time energy and zero-crossing rate, and extract the i-vector of every section of voice
Identify speak artificial teacher and different students.
Described sound identification module 24, is changed to word for carrying out automatic speech recognition the audio conversion after segmentation, and voice is known
Shi Yong acoustics self adaptation and language model adaptation.
Described sound identification module 24 is for using automatic speech recognition to obtain every text corresponding to audio frequency, described acoustics
Self adaptation is for the adaptation to playback environ-ment, noise types, speaker's type etc.；Described language model adaptation is for speech
The adaptation of specialized vocabulary in lantern slide and relevant reading material.
Specifically, during training acoustic model, audio frequency is clustered by i-vector, the audio frequency of each cluster is trained one based on
The acoustic model of deep neural network, finds cluster nearest for its i-vector when identifying audio frequency, and uses this acoustic model clustered.
Use mass text to extract the reverse document-frequency of each word, use TF-IDF statistics degree of depth study courseware and extension to read
In key word.As for extension read " gradient decline (GD) be a kind of common method minimizing risk function, loss function,
It is two kinds of iterative thinkings that stochastic gradient descent and batch gradient decline.Gradient declines in batches---minimize the damage of all training samples
Losing function so that finally solve is the optimal solution of the overall situation, the parameter i.e. solved is so that risk function is minimum.Stochastic gradient descent
---minimize the loss function of every sample, although be not the loss function that obtains of each iteration all towards global optimum direction, but
Big overall direction to globally optimal solution, final result is often near globally optimal solution.", expansion can be extracted out
Open and read the keyword " gradient decline " in reading, " stochastic gradient descent ", " gradient decline in batches ", " loss function " etc., and some are conventional
Word is such as " common method ", and " a kind of ", " minimizing " etc. then can be not listed as key word because TF-IDF weights are the lowest.
When using language model based on recurrent neural network to calculate complexity (perplexity) in short, it is assumed that model parameter is
θ, then the former computing formula of complexity perplexity is:Wherein: N is this sentence
The length of son, for the keyword in this field, then complexity perplexity can be written as:
Work as wiFor the keyword in this field, then q (wi) it is 1, it is otherwise 0.λ is hyper parameter.The method is used to improve right
Discrimination in specialized vocabulary.
Keyword and content notes extraction module 25, extract keyword from word for server and generate content notes.
Described keyword and content notes extraction module 25 in the text extracting speech recognition with pass that speech content is relevant
Key word, and extract and relevant notes of giving a lecture to the degree of association of speech content according in text every.
In this instance, the text later such as through speech recognition be " for a lot of machine learning algorithms, including linear regression,
Logistic regression, neutral net etc., the realization of algorithm is all by showing that certain cost function or certain optimized target are come real
Existing, then use the such method of gradient decline to be used as optimized algorithm and try to achieve the minima of cost function.Training set when us
Time bigger, gradient descent algorithm then seems that amount of calculation is the biggest in batches.Assume that you have the picture of 10,000,000 cats, carry out once batch
Gradient descent algorithm is equivalent to read through this ten million photo, and we need to look for some the most shorter methods to find most cats
Characteristic.In this course, I wants to introduce and a kind of declines different methods with batch gradient: stochastic gradient descent.”
Being similar to, analyzed by TF-IDF, we can draw occur seldom tying this section of speech recognition in daily text
Occurring more word " gradient decline " in Guo, " stochastic gradient descent ", " neutral net " is as key word, and obtains their TF-IDF
Weights.
The weights calculating sentence afterwards are the meansigma methods of each word TF-IDF weights in sentence, and export the sentence work that weights are the highest
Take down notes for content, " for a lot of machine learning algorithms, including linear regression, logistic regression, neutral net etc., the realization of algorithm
All by showing that certain cost function or certain optimized target realize, then use gradient to decline such method and come
The minima of cost function is tried to achieve as optimized algorithm.When our training set is bigger, gradient descent algorithm then seems calculating in batches
Measure the biggest.In this course, I wants to introduce and a kind of declines different methods with batch gradient: stochastic gradient descent.”
The device that the embodiment of the present invention provides, is become audio identification by speech recognition the textual form that can repeatedly read, makes term
Speech model adaptation and acoustics model adaptation improve recognition accuracy.And carry out Knowledge Integration, it is to avoid time flower is being read redundancy
In information.
One of ordinary skill in the art will appreciate that all or part of step realizing above-described embodiment can be completed by hardware,
Can also instruct relevant hardware by program to complete, this program can be stored in a kind of computer-readable recording medium, above-mentioned
The storage medium mentioned can be read only memory, disk or CD etc..
Above-mentioned be embodied as can by those skilled in the art on the premise of without departing substantially from the principle of the invention and objective in a different manner
It is carried out local directed complete set, and protection scope of the present invention is as the criterion with claims and is not limited by above-mentioned being embodied as, in its scope
Each interior implementation is all by the constraint of the present invention.
Claims (10)
1. a speech content extracting method based on cloud platform, it is characterised in that including:
Step 1) gather the audio frequency and video of speech, by the audio and video buffer that collects to PC, and carry out pretreatment；
Step 2) send pretreated audio frequency and video and related data include giving a lecture lantern slide, relevant reading material waits until server；
Step 3) server to receive audio frequency carry out phonetic segmentation, audio frequency by speaker split；
Step 4) carry out automatic speech recognition the audio conversion after segmentation is changed to word, speech recognition uses acoustics self adaptation and language
Model adaptation；
Step 5) from the text of speech recognition, extract keyword and generate content notes.
Method the most according to claim 1, is characterized in that, described collection includes: uses mike, image first-class setting
The standby audio frequency and video gathering speech, utilize wired or wireless network to be cached in PC simultaneously；Use PC that audio frequency is carried out speech enhan-cement
Except noise, and it is compressed audio frequency and video processing.
Method the most according to claim 1, is characterized in that, the mode of described phonetic segmentation is the server sound to receiving
Frequency carries out voice activity detection, carries out cutting according to the pause of voice；Described splits the mode of voice for identifying every section by speaker
The speaker of voice, splits audio frequency by speaker.
Method the most according to claim 1, is characterized in that, described acoustics self adaptation includes playback environ-ment, noise-like
The adaptation of type, speaker's type etc.；Described language model adaptation includes specialty word in speech lantern slide and relevant reading material
The adaptation converged.
Method the most according to claim 1, is characterized in that, described extraction includes: extract speech recognition text in
The keyword that speech content is relevant, and extract and relevant notes of giving a lecture to the degree of association of speech content according in text every.
6. the speech content extraction element realizing method described in any of the above-described claim, it is characterised in that including:
Module is recorded in speech, is used for gathering speech audio frequency and video, by the audio and video buffer that collects to the PC in classroom, and carries out
Pretreatment,
Material sending module, be used for sending pretreated audio frequency and video and speech lantern slide, relevant reading material to server,
Voice segmentation module, for receive audio frequency carry out phonetic segmentation, audio frequency by speaker segmentation,
Sound identification module, is changed to word for carrying out automatic speech recognition the audio conversion after segmentation, and speech recognition uses acoustics
Self adaptation and language model adaptation,
Keyword and content notes extraction module, extract keyword from word for server and generate content notes.
Device the most according to claim 6, is characterized in that, described speech is recorded module and adopted by mike, photographic head
The audio frequency and video of collection speech, utilize wired or wireless network to be cached in PC simultaneously, use PC that audio frequency carries out speech enhan-cement removal and make an uproar
Sound, and be compressed audio frequency and video processing.
Device the most according to claim 6, is characterized in that, described phonetic segmentation carries out speech activity to the audio frequency received
Detection, carries out cutting according to the pause of voice；Described splits voice for identifying the speaker of every section of voice, normally by speaker
Words people splits audio frequency.
Device the most according to claim 6, is characterized in that, described sound identification module is used for using automatic speech recognition
Obtaining every text corresponding to audio frequency, described acoustics self adaptation is for playback environ-ment, noise types, the adaptation of speaker's type；
Described language model adaptation is for the adaptation of specialized vocabulary in speech lantern slide and relevant reading material.
Device the most according to claim 6, is characterized in that, described keyword and content notes extraction module are used for carrying
Take keyword relevant with speech content in the text of speech recognition, and according to the degree of association of in text every with speech content extract with
The notes that speech is relevant.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610260647.7A CN105957531B (en) | 2016-04-25 | 2016-04-25 | Speech content extraction method and device based on cloud platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610260647.7A CN105957531B (en) | 2016-04-25 | 2016-04-25 | Speech content extraction method and device based on cloud platform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105957531A true CN105957531A (en) | 2016-09-21 |
| CN105957531B CN105957531B (en) | 2019-12-31 |
Family
ID=56915289
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610260647.7A Active CN105957531B (en) | 2016-04-25 | 2016-04-25 | Speech content extraction method and device based on cloud platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105957531B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107818797A (en) * | 2017-12-07 | 2018-03-20 | 苏州科达科技股份有限公司 | Voice quality assessment method, apparatus and its system |
| CN108022583A (en) * | 2017-11-17 | 2018-05-11 | 平安科技（深圳）有限公司 | Meeting summary generation method, application server and computer-readable recording medium |
| CN108256512A (en) * | 2018-03-22 | 2018-07-06 | 长春大学 | Listen the raw inclusive education classroom auxiliary system of barrier and device |
| CN108335693A (en) * | 2017-01-17 | 2018-07-27 | 腾讯科技（深圳）有限公司 | A kind of Language Identification and languages identification equipment |
| CN108597521A (en) * | 2018-05-04 | 2018-09-28 | 徐涌 | Audio role divides interactive system, method, terminal and the medium with identification word |
| CN109934188A (en) * | 2019-03-19 | 2019-06-25 | 上海大学 | A kind of lantern slide switching detection method, system, terminal and storage medium |
| WO2019237708A1 (en) * | 2018-06-15 | 2019-12-19 | 山东大学 | Interpersonal interaction body language automatic generation method and system based on deep learning |
| WO2020103447A1 (en) * | 2018-11-21 | 2020-05-28 | 平安科技（深圳）有限公司 | Link-type storage method and apparatus for video information, computer device and storage medium |
| CN111723816A (en) * | 2020-06-28 | 2020-09-29 | 北京联想软件有限公司 | Teaching note acquisition method and electronic equipment |
| CN111897918A (en) * | 2020-07-28 | 2020-11-06 | 扬州大学 | Online teaching classroom note generation method |
| CN111932964A (en) * | 2020-08-21 | 2020-11-13 | 扬州大学 | Online live broadcast teaching method |
| CN112767753A (en) * | 2021-01-08 | 2021-05-07 | 中国石油大学胜利学院 | Supervision type intelligent online teaching system and action method thereof |
| CN114501112A (en) * | 2022-01-24 | 2022-05-13 | 北京百度网讯科技有限公司 | Method, apparatus, device, medium and product for generating video notes |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1881415A (en) * | 2003-08-15 | 2006-12-20 | 株式会社东芝 | Information processing apparatus and method therefor |
| CN101923854A (en) * | 2010-08-31 | 2010-12-22 | 中国科学院计算技术研究所 | Interactive speech recognition system and method |
| CN102280106A (en) * | 2010-06-12 | 2011-12-14 | 三星电子株式会社 | VWS method and apparatus used for mobile communication terminal |
| CN103165130A (en) * | 2013-02-06 | 2013-06-19 | 湘潭安道致胜信息科技有限公司 | Voice text matching cloud system |
| CN105159870A (en) * | 2015-06-26 | 2015-12-16 | 徐信 | Processing system for precisely completing continuous natural speech textualization and method for precisely completing continuous natural speech textualization |
| CN105427858A (en) * | 2015-11-06 | 2016-03-23 | 科大讯飞股份有限公司 | Method and system for achieving automatic voice classification |
-
2016
- 2016-04-25 CN CN201610260647.7A patent/CN105957531B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1881415A (en) * | 2003-08-15 | 2006-12-20 | 株式会社东芝 | Information processing apparatus and method therefor |
| CN102280106A (en) * | 2010-06-12 | 2011-12-14 | 三星电子株式会社 | VWS method and apparatus used for mobile communication terminal |
| CN101923854A (en) * | 2010-08-31 | 2010-12-22 | 中国科学院计算技术研究所 | Interactive speech recognition system and method |
| CN103165130A (en) * | 2013-02-06 | 2013-06-19 | 湘潭安道致胜信息科技有限公司 | Voice text matching cloud system |
| CN105159870A (en) * | 2015-06-26 | 2015-12-16 | 徐信 | Processing system for precisely completing continuous natural speech textualization and method for precisely completing continuous natural speech textualization |
| CN105427858A (en) * | 2015-11-06 | 2016-03-23 | 科大讯飞股份有限公司 | Method and system for achieving automatic voice classification |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108335693B (en) * | 2017-01-17 | 2022-02-25 | 腾讯科技（深圳）有限公司 | Language identification method and language identification equipment |
| CN108335693A (en) * | 2017-01-17 | 2018-07-27 | 腾讯科技（深圳）有限公司 | A kind of Language Identification and languages identification equipment |
| CN108022583A (en) * | 2017-11-17 | 2018-05-11 | 平安科技（深圳）有限公司 | Meeting summary generation method, application server and computer-readable recording medium |
| CN107818797A (en) * | 2017-12-07 | 2018-03-20 | 苏州科达科技股份有限公司 | Voice quality assessment method, apparatus and its system |
| CN108256512A (en) * | 2018-03-22 | 2018-07-06 | 长春大学 | Listen the raw inclusive education classroom auxiliary system of barrier and device |
| CN108597521A (en) * | 2018-05-04 | 2018-09-28 | 徐涌 | Audio role divides interactive system, method, terminal and the medium with identification word |
| WO2019237708A1 (en) * | 2018-06-15 | 2019-12-19 | 山东大学 | Interpersonal interaction body language automatic generation method and system based on deep learning |
| WO2020103447A1 (en) * | 2018-11-21 | 2020-05-28 | 平安科技（深圳）有限公司 | Link-type storage method and apparatus for video information, computer device and storage medium |
| CN109934188A (en) * | 2019-03-19 | 2019-06-25 | 上海大学 | A kind of lantern slide switching detection method, system, terminal and storage medium |
| CN109934188B (en) * | 2019-03-19 | 2020-10-30 | 上海大学 | Slide switching detection method, system, terminal and storage medium |
| CN111723816A (en) * | 2020-06-28 | 2020-09-29 | 北京联想软件有限公司 | Teaching note acquisition method and electronic equipment |
| CN111723816B (en) * | 2020-06-28 | 2023-10-27 | 北京联想软件有限公司 | Acquisition method of teaching notes and electronic equipment |
| CN111897918A (en) * | 2020-07-28 | 2020-11-06 | 扬州大学 | Online teaching classroom note generation method |
| CN111932964A (en) * | 2020-08-21 | 2020-11-13 | 扬州大学 | Online live broadcast teaching method |
| CN112767753A (en) * | 2021-01-08 | 2021-05-07 | 中国石油大学胜利学院 | Supervision type intelligent online teaching system and action method thereof |
| CN112767753B (en) * | 2021-01-08 | 2022-07-22 | 中国石油大学胜利学院 | Supervision type intelligent online teaching system and action method thereof |
| CN114501112A (en) * | 2022-01-24 | 2022-05-13 | 北京百度网讯科技有限公司 | Method, apparatus, device, medium and product for generating video notes |
| CN114501112B (en) * | 2022-01-24 | 2024-03-22 | 北京百度网讯科技有限公司 | Method, apparatus, device, medium, and article for generating video notes |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105957531B (en) | 2019-12-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105957531A (en) | Speech content extracting method and speech content extracting device based on cloud platform | |
| CN108399923B (en) | More human hairs call the turn spokesman's recognition methods and device | |
| CN102436812B (en) | Conference recording device and conference recording method using same | |
| CN108305632A (en) | A kind of the voice abstract forming method and system of meeting | |
| US20100057452A1 (en) | Speech interfaces | |
| CN109192224A (en) | A kind of speech evaluating method, device, equipment and readable storage medium storing program for executing | |
| Ding et al. | Audio-visual keyword spotting based on multidimensional convolutional neural network | |
| CN110728991B (en) | Improved recording equipment identification algorithm | |
| WO2017166483A1 (en) | Method and system for processing dynamic picture | |
| CN109712612A (en) | A kind of voice keyword detection method and device | |
| CN110970036A (en) | Voiceprint recognition method and device, computer storage medium and electronic equipment | |
| CN109452932A (en) | A kind of Constitution Identification method and apparatus based on sound | |
| CN116246610A (en) | Conference record generation method and system based on multi-mode identification | |
| CN206672635U (en) | A kind of voice interaction device based on book service robot | |
| CN108364655A (en) | Method of speech processing, medium, device and computing device | |
| CN113053361B (en) | Speech recognition method, model training method, device, equipment and medium | |
| CN114330454A (en) | Live pig cough sound identification method based on DS evidence theory fusion characteristics | |
| KR20170086233A (en) | Method for incremental training of acoustic and language model using life speech and image logs | |
| CN117078094A (en) | Teacher comprehensive ability assessment method based on artificial intelligence | |
| Wan | Research on speech separation and recognition algorithm based on deep learning | |
| Blunt et al. | A model for incorporating an automatic speech recognition system in a noisy educational environment | |
| KR102429365B1 (en) | System and method for analyzing emotion of speech | |
| CN112837688B (en) | Voice transcription method, device, related system and equipment | |
| CN116524910B (en) | Manuscript prefabrication method and system based on microphone | |
| Donai et al. | Classification of indexical and segmental features of human speech using low-and high-frequency energy |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200617 Address after: Room 105G, 199 GuoShoujing Road, Pudong New Area, Shanghai, 200120 Patentee after: Shanghai Jiaotong University Intellectual Property Management Co.,Ltd. Address before: 200240 Dongchuan Road, Shanghai, No. 800, No. Patentee before: SHANGHAI JIAO TONG University |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20201028 Address after: Building 14, Tengfei Innovation Park, no.388 Xinping street, Suzhou Industrial Park, Jiangsu Province, 215000 Patentee after: AI SPEECH Ltd. Address before: Room 105G, 199 GuoShoujing Road, Pudong New Area, Shanghai, 200120 Patentee before: Shanghai Jiaotong University Intellectual Property Management Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 215000 building 14, Tengfei Innovation Park, 388 Xinping street, Suzhou Industrial Park, Suzhou City, Jiangsu Province Patentee after: Sipic Technology Co.,Ltd. Address before: 215000 building 14, Tengfei Innovation Park, 388 Xinping street, Suzhou Industrial Park, Suzhou City, Jiangsu Province Patentee before: AI SPEECH Ltd. |
|
| CP01 | Change in the name or title of a patent holder |