CN108091325A - A kind of speech recognition system and method based on surname - Google Patents
A kind of speech recognition system and method based on surname Download PDFInfo
- Publication number
- CN108091325A CN108091325A CN201711440127.5A CN201711440127A CN108091325A CN 108091325 A CN108091325 A CN 108091325A CN 201711440127 A CN201711440127 A CN 201711440127A CN 108091325 A CN108091325 A CN 108091325A
- Authority
- CN
- China
- Prior art keywords
- surname
- phonetic
- unit
- list
- chinese character
- 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
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/02—Feature extraction for speech recognition; Selection of recognition unit
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
- G10L15/063—Training
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/18—Speech classification or search using natural language modelling
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L2015/086—Recognition of spelled words
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)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Document Processing Apparatus (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a kind of speech recognition system and method based on surname, including main control unit, voice acquisition unit, Audio Processing Unit, surname model unit and pre-training unit：The main control unit connects voice acquisition unit, Audio Processing Unit, surname model unit and pre-training unit respectively, the present invention is not only simple and convenient, effectively realize the identification of the quick surname Chinese character of high-accuracy, simultaneously confusing a series of processing mode may be designed for surname is a variety of, the performance and discrimination of speech recognition are improved, alleviates system-computed burden.
Description
Technical field
The present invention relates to technical field of voice recognition more particularly to a kind of speech recognition systems and method based on surname.
Background technology
Language is that the mankind mutually exchange the most frequently used, most effective, most important and most convenient communication form, and voice is language
Acoustics shows, and it is the dream of the mankind all the time to carry out speech exchange with machine.With the rapid development of computer technology, voice
The achievement of identification technology also making a breakthrough property, the dream that people is engaged in the dialogue with machine with natural language is progressively close to realizing.Voice
The application range of identification technology is extremely wide, is not only related to the every aspect of daily life, and pole is also played in military field
Its important role.It is information-intensive society towards intelligent and automation development key technology, makes processing of the people to information
It is more convenient with obtaining, so as to improve the work efficiency of people.
The content of the invention
It is an object of the invention to provide a kind of speech recognition system and method based on surname, to solve above-mentioned background skill
The problem of being proposed in art.
To achieve the above object, the present invention provides following technical solution：
A kind of speech recognition system based on surname, including main control unit, voice acquisition unit, Audio Processing Unit, surname
Model unit and pre-training unit：The main control unit connects voice acquisition unit, Audio Processing Unit, surname model respectively
Unit and pre-training unit.
Further technical solution as the present invention：The voice acquisition unit is M6027 microphones.
Further technical solution as the present invention：The Audio Processing Unit be divided into acoustic model characteristic extracting module,
Language model characteristic extracting module obscures processing module and list Shuan Xingshi processing modules.
A kind of audio recognition method based on surname, comprises the steps of：
A, user inputs a string of voices；
B, system obtains voice signal, carries out feature extraction to voice by acoustic model, pronunciation sequence is drawn in acoustic model
Afterwards, the character string sequence of maximum probability is found out from candidate character sequence using language model；
C, the Chinese character for representing surname is extracted from chinese character string sequence, by the whole Chinese characters and its spelling book of collection,
Chinese character can be converted to phonetic, if it is polyphone, only be converted to wherein some phonetic；
If D, the surname is individual character surname, its phonetic is added in into list list, is done for surname phonetic and easily obscures pronunciation
Processing and the processing for searching polyphone, and result is all saved in list；
If E, the surname is double word surname, the phonetic of each word in two word surnames is individually taken out to progress and easily obscures pronunciation
Processing and lookup multitone, each word can obtain a list, two lists are combined two-by-two, find out these pinyin-groups
Close corresponding all surnames；
F, according to obtained phonetic list, the corresponding surname Chinese character of each phonetic can be found, is listed complete expressed by speaker
It portion may surname Chinese character；
G, it is final to obtain corresponding surname Chinese character list.
Compared with prior art, the beneficial effects of the invention are as follows：The present invention is not only simple and convenient, effectively realizes high precision
The identification of the quick surname Chinese character of rate, at the same for surname it is a variety of it is confusing may design a series of processing mode, improve
The performance and discrimination of speech recognition alleviate system-computed burden.
Description of the drawings
Fig. 1 is a kind of structure diagram of the speech recognition system based on surname
Fig. 2 is a kind of flow chart of the audio recognition method based on surname.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other without making creative work
Embodiment belongs to the scope of protection of the invention.
Please refer to Fig.1-2, in the embodiment of the present invention, a kind of speech recognition system based on surname is obtained including voice
Unit, Audio Processing Unit, surname model unit and pre-training unit.
Voice acquisition unit：The unit is responsible for adopting the order progress voice signal that user sends using M6027 microphones
Sample, and user voice signal is sent to Audio Processing Unit and is handled, speech recognition is passed to, is converted to chinese character
String.
Audio Processing Unit：Unit is divided into acoustic model characteristic extracting module, language model characteristic extracting module, obscures place
Manage module and list Shuan Xingshi processing modules.
Wherein acoustic model characteristic extracting module is to carry out feature extraction to voice signal using acoustic model, and voice is turned
The output of acoustics expression is changed to, it is the probability for belonging to some delimiter symbol to provide voice；
Language model characteristic extracting module is after acoustic model provides pronunciation sequence, and probability is found out from candidate character sequence most
Big character string sequence；
Obscure there are four types of processing form in processing module, one is that flat tongue consonant and cacuminal are handled, and for the phonetic of surname, judgement is
No, with z, s, either if c beginnings are started with z, s or c, continue to judge to whether there is h in phonetic, if in the presence of removing；If no
In the presence of then in second position of phonetic plus h；Two are handled for pre-nasal sound and rear nasal sound, for rear nasal sound：Alphabetical g generally goes out
Either whether ending is judged comprising ang, eng or ing in surname phonetic for the beginning of present phonetic, if comprising removing phonetic
The g of ending, for pre-nasal sound：An, en and in are generally present in the ending of phonetic, judge to whether there is an, en in surname phonetic
Or in, if in the presence of whether last position for continuing to judge phonetic is g, if it is not, then adding g in the ending of phonetic.Three be nose
Sound n and lateral l processing, whether the beginning for judging surname phonetic is n, if n, then the n of beginning is changed to l.Judge opening for phonetic
Whether head is l, and n is changed to if l, the then l started；Four be multitone word processing, and surname Chinese character is being converted to the process of phonetic, if
The surname that user says is polyphone, and it does not read according to correct phonetic of the polyphone in surname, and system is according to 35
The correct surname pronunciation of polyphone and the list of other pronunciations, judge institute's input Pinyin whether in other pronunciations, if at other
In pronunciation, then correct surname pronunciation is taken out.The list of the pronunciation of correct surname and other pronunciations of 35 polyphones is by being system
All polyphones in surname are found out in the list of existing whole Chinese character and its phonetic, wherein there are 28 multitones in individual character surname
Word has 7 polyphones in double word surname.
The phonetic of individual character surname wherein for Dan Xingshi, is first added in list list, for surname by single Shuan Xingshi processing modules
Family name's phonetic does the processing for easily obscuring pronunciation and the processing for searching polyphone, and result is all saved in list.And for
The phonetic of each word in two word surnames is individually taken out the processing for carrying out easily obscuring pronunciation and searches multitone, each word by Shuan Xingshi
A list can be all obtained, two lists are combined two-by-two, these pinyin combinations is found out and corresponds to all surnames.
Surname model unit：The whole Chinese characters and its spelling book of collection, will cover Chinese character as much as possible, and with this
Build surname model.
Pre-training unit：The recognition mode of the surname of training extraction in advance is carried out by the structure surname model of system constructing,
Carry out the extraction of surname Chinese character.
The present invention operation principle be：Its workflow is as shown in Figure 2：
1. user inputs a string of voices.
2. system obtains voice signal, feature extraction is carried out to voice by acoustic model, pronunciation is drawn in acoustic model
After sequence, the character string sequence of maximum probability is found out from candidate character sequence using language model.
3. extracting the Chinese character for representing surname from chinese character string sequence, pass through whole Chinese characters of collection and its phonetic word
Chinese character can be converted to phonetic by allusion quotation, if it is polyphone, only be converted to wherein some phonetic.
4. if the surname is individual character surname, its phonetic is added in into list list, is done for surname phonetic and easily obscures hair
The processing of sound and the processing for searching polyphone, and result is all saved in list.
5. if the surname is double word surname, the phonetic of each word in two word surnames is individually taken out to progress and easily obscures hair
The processing of sound and lookup multitone, each word can obtain a list, two lists are combined two-by-two, find out these spellings
The corresponding all surnames of sound combination.
6. according to obtained phonetic list, the corresponding surname Chinese character of each phonetic can be found, is listed expressed by speaker
All may surname Chinese character.
7. final obtain corresponding surname Chinese character list.
Claims (4)
1. a kind of speech recognition system based on surname, including main control unit, voice acquisition unit, Audio Processing Unit, surname
Family name's model unit and pre-training unit：It is characterized in that, the main control unit connects voice acquisition unit, speech processes respectively
Unit, surname model unit and pre-training unit.
2. a kind of speech recognition system based on surname according to claim 1, which is characterized in that the voice obtains single
Member is M6027 microphones.
A kind of 3. speech recognition system based on surname according to claim 1, which is characterized in that the speech processes list
Member is divided into acoustic model characteristic extracting module, language model characteristic extracting module, obscures processing module and list Shuan Xingshi processing moulds
Block.
4. a kind of audio recognition method based on surname, which is characterized in that comprise the steps of：
A, user inputs a string of voices；
B, system obtains voice signal, carries out feature extraction to voice by acoustic model, pronunciation sequence is drawn in acoustic model
Afterwards, the character string sequence of maximum probability is found out from candidate character sequence using language model；
C, the Chinese character for representing surname is extracted from chinese character string sequence, by the whole Chinese characters and its spelling book of collection,
Chinese character can be converted to phonetic, if it is polyphone, only be converted to wherein some phonetic；
If D, the surname is individual character surname, its phonetic is added in into list list, is done for surname phonetic and easily obscures pronunciation
Processing and the processing for searching polyphone, and result is all saved in list；
If E, the surname is double word surname, the phonetic of each word in two word surnames is individually taken out to progress and easily obscures pronunciation
Processing and lookup multitone, each word can obtain a list, two lists are combined two-by-two, find out these pinyin-groups
Close corresponding all surnames；
F, according to obtained phonetic list, the corresponding surname Chinese character of each phonetic can be found, is listed complete expressed by speaker
It portion may surname Chinese character；
G, it is final to obtain corresponding surname Chinese character list.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711440127.5A CN108091325A (en) | 2017-12-27 | 2017-12-27 | A kind of speech recognition system and method based on surname |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711440127.5A CN108091325A (en) | 2017-12-27 | 2017-12-27 | A kind of speech recognition system and method based on surname |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108091325A true CN108091325A (en) | 2018-05-29 |
Family
ID=62178515
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711440127.5A Pending CN108091325A (en) | 2017-12-27 | 2017-12-27 | A kind of speech recognition system and method based on surname |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108091325A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110097880A (en) * | 2019-04-20 | 2019-08-06 | 广东小天才科技有限公司 | A kind of answer determination method and device based on speech recognition |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1342942A (en) * | 2000-09-08 | 2002-04-03 | 百度在线网络技术（北京）有限公司 | Computer recognizing and indexing method of Chinese names |
| CN103578467A (en) * | 2013-10-18 | 2014-02-12 | 威盛电子股份有限公司 | Acoustic model building method, voice recognition method and electronic device |
| CN104331148A (en) * | 2014-09-23 | 2015-02-04 | 普强信息技术（北京）有限公司 | Voice user interface information interaction method |
| CN105988989A (en) * | 2015-02-26 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Chinese surname recognition method and device, as well as server |
| CN106447346A (en) * | 2016-08-29 | 2017-02-22 | 北京中电普华信息技术有限公司 | Method and system for construction of intelligent electric power customer service system |
-
2017
- 2017-12-27 CN CN201711440127.5A patent/CN108091325A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1342942A (en) * | 2000-09-08 | 2002-04-03 | 百度在线网络技术（北京）有限公司 | Computer recognizing and indexing method of Chinese names |
| CN103578467A (en) * | 2013-10-18 | 2014-02-12 | 威盛电子股份有限公司 | Acoustic model building method, voice recognition method and electronic device |
| CN104331148A (en) * | 2014-09-23 | 2015-02-04 | 普强信息技术（北京）有限公司 | Voice user interface information interaction method |
| CN105988989A (en) * | 2015-02-26 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Chinese surname recognition method and device, as well as server |
| CN106447346A (en) * | 2016-08-29 | 2017-02-22 | 北京中电普华信息技术有限公司 | Method and system for construction of intelligent electric power customer service system |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110097880A (en) * | 2019-04-20 | 2019-08-06 | 广东小天才科技有限公司 | A kind of answer determination method and device based on speech recognition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105957518B (en) | A kind of method of Mongol large vocabulary continuous speech recognition | |
| CN109829058B (en) | Classification recognition method for improving dialect recognition accuracy based on multitask learning | |
| CN110675854B (en) | Chinese and English mixed speech recognition method and device | |
| CN108847241A (en) | It is method, electronic equipment and the storage medium of text by meeting speech recognition | |
| CN104575497B (en) | A kind of acoustic model method for building up and the tone decoding method based on the model | |
| CN104078044A (en) | Mobile terminal and sound recording search method and device of mobile terminal | |
| WO2003010754A1 (en) | Speech input search system | |
| CN104485107B (en) | Audio recognition method, speech recognition system and the speech recognition apparatus of title | |
| CN109256150A (en) | Speech emotion recognition system and method based on machine learning | |
| CN110277088B (en) | Intelligent voice recognition method, intelligent voice recognition device and computer readable storage medium | |
| CN103632663B (en) | A kind of method of Mongol phonetic synthesis front-end processing based on HMM | |
| CN102063900A (en) | Speech recognition method and system for overcoming confusing pronunciation | |
| CN112818680B (en) | Corpus processing method and device, electronic equipment and computer readable storage medium | |
| CN103219007A (en) | Voice recognition method and voice recognition device | |
| CN104199825A (en) | Information inquiry method and system | |
| CN109102800A (en) | A kind of method and apparatus that the determining lyrics show data | |
| CN111192572A (en) | Semantic recognition method, device and system | |
| CN111933116B (en) | Speech recognition model training method, system, mobile terminal and storage medium | |
| CN110390929A (en) | Chinese and English civil aviaton land sky call acoustic model construction method based on CDNN-HMM | |
| CN108091325A (en) | A kind of speech recognition system and method based on surname | |
| CN112489634A (en) | Language acoustic model training method and device, electronic equipment and computer medium | |
| CN107562907A (en) | A kind of intelligent lawyer's expert system and case answering device | |
| CN109213988A (en) | Barrage subject distillation method, medium, equipment and system based on N-gram model | |
| CN104424942A (en) | Method for improving character speed input accuracy | |
| CN109523992A (en) | Tibetan dialect speech processing system |
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: 20180529 |
|
| RJ01 | Rejection of invention patent application after publication |