CN112509581B - Error correction method and device for text after voice recognition, readable medium and electronic equipment - Google Patents
Error correction method and device for text after voice recognition, readable medium and electronic equipment Download PDFInfo
- Publication number
- CN112509581B CN112509581B CN202011311695.7A CN202011311695A CN112509581B CN 112509581 B CN112509581 B CN 112509581B CN 202011311695 A CN202011311695 A CN 202011311695A CN 112509581 B CN112509581 B CN 112509581B
- Authority
- CN
- China
- Prior art keywords
- word
- corrected
- text
- correcting
- voice recognition
- 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.)
- Active
Links
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
- G10L15/26—Speech to text systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/242—Dictionaries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/253—Grammatical analysis; Style critique
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Human Computer Interaction (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Probability & Statistics with Applications (AREA)
- Machine Translation (AREA)
Abstract
The disclosure relates to an error correction method and device for text after voice recognition, a readable medium and electronic equipment, belongs to the field of voice recognition, and can improve the sentence accuracy of voice recognition. A method of error correction of text after speech recognition, comprising: based on the expression characteristics of the word text, checking whether the text after the voice recognition belongs to the word text; extracting a word to be corrected and a word for correcting the word to be corrected from the text after the voice recognition if the text after the voice recognition belongs to the word text in the word, wherein the relationship between the word to be corrected and the word for correcting the word to be corrected is the word in the word; searching candidate words for correcting the word to be corrected from words to be corrected by utilizing a pronunciation confusion matrix, wherein the pronunciation confusion matrix comprises the probability that each pronunciation is identified as other pronunciation; and correcting the word to be corrected by using the candidate word.
Description
Technical Field
The present disclosure relates to the field of speech recognition, and in particular, to a method and apparatus for correcting text error after speech recognition, a readable medium, and an electronic device.
Background
With the popularization of intelligent devices and the development of natural language processing technology, voice input is becoming an increasingly important human-computer interaction means due to the characteristics of convenience and rapidness. However, due to the complex diversity of languages and the influence of ambient noise, the speech recognition result often has a larger deviation from what the user actually wants to input, and thus further error correction processing is required for the text after speech recognition, so that the speech recognition result can be applied to an actual system.
Therefore, how to provide a text error correction scheme after voice recognition can effectively solve the problem of inaccurate voice recognition, and is a technical problem to be solved urgently at present.
Disclosure of Invention
This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.
In a first aspect, the present disclosure provides a method for correcting text after speech recognition, including: based on the expression characteristics of the word text, checking whether the text after the voice recognition belongs to the word text; extracting a word to be corrected and a word for correcting the word to be corrected from the text after the voice recognition if the text after the voice recognition belongs to the word-in-word text, wherein the word to be corrected and the word for correcting the word to be corrected are in word-in-word relation; searching candidate words for correcting the word to be corrected from the words to be corrected by utilizing a pronunciation confusion matrix, wherein the pronunciation confusion matrix comprises the probability that each pronunciation is identified as other pronunciation; and correcting the word to be corrected by using the candidate word.
In a second aspect, the present disclosure provides an error correction apparatus for text after speech recognition, including: the checking module is used for checking whether the text after the voice recognition belongs to the word text or not based on the expression characteristics of the word text; the extraction module is used for extracting a word to be corrected and a word for correcting the word to be corrected from the text after the voice recognition if the text after the voice recognition belongs to the word text, wherein the relationship between the word to be corrected and the word for correcting the word to be corrected is a word-in-word relationship; a searching module, configured to search for a candidate word for correcting the word to be corrected from the word to be corrected by using a pronunciation confusion matrix, where the pronunciation confusion matrix includes a probability that each pronunciation is identified as another pronunciation; and the error correction module is used for correcting the word needing to be corrected by utilizing the candidate word.
In a third aspect, the present disclosure provides a computer readable medium having stored thereon a computer program which when executed by a processing device performs the steps of the method of the first aspect of the present disclosure.
In a fourth aspect, the present disclosure provides an electronic device comprising: a storage device having a computer program stored thereon; processing means for executing said computer program in said storage means to carry out the steps of the method of the first aspect of the disclosure.
By adopting the technical scheme, firstly, whether the text after voice recognition belongs to the word text or not is checked based on the expression characteristics of the word text, then if the text after voice recognition belongs to the word text, the word needing error correction and the word used for correcting the word needing error correction are extracted from the text after voice recognition, then candidate words used for correcting the word needing error correction are searched from the word needing error correction by utilizing a pronunciation confusion matrix, finally the word needing error correction is corrected by utilizing the candidate words, so that the word needing error correction can be corrected by utilizing the word with correct voice recognition in the word text after voice recognition, thereby greatly improving the sentence accuracy of voice recognition, correcting the word in the recognition result and improving the user experience of voice recognition.
Additional features and advantages of the present disclosure will be set forth in the detailed description which follows.
Drawings
The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. The same or similar reference numbers will be used throughout the drawings to refer to the same or like elements. It should be understood that the figures are schematic and that elements and components are not necessarily drawn to scale. In the drawings:
fig. 1 is a flow chart of a method of error correction of text after speech recognition according to one embodiment of the present disclosure.
Fig. 2 is a schematic block diagram of an error correction apparatus for text after speech recognition according to one embodiment of the present disclosure.
Fig. 3 is a schematic structural view of an electronic device according to an embodiment of the present disclosure.
Detailed Description
Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure have been shown in the accompanying drawings, it is to be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for illustration purposes only and are not intended to limit the scope of the present disclosure.
It should be understood that the various steps recited in the method embodiments of the present disclosure may be performed in a different order and/or performed in parallel. Furthermore, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.
The term "including" and variations thereof as used herein are intended to be open-ended, i.e., including, but not limited to. The term "based on" is based at least in part on. The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments. Related definitions of other terms will be given in the description below.
It should be noted that the terms "first," "second," and the like in this disclosure are merely used to distinguish between different devices, modules, or units and are not used to define an order or interdependence of functions performed by the devices, modules, or units.
It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.
The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.
Fig. 1 is a flow chart of a method of error correction of text after speech recognition according to one embodiment of the present disclosure. As shown in fig. 1, the method includes the following steps S11 to S14.
In step S11, it is checked whether the text after speech recognition belongs to the word-in-word text based on the expression characteristics of the word-in-word text.
The word text in a word typically has the expression of, for example, "B … of … a", where a represents a word, such as idioms, verses, etc., and B represents the word in a. For example, "please love house can group what words", "how two Orioles are singing for green's Orioles are writing", "what meaning is looking for in the three-way, whether the number is the third or fourth sound in the nine-cold days", "what is the winding pinyin", "how many strokes are right in the winding", "how is the wake feeling is writing", etc., these are word-in-word text, where "loving" and "two Orunder" and "three-year" and "nine-day" and "winding" and "correct" and "wake" are words in word-in-word text, and "house", "Orunder", "looking" and "number", "winding", "positive" and "feel" are words in the aforementioned words.
If the text after the voice recognition accords with the expression characteristics of the word text in words, the text after the voice recognition is considered to belong to the word text in words, otherwise, the text after the voice recognition is considered not to belong to the word text in words. If it is not considered to belong to word-in-word text, no subsequent steps are performed, but rather the examination of other speech recognized text is continued.
In speech recognition, words in the word text are usually recognized correctly, and words in the word text as described above may be recognized incorrectly. For example, "what words can be grouped in a loving house" may be recognized as "what words can be grouped in a loving house", "how two Orioles write how many times the Oriole green willow" may be recognized as "how many times the two Oriole green willow write how many times apart". Therefore, it is necessary to correct the text of the word after the speech recognition, so that the result of the speech recognition becomes more accurate and reasonable.
In step S12, if the text after speech recognition belongs to the word-in-word text, a word to be corrected and a word for correcting the word to be corrected are extracted from the text after speech recognition, wherein the relationship between the word to be corrected and the word for correcting the word to be corrected is the word-in-word relationship.
In this step, extraction is mainly based on the expression characteristics of the text of the word. For example, the word in the text after speech recognition may be first boundary-divided based on the expression characteristics of the word text in the word; then, based on the expression characteristics of the text of the word, the word to be corrected and the word for correcting the word to be corrected are extracted from the word after boundary division.
Still taking the general word in word text "… a B …" as described above as an example. In the boundary segmentation, the word text may be divided into 5 parts, namely: the part before A, the part after A, "B, B; b is a word that may be incorrect by speech recognition, and a is a candidate word for correcting B. Taking "ask for love, set what words can be assembled" as an example, the part before A is "ask for questions", A is "love, B is" set ", and the part after B is" what words can be assembled ". In the extraction, a and B are extracted from the divided 5 parts.
In addition, based on the length characteristics of Chinese words, the maximum number of words to be corrected is two, and the maximum number of words to be corrected is 7.
In step S13, candidate words for correcting the word to be corrected are found from words corrected by the word to be corrected using a pronunciation confusion matrix including probabilities that each pronunciation is recognized as another pronunciation.
The pronunciation confusion matrix may be obtained as follows.
Firstly, obtaining probability distribution y E R of each sentence per frame by utilizing an acoustic model and a labeling text v Where v is the size of the pronunciation dictionary, the subscript of the highest probability in y is idx, the list of pronunciation dictionaries (e.g., "ai1", "ai2", etc., where the number indicates the tone) is token, and y indicates token [ idx ]]Probability distribution of the recognition result of this word. The acoustic model refers to a model which can return a pronunciation sequence corresponding to the audio aiming at a section of input audio characteristics, and the labeling text refers to a text corresponding to the audio.
Then, by means of the average value of y in a plurality of test sets, a probability matrix, namely a pronunciation confusion matrix, of each pronunciation being identified as other pronunciation can be obtained. By using the pronunciation confusion matrix in the text error correction process after voice recognition, the acquired pronunciation information contained in the pronunciation similarity can be more complete, and the voice recognition error correction effect is better.
For example, assuming that the pronunciation dictionary has 3 kinds of pronunciations, "ai1", "ai2" and "ai3", respectively, a pronunciation confusion matrix as shown in table 1 below is obtained. As can be seen from the pronunciation confusion matrix, the probability that "ai1" is read as "ai1" is 0.8, the probability that "ai1" is read as "ai2" is 0.15, and the probability that "ai1" is read as "ai3" is 0.05.
| ai1 | ai2 | ai3 | |
| ai1 | 0.8 | 0.15 | 0.05 |
| ai2 | 0.2 | 0.75 | 0.05 |
| ai3 | 0.15 | 0.05 | 0.8 |
TABLE 1
Step S13 may have a plurality of implementations, and one implementation may be: firstly, acquiring the similarity between the pinyin of each word in words which need to be corrected and the pinyin of the word which needs to be corrected by utilizing a pronunciation confusion matrix; then, if the similarity is greater than a preset threshold, it is determined that the word is a candidate word for correcting the word that needs to be corrected.
In one embodiment, where the word to be corrected is a word, the pronunciation confusion matrix may be used to obtain the similarity of the pinyin of each word in the word to be corrected to the pinyin of the word to be corrected.
For example, taking the example of how the text after speech recognition is "awake", after the word to be corrected is "awake" is extracted in step S12, the word to be corrected is "awake", and then the similarity between the pinyin of each word in "awake", that is, "sense" and "awake", and the pinyin of the word to be corrected "absolute" is obtained in step S13 using the pronunciation confusion matrix.
In yet another embodiment, in the case where the word to be corrected is two words, the pronunciation confusion matrix may be used to obtain the similarity between the pinyin of two words sequentially selected from the first word among words corrected for the word to be corrected and the pinyin of two words of the word to be corrected.
For example, taking how the text after the voice recognition is "how yellow" of two yellow green willows is written as an example, in step S12, the word to be corrected is "yellow, after the word to be corrected is" yellow green willows ", in step S13, the pinyin similarity of" two "of two yellow green willows" and the word to be corrected "yellow" is first obtained by using the pronunciation confusion matrix, then the pinyin similarity of "yellow" of "two yellow green willows" and the word to be corrected "yellow" is obtained by using the pronunciation confusion matrix, then the pinyin similarity of "yellow" of "two yellow green willows" and the word to be corrected "yellow" is obtained by using the pronunciation confusion matrix, and then the pinyin similarity of "yellow green willows" of "two yellow green willows" and the word to be corrected "is obtained by using the pronunciation confusion matrix, and the pinyin similarity of" yellow green willows "of" two yellow green willows "is obtained by using the pronunciation confusion matrix.
In addition, in the case where the word to be corrected is one word or two words, there may be cases where the word to be corrected is a polyphone word, that is, there are a plurality of pronunciations. In the case of polyphones, it is necessary to obtain the similarity of the pinyin of each word in the word to be corrected to each pronunciation of the word to be corrected.
Taking the example of how the text after speech recognition is "endless" as an example, since "endless" has a plurality of pronunciations, for example, "jube" or "ji a o" can be read, when the similarity is obtained, it is necessary to obtain the similarity between each pronunciation of "endless" and the pinyin of each word in "endless".
Taking the text after the speech recognition as an example of how the Chinese character is written in the great drought meeting, because the Chinese character has a plurality of pronunciations, for example, D-a-h n can be read, D-a-h-n can also be read, and therefore, when the similarity is obtained, the similarity between each pronunciations of the Chinese character and the pinyin of every two continuous characters in the great drought meeting is required to be obtained.
In step S14, the word to be corrected is corrected with the candidate word.
That is, in this step, the candidate word having the greatest similarity is selected from among the candidate words, and the word to be corrected is corrected. Still taking the example of "two yellow-green-willows" as an example, assuming that the spelling similarity between the "yellow-green-willows" and the word "yellow calendar" to be corrected is the greatest obtained in step S13, the word "yellow calendar" to be corrected is corrected to "yellow-green-willows" in step S14. Thus, after post-processing, the text after speech recognition is corrected from "how the yellow calendar of two green willows" to "how the yellow balance of two green willows writes".
By adopting the technical scheme, firstly, whether the text after voice recognition belongs to the word text or not is checked based on the expression characteristics of the word text, then if the text after voice recognition belongs to the word text, the word needing error correction and the word used for correcting the word needing error correction are extracted from the text after voice recognition, then candidate words used for correcting the word needing error correction are searched from the word needing error correction by utilizing a pronunciation confusion matrix, finally the word needing error correction is corrected by utilizing the candidate words, so that the word needing error correction can be corrected by utilizing the word with correct voice recognition in the word text after voice recognition, thereby greatly improving the sentence accuracy of voice recognition, correcting the word in the recognition result and improving the user experience of voice recognition.
Fig. 2 is a schematic block diagram of an error correction apparatus for text after speech recognition according to one embodiment of the present disclosure. As shown in fig. 2, the apparatus includes: a checking module 21, configured to check whether the text after speech recognition belongs to the word text based on the expression characteristics of the word text; the extracting module 22 is configured to extract, from the text after speech recognition, a word to be corrected and a word for correcting the word to be corrected, where a relationship between the word to be corrected and the word for correcting the word to be corrected is a word-in-word relationship; a searching module 23, configured to search for a candidate word for correcting the word to be corrected from words that correct the word to be corrected, using a pronunciation confusion matrix, where the pronunciation confusion matrix includes a probability that each pronunciation is identified as another pronunciation; an error correction module 24 for correcting the word to be corrected with the candidate word.
By adopting the technical scheme, firstly, whether the text after voice recognition belongs to the word text or not is checked based on the expression characteristics of the word text, then if the text after voice recognition belongs to the word text, the word needing error correction and the word used for correcting the word needing error correction are extracted from the text after voice recognition, then candidate words used for correcting the word needing error correction are searched from the word needing error correction by utilizing a pronunciation confusion matrix, finally the word needing error correction is corrected by utilizing the candidate words, so that the word needing error correction can be corrected by utilizing the word with correct voice recognition in the word text after voice recognition, thereby greatly improving the sentence accuracy of voice recognition, correcting the word in the recognition result and improving the user experience of voice recognition.
Optionally, the inspection module 21 is further configured to: checking whether the text after voice recognition meets the expression characteristics of B of A, wherein A is a word and B is a word in A; if the text after voice recognition meets the expression characteristics of the B of the A, the text after voice recognition belongs to word-in-word text.
Optionally, the extraction module 22 includes: the boundary dividing sub-module is used for dividing the boundary of the words in the text after the voice recognition based on the expression characteristics of the word text in the words; and the extraction sub-module is used for extracting the words needing to be corrected from the words after boundary division and the words needing to be corrected based on the expression characteristics of the word text in the words.
Alternatively, the words to be corrected are at most two words, and the words to be corrected are at most 7 words.
Optionally, the searching module 23 includes: the acquisition sub-module is used for acquiring the similarity between the pinyin of each word in the word corrected by the word to be corrected and the pinyin of the word to be corrected by utilizing the pronunciation confusion matrix; and the determining submodule is used for determining that the word is a candidate word for correcting the word needing to be corrected if the similarity is larger than a preset threshold value.
Optionally, the obtaining submodule is configured to: in the case that the word to be corrected is one word, the pronunciation confusion matrix is used to obtain the similarity between the pinyin of each word in the word to be corrected and the pinyin of the word to be corrected.
Optionally, the obtaining submodule is configured to: when the character to be corrected is two characters, the pronunciation confusion matrix is used to obtain the similarity between the spelling of the two characters sequentially selected from the first characters in the corrected words and the spelling of the two characters.
Optionally, the obtaining sub-module is further configured to: and if the character to be corrected has a plurality of pronunciations, acquiring the similarity between the pinyin of each character in the word corrected by the character to be corrected and each pronunciation of the character to be corrected.
Optionally, the error correction module 24 is further configured to: and selecting the candidate word with the largest similarity from the candidate words to correct the word to be corrected.
Referring now to fig. 3, a schematic diagram of an electronic device 600 suitable for use in implementing embodiments of the present disclosure is shown. The terminal devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and the like, and stationary terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 3 is merely an example and should not be construed to limit the functionality and scope of use of the disclosed embodiments.
As shown in fig. 3, the electronic device 600 may include a processing means (e.g., a central processing unit, a graphics processor, etc.) 601, which may perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 602 or a program loaded from a storage means 608 into a Random Access Memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic apparatus 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other through a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.
In general, the following devices may be connected to the I/O interface 605: input devices 606 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, and the like; an output device 607 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 608 including, for example, magnetic tape, hard disk, etc.; and a communication device 609. The communication means 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. While fig. 3 shows an electronic device 600 having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead.
In particular, according to embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a non-transitory computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via communication means 609, or from storage means 608, or from ROM 602. The above-described functions defined in the methods of the embodiments of the present disclosure are performed when the computer program is executed by the processing device 601.
It should be noted that the computer readable medium described in the present disclosure may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this disclosure, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present disclosure, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.
In some implementations, the clients, servers may communicate using any currently known or future developed network protocol, such as HTTP (HyperText Transfer Protocol ), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the internet (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed networks.
The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device.
The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: based on the expression characteristics of the word text, checking whether the text after the voice recognition belongs to the word text; extracting a word to be corrected and a word for correcting the word to be corrected from the text after the voice recognition if the text after the voice recognition belongs to the word-in-word text, wherein the word to be corrected and the word for correcting the word to be corrected are in word-in-word relation; searching candidate words for correcting the word to be corrected from the words to be corrected by utilizing a pronunciation confusion matrix, wherein the pronunciation confusion matrix comprises the probability that each pronunciation is identified as other pronunciation; and correcting the word to be corrected by using the candidate word.
Computer program code for carrying out operations of the present disclosure may be written in one or more programming languages, including, but not limited to, an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).
The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The modules described in the embodiments of the present disclosure may be implemented in software or hardware. The name of a module does not in some cases define the module itself.
The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a Complex Programmable Logic Device (CPLD), and the like.
In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
In accordance with one or more embodiments of the present disclosure, example 1 provides a method of error correction of text after speech recognition, comprising: based on the expression characteristics of the word text, checking whether the text after the voice recognition belongs to the word text; extracting a word to be corrected and a word for correcting the word to be corrected from the text after the voice recognition if the text after the voice recognition belongs to the word-in-word text, wherein the word to be corrected and the word for correcting the word to be corrected are in word-in-word relation; searching candidate words for correcting the word to be corrected from the words to be corrected by utilizing a pronunciation confusion matrix, wherein the pronunciation confusion matrix comprises the probability that each pronunciation is identified as other pronunciation; and correcting the word to be corrected by using the candidate word.
According to one or more embodiments of the present disclosure, example 2 provides the method of example 1, wherein the checking whether the speech-recognized text belongs to the word-in-word text based on the expression characteristics of the word-in-word text includes: checking whether the text after voice recognition meets the expression characteristics of B of A, wherein A is a word and B is a word in A; if the text after voice recognition meets the expression characteristics of the B of the A, the text after voice recognition belongs to word-in-word text.
According to one or more embodiments of the present disclosure, example 3 provides the method of example 1, wherein the extracting the word to be corrected from the text after speech recognition and the word for correcting the word to be corrected include: based on the expression characteristics of the word text in the words, carrying out boundary division on the words in the text after the voice recognition; and extracting the word to be corrected and the word used for correcting the word to be corrected from the word after boundary division based on the expression characteristics of the word text in the word.
In accordance with one or more embodiments of the present disclosure, example 4 provides the method of example 1, wherein the searching for a candidate word for correcting the word to be corrected from the words correcting the word to be corrected using a pronunciation confusion matrix, comprises: acquiring the similarity between the pinyin of each word in the word corrected by the word to be corrected and the pinyin of the word to be corrected by using the pronunciation confusion matrix; and if the similarity is larger than a preset threshold value, determining that the word is a candidate word for correcting the word needing to be corrected.
In accordance with one or more embodiments of the present disclosure, example 5 provides the method of example 4, wherein the obtaining, using the pronunciation confusion matrix, a similarity of pinyin for each of the words that correct the word to be corrected and pinyin for the word to be corrected includes: and under the condition that the word to be corrected is one word, acquiring the similarity of the pinyin of each word in the word to be corrected and the pinyin of the word to be corrected by utilizing the pronunciation confusion matrix.
In accordance with one or more embodiments of the present disclosure, example 6 provides the method of example 4, wherein the obtaining, using the pronunciation confusion matrix, a similarity of pinyin for each of the words that correct the word to be corrected and pinyin for the word to be corrected includes: and under the condition that the characters needing to be corrected are two characters, acquiring the similarity between the pinyin of the two characters and the pinyin of the two characters needing to be corrected, which are sequentially selected from the first characters in the words correcting the characters needing to be corrected, by utilizing the pronunciation confusion matrix.
In accordance with one or more embodiments of the present disclosure, example 7 provides the method of example 5 or 6, wherein if the word to be corrected has a plurality of pronunciations, a similarity of the pinyin of each of the words corrected for the word to be corrected and each of the pronunciations of the word to be corrected is obtained.
According to one or more embodiments of the present disclosure, example 8 provides the method of example 1, wherein said correcting said word to be corrected with said candidate word comprises: and selecting the candidate word with the largest similarity from the candidate words, and correcting the word to be corrected.
The foregoing description is only of the preferred embodiments of the present disclosure and description of the principles of the technology being employed. It will be appreciated by persons skilled in the art that the scope of the disclosure referred to in this disclosure is not limited to the specific combinations of features described above, but also covers other embodiments which may be formed by any combination of features described above or equivalents thereof without departing from the spirit of the disclosure. Such as those described above, are mutually substituted with the technical features having similar functions disclosed in the present disclosure (but not limited thereto).
Moreover, although operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order. In certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the present disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.
Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are example forms of implementing the claims. The specific manner in which the various modules perform the operations in the apparatus of the above embodiments have been described in detail in connection with the embodiments of the method, and will not be described in detail herein.
Claims (11)
1. A method for error correction of text after speech recognition, comprising:
based on the expression characteristics of the word text, checking whether the text after the voice recognition belongs to the word text;
extracting a word to be corrected and a word for correcting the word to be corrected from the text after the voice recognition if the text after the voice recognition belongs to the word-in-word text, wherein the word to be corrected and the word for correcting the word to be corrected are in word-in-word relation;
searching candidate words for correcting the word to be corrected from the words to be corrected by utilizing a pronunciation confusion matrix, wherein the pronunciation confusion matrix comprises the probability that each pronunciation is identified as other pronunciation;
and correcting the word to be corrected by using the candidate word.
2. The method of claim 1, wherein the checking whether the text after the speech recognition belongs to the word-in-word text based on the expression characteristics of the word-in-word text comprises:
checking whether the text after voice recognition meets the expression characteristics of B of A, wherein A is a word and B is a word in A;
if the text after voice recognition meets the expression characteristics of the B of the A, the text after voice recognition belongs to word-in-word text.
3. The method of claim 1, wherein the extracting the words to be corrected and the words for correcting the words to be corrected from the speech-recognized text comprises:
based on the expression characteristics of the word text in the words, carrying out boundary division on the words in the text after the voice recognition;
and extracting the word to be corrected and the word used for correcting the word to be corrected from the word after boundary division based on the expression characteristics of the word text in the word.
4. The method of claim 1, wherein the searching for a candidate word for correcting the word to be corrected from the word for correcting the word to be corrected using a pronunciation confusion matrix comprises:
acquiring the similarity between the pinyin of each word in the word corrected by the word to be corrected and the pinyin of the word to be corrected by using the pronunciation confusion matrix;
and if the similarity is larger than a preset threshold value, determining that the word is a candidate word for correcting the word needing to be corrected.
5. The method of claim 4, wherein the obtaining, using the pronunciation confusion matrix, a similarity of the pinyin for each of the words that correct the word to be corrected to the pinyin for the word to be corrected includes:
and under the condition that the word to be corrected is one word, acquiring the similarity of the pinyin of each word in the word to be corrected and the pinyin of the word to be corrected by utilizing the pronunciation confusion matrix.
6. The method of claim 4, wherein the obtaining, using the pronunciation confusion matrix, a similarity of the pinyin for each of the words that correct the word to be corrected to the pinyin for the word to be corrected includes:
and under the condition that the characters needing to be corrected are two characters, acquiring the similarity between the pinyin of the two characters and the pinyin of the two characters needing to be corrected, which are sequentially selected from the first characters in the words correcting the characters needing to be corrected, by utilizing the pronunciation confusion matrix.
7. The method of claim 5 or 6, wherein if the word to be corrected has a plurality of pronunciations, a similarity of the pinyin of each of the words corrected for the word to be corrected to each of the pronunciations of the word to be corrected is obtained.
8. The method of claim 1, wherein said correcting the word to be corrected using the candidate word comprises:
and selecting the candidate word with the largest similarity from the candidate words, and correcting the word to be corrected.
9. An error correction device for text after speech recognition, comprising:
the checking module is used for checking whether the text after the voice recognition belongs to the word text or not based on the expression characteristics of the word text;
the extraction module is used for extracting a word to be corrected and a word for correcting the word to be corrected from the text after the voice recognition if the text after the voice recognition belongs to the word text, wherein the relationship between the word to be corrected and the word for correcting the word to be corrected is a word-in-word relationship;
a searching module, configured to search for a candidate word for correcting the word to be corrected from the word to be corrected by using a pronunciation confusion matrix, where the pronunciation confusion matrix includes a probability that each pronunciation is identified as another pronunciation;
and the error correction module is used for correcting the word needing to be corrected by utilizing the candidate word.
10. A computer readable medium on which a computer program is stored, characterized in that the program, when being executed by a processing device, carries out the steps of the method according to any one of claims 1-8.
11. An electronic device, comprising:
a storage device having a computer program stored thereon;
processing means for executing said computer program in said storage means to carry out the steps of the method according to any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011311695.7A CN112509581B (en) | 2020-11-20 | 2020-11-20 | Error correction method and device for text after voice recognition, readable medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011311695.7A CN112509581B (en) | 2020-11-20 | 2020-11-20 | Error correction method and device for text after voice recognition, readable medium and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112509581A CN112509581A (en) | 2021-03-16 |
| CN112509581B true CN112509581B (en) | 2024-03-01 |
Family
ID=74959146
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011311695.7A Active CN112509581B (en) | 2020-11-20 | 2020-11-20 | Error correction method and device for text after voice recognition, readable medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112509581B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113192497B (en) * | 2021-04-28 | 2024-03-01 | 平安科技（深圳）有限公司 | Speech recognition method, device, equipment and medium based on natural language processing |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6064959A (en) * | 1997-03-28 | 2000-05-16 | Dragon Systems, Inc. | Error correction in speech recognition |
| US6513005B1 (en) * | 1999-07-27 | 2003-01-28 | International Business Machines Corporation | Method for correcting error characters in results of speech recognition and speech recognition system using the same |
| JP2004258531A (en) * | 2003-02-27 | 2004-09-16 | Nippon Telegr & Teleph Corp <Ntt> | Voice recognition error correction method, system, and program |
| WO2014117549A1 (en) * | 2013-01-29 | 2014-08-07 | Tencent Technology (Shenzhen) Company Limited | Method and device for error correction model training and text error correction |
| CN106598939A (en) * | 2016-10-21 | 2017-04-26 | 北京三快在线科技有限公司 | Method and device for text error correction, server and storage medium |
| CN109243433A (en) * | 2018-11-06 | 2019-01-18 | 北京百度网讯科技有限公司 | Audio recognition method and device |
| CN110765763A (en) * | 2019-09-24 | 2020-02-07 | 金蝶软件（中国）有限公司 | Error correction method and device for speech recognition text, computer equipment and storage medium |
| CN111243593A (en) * | 2018-11-09 | 2020-06-05 | 奇酷互联网络科技（深圳）有限公司 | Speech recognition error correction method, mobile terminal and computer-readable storage medium |
| CN111369996A (en) * | 2020-02-24 | 2020-07-03 | 网经科技（苏州）有限公司 | Method for correcting text error in speech recognition in specific field |
| WO2020186778A1 (en) * | 2019-03-15 | 2020-09-24 | 平安科技（深圳）有限公司 | Error word correction method and device, computer device, and storage medium |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090326938A1 (en) * | 2008-05-28 | 2009-12-31 | Nokia Corporation | Multiword text correction |
| TWI391832B (en) * | 2008-09-09 | 2013-04-01 | Inst Information Industry | Error detection apparatus and methods for chinese articles, and storage media |
| CN103714048B (en) * | 2012-09-29 | 2017-07-21 | 国际商业机器公司 | Method and system for correcting text |
| US9747897B2 (en) * | 2013-12-17 | 2017-08-29 | Google Inc. | Identifying substitute pronunciations |
-
2020
- 2020-11-20 CN CN202011311695.7A patent/CN112509581B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6064959A (en) * | 1997-03-28 | 2000-05-16 | Dragon Systems, Inc. | Error correction in speech recognition |
| US6513005B1 (en) * | 1999-07-27 | 2003-01-28 | International Business Machines Corporation | Method for correcting error characters in results of speech recognition and speech recognition system using the same |
| JP2004258531A (en) * | 2003-02-27 | 2004-09-16 | Nippon Telegr & Teleph Corp <Ntt> | Voice recognition error correction method, system, and program |
| WO2014117549A1 (en) * | 2013-01-29 | 2014-08-07 | Tencent Technology (Shenzhen) Company Limited | Method and device for error correction model training and text error correction |
| CN106598939A (en) * | 2016-10-21 | 2017-04-26 | 北京三快在线科技有限公司 | Method and device for text error correction, server and storage medium |
| CN109243433A (en) * | 2018-11-06 | 2019-01-18 | 北京百度网讯科技有限公司 | Audio recognition method and device |
| CN111243593A (en) * | 2018-11-09 | 2020-06-05 | 奇酷互联网络科技（深圳）有限公司 | Speech recognition error correction method, mobile terminal and computer-readable storage medium |
| WO2020186778A1 (en) * | 2019-03-15 | 2020-09-24 | 平安科技（深圳）有限公司 | Error word correction method and device, computer device, and storage medium |
| CN110765763A (en) * | 2019-09-24 | 2020-02-07 | 金蝶软件（中国）有限公司 | Error correction method and device for speech recognition text, computer equipment and storage medium |
| CN111369996A (en) * | 2020-02-24 | 2020-07-03 | 网经科技（苏州）有限公司 | Method for correcting text error in speech recognition in specific field |
Non-Patent Citations (5)
| Title |
|---|
| Post Text Processing of Chinese Speech Recognition Based on Bidirectional LSTM Networks and CRF;Li Yang;electronics;20191130;第8卷(第11期);全文 * |
| 一种基于语义分析的汉语语音识别纠错方法;韦向峰;张全;熊亮;;计算机科学;20061025(10);全文 * |
| 基于word2vec的语音识别后文本纠错;张佳宁;严冬梅;王勇;计算机工程与设计;20201116;第41卷(第11期);全文 * |
| 汉语语音理解中自动纠错系统的研究;李晶皎, 张王利, 姚天顺;软件学报;19990405(04);全文 * |
| 语音识别技术研究进展与挑战;刘庆峰;高建清;万根顺;;数据与计算发展前沿;20191215(06);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112509581A (en) | 2021-03-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112115706B (en) | Text processing method and device, electronic equipment and medium | |
| CN110969012B (en) | Text error correction method and device, storage medium and electronic equipment | |
| CN113470619B (en) | Speech recognition method, device, medium and equipment | |
| CN113378586B (en) | Speech translation method, translation model training method, device, medium, and apparatus | |
| CN111489735B (en) | Voice recognition model training method and device | |
| CN112906380B (en) | Character recognition method and device in text, readable medium and electronic equipment | |
| CN111986655B (en) | Audio content identification method, device, equipment and computer readable medium | |
| CN111667810B (en) | Method and device for acquiring polyphone corpus, readable medium and electronic equipment | |
| CN112883968A (en) | Image character recognition method, device, medium and electronic equipment | |
| CN114445754A (en) | Video processing method and device, readable medium and electronic equipment | |
| CN112380876B (en) | Translation method, device, equipment and medium based on multilingual machine translation model | |
| CN109558600B (en) | Translation processing method and device | |
| CN111815274A (en) | Information processing method and device and electronic equipment | |
| CN112509581B (en) | Error correction method and device for text after voice recognition, readable medium and electronic equipment | |
| CN115618808A (en) | Document typesetting method and device, electronic equipment and storage medium | |
| CN112489652B (en) | Text acquisition method and device for voice information and storage medium | |
| US20210201895A1 (en) | Method and apparatus for outputting analysis abnormality information in spoken language understanding | |
| CN110659208A (en) | Test data set updating method and device | |
| WO2023138361A1 (en) | Image processing method and apparatus, and readable storage medium and electronic device | |
| CN116956933A (en) | Event extraction method and device, electronic equipment and computer readable storage medium | |
| CN110502630B (en) | Information processing method and device | |
| CN110728137B (en) | Method and device for word segmentation | |
| CN116821327A (en) | Text data processing method, apparatus, device, readable storage medium and product | |
| CN113221514A (en) | Text processing method and device, electronic equipment and storage medium | |
| CN112699687A (en) | Content cataloging method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |