Speech error correction method, device, apparatus and storage medium

Abstract

The application belongs to the technical field of intelligent interaction, and discloses a voice correction method, device, equipment and storage medium. When it is detected that the voice information has errors, the application determines the to-be-corrected character in the voice information, acquires the to-be-corrected syllable corresponding to the to-be-corrected character, determines the to-be-replaced phoneme in the to-be-corrected syllable and the target phoneme corresponding to the to-be-replaced phoneme according to a preset replacement strategy, replaces the to-be-replaced phoneme with the target phoneme, acquires the target syllable, and corrects the to-be-corrected character according to the target syllable to acquire the corrected voice information. Since the application determines the to-be-corrected syllable in the voice information, determines the to-be-replaced phoneme in the to-be-corrected syllable and the target phoneme corresponding to the to-be-replaced phoneme according to a preset replacement strategy, and then replaces the to-be-replaced phoneme, the to-be-corrected syllable of the error character in the voice information is corrected, so that the error of the voice information is quickly corrected, and the voice correction efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent interactive technology, and in particular to a voice error correction method, apparatus, device, and storage medium. Background Technology

[0002] With the development of science and technology and the progress of society, intelligent interaction is becoming increasingly popular. In this process, Automatic Speech Recognition (ASR) technology, which is closely related to intelligent interaction, is also gradually entering people's work and lives. ASR technology is often used in intelligent interaction scenarios; however, it has a certain probability of error recognition. In Chinese application scenarios, the most common speech recognition error is Pinyin error. Currently, the method for correcting Chinese Pinyin is to directly identify the Pinyin components of the speech information, then calculate the correction edit distance, and directly correct the speech information based on the correction edit distance. However, the existing Pinyin correction method has a relatively long correction edit distance, resulting in low correction efficiency and long correction time.

[0003] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main objective of this invention is to provide a speech error correction method, apparatus, device, and storage medium, aiming to solve the technical problems of low speech error correction efficiency and long correction time in the prior art.

[0005] To achieve the above objectives, the present invention provides a speech error correction method, the method comprising the following steps: When an error is detected in the voice information, the characters in the voice information that need to be corrected are identified. Obtain the syllable to be corrected corresponding to the character to be corrected; The phonemes to be replaced in the syllable to be corrected and the target phonemes corresponding to the phonemes to be replaced are determined according to a preset replacement strategy. The phoneme to be replaced is replaced with the target phoneme to obtain the target syllable, and the character to be corrected is corrected according to the target syllable to obtain the corrected speech information.

[0006] Optionally, determining the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced according to a preset replacement strategy includes: Phoneme matching is performed on each phoneme in the syllable to be corrected according to a preset replacement strategy; Based on the phoneme matching results, determine the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced.

[0007] Optionally, determining the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced based on the phoneme matching result includes: When phoneme matching is successful, the phoneme to be replaced in the syllable to be corrected is determined based on the phoneme matching result; The target phoneme corresponding to the phoneme to be replaced is determined according to a preset replacement mapping table.

[0008] Optionally, when phoneme matching is successful, determining the phonemes to be replaced in the syllable to be corrected based on the phoneme matching result includes: When phoneme matching is successful, the phoneme composition type corresponding to the syllable to be corrected is determined based on the matching result; The phonemes to be replaced in the syllable to be corrected are determined based on the phoneme composition type.

[0009] Optionally, after performing phoneme matching on each phoneme in the syllable to be corrected according to a preset replacement strategy, the method further includes: When phoneme matching fails, determine the phoneme composition type corresponding to the syllable to be corrected; Determine whether there are any phonemes that need to be deleted in the syllable to be corrected based on the phoneme composition type; If it exists, delete the phoneme that needs to be deleted to obtain the modified syllable.

[0010] Optionally, if the phoneme to be deleted exists, and the modified syllable is obtained, the method includes: The phonemes to be replaced in the modified syllable and the target phonemes corresponding to the phonemes to be replaced are determined according to the phoneme composition type.

[0011] Optionally, determining the phoneme to be replaced in the modified syllable and the target phoneme corresponding to the phoneme to be replaced based on the phoneme composition type includes: The phonemes to be replaced in the modified syllable are determined according to the phoneme composition type; The target phoneme corresponding to the phoneme to be replaced is determined according to a preset replacement mapping table.

[0012] Optionally, after determining whether there are any phonemes that need to be deleted in the syllable to be corrected based on the phoneme composition type, the method further includes: If it does not exist, then determine the phonemes to be replaced in the syllable to be corrected based on the phoneme composition type. The target phoneme corresponding to the phoneme to be replaced is determined according to the preset replacement mapping table.

[0013] Optionally, the step of replacing the phoneme to be replaced with the target phoneme to obtain the target syllable, and correcting the character to be corrected based on the target syllable to obtain the corrected speech information includes: The phoneme to be replaced is replaced with the target phoneme to obtain the target syllable; The target syllable is corrected according to a preset error correction strategy, and the corrected syllable is converted into the corresponding target character. Replace the character to be corrected with the target character to obtain the corrected speech information.

[0014] Optionally, the step of correcting the target syllable according to a preset error correction strategy and converting the corrected syllable into the corresponding target character includes: The error correction editing distance of the target syllable is determined according to a preset error correction strategy; The target syllable is corrected based on the error correction editing distance; Convert the corrected syllables into their corresponding target characters.

[0015] Optionally, after determining the error correction editing distance of the target syllable according to a preset error correction strategy, the method includes: Obtain a standard error correction strategy, and determine the standard edit distance of the target syllable based on the standard error correction strategy; When the error correction editing distance does not exceed the standard editing distance, the step of correcting the target syllable based on the error correction editing distance is performed.

[0016] Optionally, after obtaining the standard error correction strategy and determining the standard edit distance of the target syllable based on the standard error correction strategy, the method further includes: When the error correction editing distance exceeds the standard editing distance, the target syllable is corrected according to the standard editing distance; Perform the step of converting the corrected syllables into the corresponding target characters.

[0017] Furthermore, to achieve the above objectives, the present invention also proposes a speech correction device, the speech correction device comprising: The syllable acquisition module is used to acquire the syllable to be corrected corresponding to the character to be corrected. The replacement strategy module is used to determine the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced according to a preset replacement strategy. The replacement and correction module is used to replace the phoneme to be replaced with the target phoneme to obtain the target syllable, and to correct the character to be corrected based on the target syllable to obtain the corrected speech information.

[0018] Furthermore, the replacement and error correction module is also used to perform phoneme matching on each phoneme in the syllable to be corrected according to a preset replacement strategy, and determine the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced according to the phoneme matching result.

[0019] Furthermore, the replacement and error correction module is also used to determine the phoneme to be replaced in the syllable to be corrected based on the phoneme matching result when the phoneme matching is successful, and to determine the target phoneme corresponding to the phoneme to be replaced based on the preset replacement mapping table.

[0020] Furthermore, the replacement and error correction module is also used to determine the phoneme composition type corresponding to the syllable to be corrected based on the matching result when the phoneme matching is successful, and to determine the phoneme to be replaced in the syllable to be corrected based on the phoneme composition type.

[0021] Furthermore, the replacement and error correction module is also used to replace the phoneme to be replaced with the target phoneme to obtain the target syllable, correct the target syllable according to a preset error correction strategy, convert the corrected syllable into the corresponding target character, replace the character to be corrected with the target character, and obtain the corrected speech information.

[0022] Furthermore, the replacement and error correction module is also used to determine the error correction editing distance of the target syllable according to a preset error correction strategy, correct the target syllable according to the error correction editing distance, and convert the corrected syllable into the corresponding target character.

[0023] Furthermore, to achieve the above objectives, the present invention also proposes a voice error correction device, which includes: a memory, a processor, and a voice error correction program stored in the memory and executable on the processor, wherein the voice error correction program is configured to implement the steps of the voice error correction method described above.

[0024] In addition, to achieve the above objectives, the present invention also proposes a storage medium storing a speech correction program, which, when executed by a processor, implements the steps of the speech correction method as described above.

[0025] This invention, upon detecting an error in speech information, identifies the character to be corrected, obtains the corresponding syllable to be corrected, determines the phoneme to be replaced within the syllable and its corresponding target phoneme according to a preset replacement strategy, replaces the phoneme to be replaced with the target phoneme to obtain the target syllable, and corrects the character to be corrected based on the target syllable to obtain corrected speech information. Because this invention corrects erroneous characters in speech information by identifying the syllable to be corrected, determining the phoneme to be replaced within the syllable according to a preset replacement strategy, and then replacing the phoneme, it achieves rapid correction of errors in speech information and improves speech error correction efficiency. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the structure of the voice error correction device in the hardware operating environment involved in the embodiments of the present invention; Figure 2 This is a flowchart illustrating the first embodiment of the speech error correction method of the present invention; Figure 3 This is a schematic diagram of the replacement process of the preset replacement strategy in the first embodiment of the speech error correction method of the present invention; Figure 4 This is a flowchart illustrating the second embodiment of the speech error correction method of the present invention; Figure 5 This is a flowchart illustrating the third embodiment of the speech error correction method of the present invention; Figure 6 This is a structural block diagram of the first embodiment of the speech correction device of the present invention.

[0027] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0028] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0029] Reference Figure 1 , Figure 1 This is a schematic diagram of the voice error correction device structure in the hardware operating environment involved in the embodiments of the present invention.

[0030] like Figure 1As shown, the voice error correction device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk storage device. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0031] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the speech correction device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0032] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and a voice error correction program.

[0033] exist Figure 1 In the voice correction device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the voice correction device of the present invention can be set in the voice correction device, and the voice correction device calls the voice correction program stored in the memory 1005 through the processor 1001 and executes the voice correction method provided in the embodiment of the present invention.

[0034] This invention provides a speech error correction method, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of a speech error correction method according to the present invention.

[0035] In this embodiment, the voice error correction method includes the following steps: Step S10: When an error is detected in the voice information, determine the characters in the voice information that need to be corrected.

[0036] It should be noted that the execution subject of the method in this embodiment may be a speech error correction device with data processing, network communication and program operation functions, such as a mobile phone, a tablet computer, a laptop computer, etc., or other devices or equipment that can implement the same or similar functions. The following description will be given by taking the speech error correction device as an example, which is not limited in this embodiment.

[0037] It should be understood that when the present embodiment is applied to a scenario where character pinyin error correction is required for speech information, the syllable to be corrected in the speech information is determined, the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced are determined according to a preset replacement strategy, and then error correction of the syllable to be corrected corresponding to the wrong character in the speech information is realized by replacing the phoneme to be replaced.

[0038] In Chinese application scenarios, pinyin errors are the most common, which requires error correction on speech information. If error correction is only performed according to the standard pinyin error correction scheme, that is, correcting wrong characters by calculating the edit distance for error correction, it will cause the problems of low error correction efficiency and long time consumption. For example, the pinyin of "王" (wang) in the speech information is recognized as "黄" (huang). The edit distance for error correction here is a quantitative measurement of the difference between two strings, which is measured by how many operations are required at minimum to change one string into another string.

[0039] The standard pinyin error correction scheme cannot well reflect the problem of similar pronunciation of Chinese characters. For example, the pronunciations of "华" and "挖" are very similar. When the above standard pinyin error correction scheme is used for error correction, the edit distance of the pinyin is 2 bits (converting "wa" to "hua"). Therefore, the above standard pinyin error correction scheme cannot accurately correct wrong pinyin, and it also has the problems of long error correction time and low error correction efficiency.

[0040] It can be understood that, compared with the problems of long error correction time and low error correction efficiency in the standard pinyin error correction scheme, this embodiment determines the syllable to be corrected in the speech information, determines the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced according to the preset replacement strategy, and then replaces the phoneme to be replaced to realize error correction on the syllable to be corrected of the wrong character in the speech information, so that errors in the speech information can be corrected quickly and the speech error correction efficiency is improved.

[0041] It should be noted that the speech information may be information received by the speech error correction device during speech interaction, and the speech information may be Chinese speech information or English speech information. In this embodiment, Chinese speech information is taken as an example for description.

[0042] The character to be corrected may be error character information when there is an error in the voice information. For example, when a user says "navigate to Chenghua Avenue", the voice information recognized by the voice error correction device is "navigate to Chengwa Avenue", therefore, the character to be corrected at this time is "wa".

[0043] It should be understood that when a voice error correction device receives voice information sent by a user, it converts the voice information into corresponding text information, then performs error determination on the text information according to the context in the voice information and the current conversation scenario. If there is error information in the text information, the error character in the text information is determined, that is, the character to be corrected; if there is no error information in the text information, it is determined that there is no character to be corrected in the text information.

[0044] In specific implementation, when a voice error correction device receives an input of voice information, it converts the voice information into text information, judges whether there is an error in the text information corresponding to the voice information according to the context in the voice information and the current conversation scenario. When it is detected that there is an error in the text information, it determines that there is an error in the voice information, then determines the error information in the voice information, and determines the character to be corrected according to the error information.

[0045] For example, when a voice error correction device receives an input of the voice information "play the music Hotel California", it converts the voice information into the text information "play the music Zou Hotel". It judges that there is an error in the text information corresponding to the voice information according to the context in the voice information and the current conversation scenario, determines that there is an error in the voice information, then determines the error information in the voice information, and determines that the character to be corrected is "Zou" according to the error information.

[0046] Step S20: Obtain the syllable to be corrected corresponding to the character to be corrected.

[0047] It should be noted that the syllable to be corrected may be the syllable or pinyin corresponding to the character to be corrected. For example, the character to be corrected is "huang", and the syllable to be corrected corresponding to the character to be corrected is "huang".

[0048] It should be understood that when the voice error correction device performs voice interaction with a user, the accuracy of the voice information sent by the user received by the voice error correction device mainly depends on the standardization degree of the user's mandarin. However, the user may have a special accent in mandarin due to local dialect. For example, when the user says "set an alarm clock", due to the user's special dialect accent, the voice information received by the voice error correction device is "set a lao clock", therefore, the character to be corrected in the voice information is "lao", and the corresponding syllable to be corrected is "lao".

[0049] Step S30: Determine the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced according to a preset replacement strategy.

[0050] It should be noted that the preset replacement strategy may be a pre-configured replacement strategy for replacing phonemes in the syllable to be corrected. The phoneme to be replaced may be the phoneme that needs to be edited and replaced in the syllable to be corrected. The target phoneme may be the phoneme used to replace the phoneme to be replaced in the syllable to be corrected.

[0051] It should be understood that the preset replacement strategy can determine the replacement strategy for the syllable to be replaced according to the phoneme composition type of the syllable to be replaced. For example, the speech error correction device determines whether the first phoneme of the syllable to be replaced is an initial consonant; if yes, it then determines whether the ending phoneme of the syllable to be replaced is a compound final, and if yes, replaces the compound final with the target phoneme.

[0052] It should be noted that the correct speech information is "hua" (corresponding to the Chinese character "hua"), but due to the influence of the user's dialect or environmental noise, the recognized speech information is "wa" (corresponding to the Chinese character "wa"). When a standard Pinyin error correction scheme corrects this error, it is required to correct "wa" to "hua". It can be seen from this error correction process that the standard Pinyin error correction scheme requires an edit distance of 2 bits. Therefore, the edit distance required by the standard Pinyin error correction scheme is too long, resulting in problems of low error correction efficiency and long error correction time.

[0053] In this embodiment, the speech error correction device determines that the syllable to be corrected is "wa", determines that the phoneme to be replaced in the syllable to be corrected is "w" according to the preset replacement strategy, determines that the target phoneme corresponding to the phoneme to be replaced is "h", replaces the phoneme to be replaced with the target phoneme to obtain the replaced syllable "hua", that is, the corrected result. This error correction process only requires an edit distance of 1 bit, thereby improving error correction efficiency and shortening error correction time.

[0054] Step S40: replacing the phoneme to be replaced with the target phoneme to obtain a target syllable, and correcting the character to be error-corrected according to the target syllable to obtain error-corrected speech information.

[0055] It should be noted that the preset replacement strategy refers to Figure 3 , Figure 3 is a strategy flow chart of the preset replacement strategy. The speech error correction device determines whether the first phoneme of the syllable to be corrected is an initial consonant; if yes, it determines the ending phoneme of the syllable to be corrected. If the ending phoneme is "iu", replacing "iu" with "iou"; if the ending phoneme is "ui", replacing "ui" with "uei"; if the first initial consonant of the syllable to be corrected is an initial consonant other than "j", "q" or "x", determining whether the syllable to be corrected ends with "un", and if it ends with "un", replacing "un" with "uen".

[0056] If the first phoneme of the syllable to be corrected is not an initial consonant, then determine the first phoneme of the syllable. If the first phoneme is "y", then determine whether the second phoneme is "i" or "u"; if the second phoneme is "i", then delete the first phoneme "y"; if the second phoneme is "u", then replace "u" with "v". If the first phoneme of the syllable to be corrected is "w", then determine whether the second phoneme of the syllable is "u"; if the second phoneme is "u", then delete the first phoneme "w"; if the second phoneme is not "u", then replace the first phoneme "w" with "u". If the first phoneme of the syllable to be corrected is "j", "q" or "x", then determine whether the second phoneme of the syllable to be corrected is "u", and if so, replace "u" with "v".

[0057] In a specific implementation, for example, the speech correction device determines that the syllable to be replaced in the speech information is "wa", determines the phoneme to be replaced as "w" and the target phoneme corresponding to the phoneme to be replaced as "u" according to the preset replacement strategy, replaces the phoneme to be replaced "w" with the target phoneme "u", obtains the target syllable "ua", then performs error correction editing on "ua" to obtain the corrected syllable "hua", and thus obtains the corrected speech information.

[0058] This embodiment, upon detecting an error in the speech information, identifies the character to be corrected, obtains the corresponding syllable, and, based on a preset replacement strategy, determines the phoneme to be replaced and its target phoneme within that syllable. The phoneme to be replaced is then replaced with the target phoneme to obtain the target syllable. Finally, the character to be corrected is corrected based on the target syllable to obtain the corrected speech information. Because this invention identifies the syllable to be corrected in the speech information, determines the phoneme to be replaced and its target phoneme based on a preset replacement strategy, and then replaces the phoneme with the target phoneme, it achieves error correction of the syllable containing the erroneous character in the speech information, thereby rapidly correcting errors in the speech information and improving speech correction efficiency.

[0059] refer to Figure 4 , Figure 4 This is a flowchart illustrating a second embodiment of a speech error correction method according to the present invention.

[0060] Based on the first embodiment described above, in this embodiment, step S30 includes: Step S301: Perform phoneme matching on each phoneme in the syllable to be corrected according to the preset replacement strategy.

[0061] It should be noted that phoneme matching can be a regular expression matching of each phoneme according to a preset replacement strategy.

[0062] It should be understood that in order to accurately obtain the phonemes to be replaced in the syllable to be corrected, the speech correction device needs to perform phoneme matching on each phoneme in the syllable to be corrected according to a preset replacement strategy. For example, the speech correction device first determines whether the first phoneme in the syllable to be corrected is an initial consonant according to the preset replacement strategy. If it is an initial consonant, it determines the last phoneme in the syllable to be corrected, and then matches the last phoneme according to the preset replacement strategy to determine the corresponding replacement scheme and the target phoneme to be replaced.

[0063] Furthermore, in order to accurately correct the syllable to be corrected after phoneme matching fails, the above step S301 further includes: Step S3011: When phoneme matching fails, determine the phoneme composition type corresponding to the syllable to be corrected; Step S3012: Determine whether there are any phonemes that need to be deleted in the syllable to be corrected based on the phoneme composition type; Step S3013: If it exists, delete the phoneme to be deleted to obtain the modified syllable.

[0064] It should be noted that phoneme matching failure may occur when there is no replaceable phoneme in the syllable to be corrected, but there may be a phoneme that needs to be deleted. For example, if the syllable to be corrected is "wu", phoneme matching is performed on "wu" according to the preset replacement strategy. It is determined that there is no replaceable phoneme in the syllable, but there is a phoneme that can be deleted. The phoneme that needs to be deleted is "w", and the corrected syllable is "u".

[0065] It should be understood that after a matching failure occurs in the syllable to be corrected, the speech correction device determines that there is no phoneme to be replaced in the syllable to be corrected. Then, it determines the phoneme composition type of the syllable to be corrected, and then determines whether there is a phoneme to be deleted in the syllable to be corrected based on the phoneme composition type. If there is a phoneme to be deleted, then the phoneme is deleted, and the modified syllable is obtained.

[0066] In a specific implementation, for example, if the syllable to be corrected is "ying", after the speech correction fails to match and replace the phoneme for the syllable to be corrected according to the preset replacement strategy, it is determined that there is no phoneme to be replaced in the syllable to be corrected. Then, the phoneme composition type corresponding to the syllable to be corrected is determined to be "y", "i", "n" and "g". Then, based on the phoneme composition type, it is determined that there is a phoneme to be deleted in the syllable to be corrected, which is "y". Finally, the phoneme "y" is deleted, and the modified syllable is "ing".

[0067] Furthermore, in order to continue replacing the modified syllables and ensure the accuracy of speech correction, the above step S3013 also includes: Step S30131: Determine the phonemes to be replaced in the modified syllable according to the phoneme composition type; Step S30132: Determine the target phoneme corresponding to the phoneme to be replaced according to the preset replacement mapping table.

[0068] In a specific implementation, for example, the speech correction device deletes the phonemes that need to be deleted from the syllable to be corrected, and obtains the modified syllable as "ing". Then, it determines that the composition type corresponding to the modified syllable is "i", "n" and "g". Then, it determines that the phoneme to be replaced in the modified syllable is "i", and determines the target phoneme corresponding to the phoneme to be replaced as "u" according to the preset replacement mapping table.

[0069] Furthermore, to avoid phoneme matching errors, the above step S3012 also includes: Step S30121: If it does not exist, determine the phonemes to be replaced in the syllable to be corrected according to the phoneme composition type; Step S30122: Determine the target phoneme corresponding to the phoneme to be replaced according to the preset replacement mapping table.

[0070] In a specific implementation, for example, if the phoneme to be corrected that failed to match is “yun”, then the composition type corresponding to the syllable is determined to be “y”, “u” and “n”, then the phoneme to be replaced in the syllable is determined to be “u”, and the target phoneme corresponding to the phoneme to be replaced is determined to be “v” according to the preset replacement mapping table.

[0071] Step S302: Determine the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced based on the phoneme matching result.

[0072] It should be noted that the matching result can be the result of regular expression matching of each phoneme in the syllable to be corrected according to a preset replacement strategy. The matching result may include information such as the replacement scheme for each phoneme in the syllable to be corrected, the phoneme to be replaced, and the target phoneme corresponding to the phoneme to be replaced, etc., which are not limited in this embodiment.

[0073] It should be understood that in order to accurately replace the phonemes in the syllable to be corrected, the speech correction device needs to determine the replacement scheme, the phoneme to be replaced, and the target phoneme to replace the phoneme to be replaced based on the matching results obtained after phoneme matching.

[0074] It should be noted that the matching results include the correction scheme for the syllable to be corrected. The correction scheme can be a scheme to replace the phoneme to be replaced in the syllable to be corrected with the target phoneme, or a scheme to delete the phoneme to be deleted in the syllable to be corrected. For example, if the syllable to be corrected is "ying", then according to the preset replacement strategy, the syllable to be corrected is matched, and the correction scheme is determined to be to delete the phoneme "y" in the syllable to be corrected, so as to obtain the target syllable "ing".

[0075] In a specific implementation, for example, if the syllable to be corrected is “yun”, the speech correction device performs phoneme matching on each phoneme in the syllable to be corrected according to a preset replacement strategy, determines that the phoneme to be replaced is “u”, the target phoneme is “v”, and the replacement scheme is to replace the phoneme to be replaced “u” with the target phoneme “v”.

[0076] Furthermore, in order to accurately obtain the target phoneme corresponding to the phoneme to be replaced, step S302 above also includes: Step S3021: When phoneme matching is successful, determine the phonemes to be replaced in the syllable to be corrected based on the phoneme matching result; Step S3022: Determine the target phoneme corresponding to the phoneme to be replaced according to the preset replacement mapping table.

[0077] It should be noted that the preset replacement mapping table can be a replacement mapping table established according to the preset replacement strategy.

[0078] It should be understood that, in order to accurately replace the phoneme to be replaced, when the phoneme matching is successful, the speech correction device determines the phoneme to be replaced in the syllable to be corrected and the corresponding replacement scheme based on the phoneme matching result, and determines the target phoneme corresponding to the phoneme to be replaced based on the replacement scheme and the preset replacement mapping table.

[0079] In a specific implementation, for example, if the syllable to be corrected is “qun”, when the phoneme matching is successful, the speech correction device determines the phoneme to be replaced in the syllable to be corrected as “u” and the corresponding replacement scheme based on the phoneme matching result. Based on the replacement scheme and the preset replacement mapping table, the target phoneme corresponding to the phoneme to be replaced “u” is determined to be “v”.

[0080] Furthermore, in order to accurately determine the phonemes to be replaced in the syllable to be corrected, step S3021 above includes: Step S30211: When phoneme matching is successful, determine the phoneme composition type corresponding to the syllable to be corrected based on the matching result; Step S30212: Determine the phonemes to be replaced in the syllable to be corrected according to the phoneme composition type.

[0081] It should be noted that the phoneme composition type can be the composition of each phoneme in the syllable to be corrected, the order of the phonemes, and the number of phonemes.

[0082] It should be understood that in order to accurately determine the phonemes to be replaced in the syllable to be corrected, the speech correction device needs to determine the phoneme composition type of the syllable to be corrected, and determine the composition, phoneme arrangement order and number of each phoneme in the syllable to be corrected based on the phoneme composition type.

[0083] In a specific implementation, for example, if the syllable to be corrected is “huang”, the speech correction device needs to determine the phoneme composition type of the syllable to be corrected. Based on the phoneme composition type, the number of phonemes in the syllable to be corrected is determined to be 5, and the composition of each phoneme is “h”, “u”, “a”, “n” and “g”. The phoneme arrangement order of the syllable to be corrected is “huang”.

[0084] This embodiment performs phoneme matching on each phoneme in the syllable to be corrected according to a preset replacement strategy, and determines the phoneme to be replaced and the target phoneme corresponding to the phoneme in the syllable to be corrected based on the phoneme matching results. Because this invention performs phoneme matching on the syllable to be corrected according to a preset replacement strategy, and then determines the phoneme replacement scheme based on the matching results, it achieves accurate replacement of the syllable to be corrected, thus improving the efficiency of speech correction.

[0085] refer to Figure 5 , Figure 5 This is a flowchart illustrating a third embodiment of a speech error correction method according to the present invention.

[0086] Based on the first embodiment described above, in this embodiment, step S40 includes: Step S401: Replace the phoneme to be replaced with the target phoneme to obtain the target syllable.

[0087] In a specific implementation, for example, if the syllable to be corrected is "wa", the phoneme to be replaced in the syllable to be corrected is "w", and the target phoneme corresponding to the phoneme to be replaced is "u", the speech correction device replaces the phoneme to be replaced "w" with the target phoneme "u" to obtain the target syllable "ua".

[0088] Step S402: Correct the target syllable according to the preset error correction strategy, and convert the corrected syllable into the corresponding target character.

[0089] It should be noted that the preset error correction strategy can be a pre-set strategy for correcting erroneous characters in speech information. This preset error correction strategy can be an error correction network model, which is pre-trained and generates the preset error correction strategy based on a large amount of speech information, semantic information, dialogue environment, and error correction history. The target character can be the correct character obtained after error correction.

[0090] It should be understood that the speech error correction device determines the correct character corresponding to the wrong character in the speech information according to a preset error correction strategy, then determines the correct syllable corresponding to the target syllable based on the correct character, performs error correction editing on the target syllable according to the correct syllable, and converts the error-corrected syllable into the corresponding target character.

[0091] In specific implementation, for example, the speech error correction device determines that the syllable to be corrected is "wa", the phoneme to be replaced is "w", and the target phoneme is "u"; it replaces the to-be-replaced phoneme "w" with the target phoneme "u" to obtain the target syllable "ua", determines the correct syllable "hua" according to the preset error correction strategy, then corrects the target syllable "ua" to obtain the error-corrected syllable "hua", and then converts the error-corrected syllable into the corresponding target character "华 (hua)".

[0092] Further, in order to accurately perform error correction editing on the target syllable, the above step S402 includes: Step S4021: determining the error correction edit distance of the target syllable according to the preset error correction strategy; Step S4022: correcting the target syllable according to the error correction edit distance; Step S4023: converting the error-corrected syllable into the corresponding target character.

[0093] It should be noted that the error correction edit distance may be a measurement distance for the degree of difference of a syllable string when correcting a syllable, and the measurement method may be the minimum number of operations required to convert the syllable to be corrected into the correct syllable. For example, when the syllable "ua" is corrected and edited to the syllable "hua", at least one edit is required, so the error correction edit distance of the syllable "ua" is 1.

[0094] In specific implementation, for example, the speech error correction device determines according to the preset error correction strategy that the correct syllable corresponding to the target syllable "uang" is "huang", thus determines that the error correction edit distance of the target syllable is 1, corrects the target syllable according to the error correction edit distance, adds the phoneme "h" to the beginning of the syllable to obtain the error-corrected syllable "huang", and then converts the syllable into the corresponding target character "黄 (huang)".

[0095] Further, in order to ensure that the target syllable is corrected and edited with the shortest-distance error correction method, after the above step S4021, the method further includes: Step S40211: acquiring a standard error correction strategy, and determining the standard edit distance of the target syllable according to the standard error correction strategy.

[0096] Step S40212: When the error correction editing distance does not exceed the standard editing distance, perform the step of correcting the target syllable according to the error correction editing distance.

[0097] It should be noted that standard error correction strategies can be strategies that directly correct erroneous syllables in speech information.

[0098] It should be understood that, in order to ensure that erroneous syllables are corrected with the shortest editing distance, speech correction devices need to compare the correction editing distance with the standard editing distance, and then select the correction method with the shorter editing distance.

[0099] In a specific implementation, for example, if the incorrect syllable in the speech information is "wang" and the correct syllable is "huang", the speech correction device edits the incorrect syllable according to the standard error correction strategy, which requires deleting the phoneme "w" and then adding the phonemes "u" and "h". Therefore, the standard edit distance is 2. However, the error correction device corrects the incorrect syllable according to the preset error correction strategy with an error correction edit distance of 1. Therefore, the error correction edit distance is less than the standard edit distance. Thus, the target syllable is corrected according to the error correction edit distance.

[0100] Furthermore, in order to select the error correction method with the shortest editing distance to correct the target syllable, the above step S40211 also includes: When the error correction editing distance exceeds the standard editing distance, the target syllable is corrected according to the standard editing distance; Perform the step of converting the corrected syllables into the corresponding target characters.

[0101] In specific implementations, for example, the error correction edit distance is 2 and the standard edit distance is 1. The speech error correction device selects the standard edit distance to correct the target syllable.

[0102] Step S403: Replace the character to be corrected with the target character to obtain the corrected speech information.

[0103] It should be noted that the corrected audio information can be the correct audio information obtained by replacing the characters in the initial erroneous audio information.

[0104] In specific implementation, for example, the voice information received by the voice error correction device is "navigate to Chengwa Avenue", it is detected that there is an error in the voice information, the character to be corrected is determined as "wa", the target syllable "ua" is obtained through phoneme replacement, then the target syllable "ua" is corrected according to a preset error correction strategy, the corrected syllable "hua" is obtained, the corrected syllable is converted into the corresponding target character "hua", then the character to be corrected in the voice information is replaced with the target character, and the corrected voice information "navigate to Chenghua Avenue" is obtained.

[0105] In this embodiment, by replacing the to-be-replaced phoneme with the target phoneme, a target syllable is obtained, the target syllable is corrected according to a preset error correction strategy, the corrected syllable is converted into a corresponding target character, the to-be-corrected character is replaced with the target character, and corrected voice information is obtained. Because the present invention obtains the target syllable through phoneme replacement, then corrects the target syllable according to the preset error correction strategy to obtain the correct target character, then replaces the wrong character with the target character to obtain correct voice information, rapid error correction for wrong characters in voice information is realized, and error correction efficiency is improved. Compared with the standard error correction scheme that requires editing each phoneme in the wrong character, the present invention reduces the editing distance in the error correction process by replacing phonemes in the wrong character, improves error correction efficiency, and saves error correction time.

[0106] In addition, an embodiment of the present invention also provides a storage medium, where a voice error correction program is stored on the storage medium, and when executed by a processor, the voice error correction program implements the steps of the voice error correction method as described above.

[0107] Since the storage medium adopts all the technical solutions of all the above embodiments, it at least has all the beneficial effects brought by the technical solutions of the above embodiments, which will not be repeated one by one herein.

[0108] Refer to Figure 6 , Figure 6 which is a structural block diagram of the first embodiment of the voice error correction apparatus of the present invention.

[0109] As shown in Figure 6 , the voice error correction apparatus provided by the embodiment of the present invention includes: an error detection module 10, configured to, when it is detected that there is an error in voice information, determine a to-be-corrected character in the voice information; a syllable acquisition module 20, configured to acquire a to-be-corrected syllable corresponding to the to-be-corrected character; a replacement strategy module 30, configured to determine, according to a preset replacement strategy, a to-be-replaced phoneme in the to-be-corrected syllable and a target phoneme corresponding to the to-be-replaced phoneme; The replacement and correction module 40 is used to replace the phoneme to be replaced with the target phoneme to obtain the target syllable, and to correct the character to be corrected according to the target syllable to obtain the corrected speech information.

[0110] Furthermore, the replacement and error correction module 40 is also used to perform phoneme matching on each phoneme in the syllable to be corrected according to a preset replacement strategy, and to determine the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced according to the phoneme matching result.

[0111] Furthermore, the replacement and error correction module 40 is also used to determine the phoneme to be replaced in the syllable to be corrected according to the phoneme matching result when the phoneme matching is successful, and to determine the target phoneme corresponding to the phoneme to be replaced according to the preset replacement mapping table.

[0112] Furthermore, the replacement and error correction module 40 is also used to determine the phoneme composition type corresponding to the syllable to be corrected based on the matching result when the phoneme matching is successful, and to determine the phoneme to be replaced in the syllable to be corrected based on the phoneme composition type.

[0113] Furthermore, the replacement and error correction module 40 is also used to replace the phoneme to be replaced with the target phoneme to obtain the target syllable, correct the target syllable according to a preset error correction strategy, convert the corrected syllable into the corresponding target character, replace the character to be corrected with the target character, and obtain the corrected speech information.

[0114] Furthermore, the replacement and error correction module 40 is also used to determine the error correction editing distance of the target syllable according to a preset error correction strategy, correct the target syllable according to the error correction editing distance, and convert the corrected syllable into the corresponding target character.

[0115] Furthermore, the replacement and error correction module 40 is also used to determine the phoneme composition type corresponding to the syllable to be corrected when phoneme matching fails, and to determine whether there are any phonemes that need to be deleted in the syllable to be corrected based on the phoneme composition type. If there are, the phonemes that need to be deleted are deleted to obtain the modified syllable.

[0116] Furthermore, the replacement and error correction module 40 is also used to determine the phoneme to be replaced in the modified syllable and the target phoneme corresponding to the phoneme to be replaced according to the phoneme composition type.

[0117] Furthermore, the replacement and error correction module 40 is also used to determine the phoneme to be replaced in the modified syllable according to the phoneme composition type, and to determine the target phoneme corresponding to the phoneme to be replaced according to a preset replacement mapping table.

[0118] Furthermore, the replacement and error correction module 40 is also used to determine the phoneme to be replaced in the syllable to be corrected according to the phoneme composition type if it does not exist, and to determine the target phoneme corresponding to the phoneme to be replaced according to the preset replacement mapping table.

[0119] Furthermore, the replacement error correction module 40 is also used to obtain a standard error correction strategy, determine the standard editing distance of the target syllable according to the standard error correction strategy, and perform the step of correcting the target syllable according to the error correction editing distance when the error correction editing distance does not exceed the standard editing distance.

[0120] Furthermore, the replacement and error correction module 40 is also used to correct the target syllable according to the standard editing distance when the error correction editing distance exceeds the standard editing distance, and to perform the step of converting the corrected syllable into the corresponding target character.

[0121] This embodiment, upon detecting an error in the speech information, identifies the character to be corrected, obtains the corresponding syllable, and, based on a preset replacement strategy, determines the phoneme to be replaced and its target phoneme within that syllable. The phoneme to be replaced is then replaced with the target phoneme to obtain the target syllable. Finally, the character to be corrected is corrected based on the target syllable to obtain the corrected speech information. Because this invention identifies the syllable to be corrected in the speech information, determines the phoneme to be replaced and its target phoneme based on a preset replacement strategy, and then replaces the phoneme with the target phoneme, it achieves error correction of the syllable containing the erroneous character in the speech information, thereby rapidly correcting errors in the speech information and improving speech correction efficiency.

[0122] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solutions of the present invention. In specific applications, those skilled in the art can make settings as needed, and the present invention does not impose any restrictions on this.

[0123] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0124] In addition, for technical details not described in detail in this embodiment, please refer to the speech error correction method provided in any embodiment of the present invention, which will not be repeated here.

[0125] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0126] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0127] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0128] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A speech error correction method, characterized by, The speech correction method includes: When an error is detected in the voice information, the characters in the voice information that need to be corrected are identified. Obtain the syllable to be corrected corresponding to the character to be corrected; The phonemes to be replaced in the syllable to be corrected and the target phonemes corresponding to the phonemes to be replaced are determined according to a preset replacement strategy. The phoneme to be replaced is replaced with the target phoneme to obtain the target syllable, and the character to be corrected is corrected according to the target syllable to obtain the corrected speech information; The step of determining the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced according to a preset replacement strategy includes: Phoneme matching is performed on each phoneme in the syllable to be corrected according to a preset replacement strategy; When phoneme matching is successful, the phoneme to be replaced in the syllable to be corrected is determined according to the phoneme matching result; the target phoneme corresponding to the phoneme to be replaced is determined according to the preset replacement mapping table. When phoneme matching fails, the phoneme composition type corresponding to the syllable to be corrected is determined; based on the phoneme composition type, it is determined whether there are any phonemes that need to be deleted in the syllable to be corrected; if so, the phonemes that need to be deleted are deleted to obtain the modified syllable; based on the phoneme composition type, the phonemes to be replaced in the modified syllable and the target phonemes corresponding to the phonemes to be replaced are determined.

2. The voice correction method of claim 1, wherein, When phoneme matching is successful, determining the phonemes to be replaced in the syllable to be corrected based on the phoneme matching result includes: When phoneme matching is successful, the phoneme composition type corresponding to the syllable to be corrected is determined based on the matching result; The phonemes to be replaced in the syllable to be corrected are determined based on the phoneme composition type.

3. The voice correction method of claim 1, wherein, The step of determining the phoneme to be replaced in the modified syllable and the target phoneme corresponding to the phoneme to be replaced based on the phoneme composition type includes: The phonemes to be replaced in the modified syllable are determined according to the phoneme composition type; The target phoneme corresponding to the phoneme to be replaced is determined according to a preset replacement mapping table.

4. The speech error correction method as described in claim 1, characterized in that, After determining whether there are any phonemes that need to be deleted in the syllable to be corrected based on the phoneme composition type, the method further includes: If it does not exist, then determine the phonemes to be replaced in the syllable to be corrected based on the phoneme composition type. The target phoneme corresponding to the phoneme to be replaced is determined according to the preset replacement mapping table.

5. The speech error correction method as described in any one of claims 1 to 4, characterized in that, The step of replacing the phoneme to be replaced with the target phoneme to obtain the target syllable, and correcting the character to be corrected based on the target syllable to obtain the corrected speech information, includes: The phoneme to be replaced is replaced with the target phoneme to obtain the target syllable; The target syllable is corrected according to a preset error correction strategy, and the corrected syllable is converted into the corresponding target character. Replace the character to be corrected with the target character to obtain the corrected speech information.

6. The speech error correction method as described in claim 5, characterized in that, The step of correcting the target syllable according to a preset error correction strategy and converting the corrected syllable into the corresponding target character includes: The error correction editing distance of the target syllable is determined according to a preset error correction strategy; The target syllable is corrected based on the error correction editing distance; Convert the corrected syllables into their corresponding target characters.

7. The speech error correction method as described in claim 6, characterized in that, After determining the error correction editing distance of the target syllable according to a preset error correction strategy, the method includes: Obtain a standard error correction strategy, and determine the standard edit distance of the target syllable based on the standard error correction strategy; When the error correction editing distance does not exceed the standard editing distance, the step of correcting the target syllable based on the error correction editing distance is performed.

8. The speech error correction method as described in claim 7, characterized in that, After obtaining the standard error correction strategy and determining the standard edit distance of the target syllable based on the standard error correction strategy, the method further includes: When the error correction editing distance exceeds the standard editing distance, the target syllable is corrected according to the standard editing distance; Perform the step of converting the corrected syllables into the corresponding target characters.

9. A voice error correction device, characterized in that, The voice error correction device includes: An error detection module is used to determine the characters to be corrected in the voice information when an error is detected. The syllable acquisition module is used to acquire the syllable to be corrected corresponding to the character to be corrected. The replacement strategy module is used to determine the phoneme to be replaced in the syllable to be corrected and the target phoneme corresponding to the phoneme to be replaced according to a preset replacement strategy. The replacement and correction module is used to replace the phoneme to be replaced with the target phoneme to obtain the target syllable, and to correct the character to be corrected according to the target syllable to obtain the corrected speech information. The replacement strategy module is used to perform phoneme matching on each phoneme in the syllable to be corrected according to a preset replacement strategy; when the phoneme matching is successful, the module determines the phoneme to be replaced in the syllable to be corrected based on the phoneme matching result; determines the target phoneme corresponding to the phoneme to be replaced based on a preset replacement mapping table; when the phoneme matching fails, the module determines the phoneme composition type corresponding to the syllable to be corrected; determines whether there is a phoneme to be deleted in the syllable to be corrected based on the phoneme composition type; if there is, the module deletes the phoneme to be deleted to obtain the modified syllable; and determines the phoneme to be replaced in the modified syllable and the target phoneme corresponding to the phoneme to be replaced based on the phoneme composition type.

10. The speech correction device as described in claim 9, characterized in that, The replacement strategy module is further configured to, when phoneme matching is successful, determine the phoneme composition type corresponding to the syllable to be corrected based on the matching result, and determine the phonemes to be replaced in the syllable to be corrected based on the phoneme composition type.

11. The speech correction device as described in any one of claims 9 to 10, characterized in that, The replacement and error correction module is further configured to replace the phoneme to be replaced with the target phoneme to obtain the target syllable, correct the target syllable according to a preset error correction strategy, convert the corrected syllable into the corresponding target character, replace the character to be corrected with the target character, and obtain the corrected speech information.

12. The speech correction device as described in claim 11, characterized in that, The replacement and error correction module is further configured to determine the error correction editing distance of the target syllable according to a preset error correction strategy, correct the target syllable according to the error correction editing distance, and convert the corrected syllable into the corresponding target character.

13. A voice error correction device, characterized in that, The voice correction device includes: a memory, a processor, and a voice correction program stored in the memory and executable on the processor, the voice correction program being configured to implement the voice correction method as described in any one of claims 1 to 8.

14. A storage medium, characterized in that, The storage medium stores a voice error correction program, which, when executed by a processor, implements the voice error correction method as described in any one of claims 1 to 8.

Images