Speech recognition detection method and device, storage medium and computer equipment
Through automated speech recognition detection methods, voice text and recognition results are obtained and detection results are calculated, and the problem of high labor and time consumption in the prior art is solved, and the detection efficiency is improved.
Patent Information
- Application Number
- CN202411191095.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-12-06
AI Technical Summary
The existing voice recognition detection methods consume a lot of manpower and time, and the detection efficiency is inefficient and limited by the number of people.
By obtaining the output speech recognition results of several speech texts and the speech recognition module to be detected, the speech recognition detection results are calculated based on the speech text and recognition results, reducing manual processing operations, reducing manual demand, and saving time.
It improves the efficiency of speech recognition detection, reduces labor and time costs, and can detect the performance of multiple speech recognition modules simultaneously.
Smart Images

Figure CN119091860A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of speech recognition detection, and in particular to a speech recognition detection method, apparatus, storage medium and computer equipment. Background Art
[0002] Most existing smart home products are configured to add voice control functions on the device side. For example, by adding a voice recognition module, the received voice is encoded and converted to obtain the corresponding voice recognition results. However, the recognition accuracy of the voice recognition module will have a direct impact on the user's control of smart home products through voice. At present, development engineers need to spend a lot of time testing the voice recognition module, and during the test process, they also need to manually record whether each voice recognition result during the test is correct. Therefore, the existing voice recognition detection method is extremely labor-intensive and time-consuming, and is limited by the number of manpower, and the detection efficiency is also very low. Summary of the invention
[0003] The embodiments of the present application provide a speech recognition detection method, device, storage medium and computer equipment, which can save manpower and time consumed in speech recognition detection and improve detection efficiency. The technical solution is as follows:
[0004] The first embodiment of the present application provides a speech recognition detection method, comprising:
[0005] Get several voice texts;
[0006] Obtaining a speech recognition result output by at least one speech recognition module to be detected; the speech recognition result is a speech recognition result obtained by the speech recognition module to be detected recognizing the speech corresponding to each speech text played by the speech playing module;
[0007] According to each of the voice texts and each of the voice recognition results, the voice recognition detection results of each voice recognition module to be detected are obtained.
[0008] A second embodiment of the present application provides a speech recognition detection device, comprising:
[0009] A playback module is used to obtain a number of voice texts;
[0010] A speech recognition result acquisition module, used to acquire a speech recognition result output by at least one speech recognition module to be detected; the speech recognition result is a speech recognition result obtained by the speech recognition module to be detected recognizing the speech corresponding to each speech text played by the speech playing module;
[0011] The speech recognition detection result acquisition module is used to obtain the speech recognition detection results of each speech recognition module to be detected according to each of the speech texts and each of the speech recognition results.
[0012] A third embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the speech recognition detection method described above are implemented.
[0013] The fourth embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the steps of the speech recognition detection method as described above when executing the computer program.
[0014] Compared with the related art, the present application can obtain several voice texts and voice recognition results output by the voice recognition module to be detected; wherein, the voice recognition result is the voice recognition result obtained by the voice recognition module to be detected recognizing the voice corresponding to each voice text played by the voice playback module; then, according to each voice text and each voice recognition result, the voice recognition detection result of each voice recognition module to be detected is obtained. In the detection of voice recognition, the operations that need to be manually processed can be reduced, the number of manpower required for voice recognition detection can be reduced, and the manpower and time consumed by voice recognition detection can be saved. Moreover, the voice recognition detection results of each voice recognition module to be detected can be obtained according to each voice text and each voice recognition result of at least one voice recognition module to be detected at the same time, and the voice recognition detection of at least one voice recognition module to be detected can be performed at the same time, which is conducive to improving the detection efficiency.
[0015] In order to provide a clearer understanding of the present application, the specific implementation of the present application will be described below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0017] Figure 1 This is a flow chart of a speech recognition detection method according to an embodiment of the present application.
[0018] Figure 2 This is a flow chart of steps S101 - S103 of a speech recognition detection method according to an embodiment of the present application.
[0019] Figure 3 This is a schematic diagram of module connections of a speech recognition detection device according to an embodiment of the present application.
[0020] 100. Speech recognition detection device; 101. Playing module; 102. Speech recognition result acquisition module; 103. Speech recognition detection result acquisition module. DETAILED DESCRIPTION
[0021] In order to make the objectives, technical solutions and advantages of the present application more clear, the embodiments of the present application will be further described in detail below with reference to the accompanying drawings.
[0022] It should be clear that the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the embodiments of the present application.
[0023] When the following description relates to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are only used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances. The singular forms of "a", "said" and "the" used in the present application and the appended claims are also intended to include the majority form, unless the context clearly indicates other meanings. The words "if" / "if" used herein can be interpreted as "at the time of" or "when" or "in response to determination".
[0024] In addition, in the description of this application, unless otherwise specified, "plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship.
[0025] See also Figure 1 , which is a flow chart of the speech recognition detection method of the first embodiment of the present application, comprising:
[0026] S1: Obtain several voice texts.
[0027] Voice text refers to text content used for voice playback according to a preset playback order. For example, the voice text can be a preset entry text or a preset sentence text. Different voice texts can also express the same meaning. For example, for a clothes drying machine in a smart home, "raise", "rise", "lift up", "rise" and so on all mean to control the clothes drying rod to rise.
[0028] S2: Obtain a speech recognition result output by at least one speech recognition module to be detected; the speech recognition result is a speech recognition result obtained by the speech recognition module to be detected recognizing the speech corresponding to each speech text played by the speech playing module.
[0029] The speech recognition module to be detected refers to a device module for implementing speech recognition. Optionally, the speech recognition module to be detected can be a device module that implements speech recognition and outputs corresponding control instructions. Therefore, the speech recognition module to be detected in step S2 can be a device module that is not applied to a smart home, or it can be a device module operated by a smart home.
[0030] The voice playback module refers to a device module for playing voice texts one by one according to a preset playback order, wherein the voice playback module can be run on an audio playback device or on the execution subject of this application. Specifically, when the voice playback module is run on an audio playback device, each time the audio playback device plays a voice of a voice text, the execution subject of this application obtains the played voice text from the audio playback device for subsequent execution of steps S2 and S3; when the voice playback module is run on the execution subject of this application, each time the execution subject of this application plays a voice of a voice text, the played voice text is recorded for subsequent execution of steps S2 and S3.
[0031] The speech recognition result refers to the result of the speech recognition module to be detected recognizing the speech corresponding to the speech text, and the speech recognition result can be expressed in the form of text, control instructions, etc. For example, when the speech recognition module to be detected recognizes the speech corresponding to the speech text, the speech is converted into recognition text according to a preset encoding program as the speech recognition result, and the corresponding control instruction can also be obtained according to the recognition text as the speech recognition result.
[0032] S3: According to each of the voice texts and each of the voice recognition results, the voice recognition detection results of each voice recognition module to be detected are obtained.
[0033] Wherein, step S3 comprises:
[0034] S31: According to the preset correspondence between the speech text and the recognition result sample, a target recognition result sample corresponding to the speech text is obtained.
[0035] Among them, the correspondence between the speech text and the recognition result sample can be pre-stored in the execution subject of the present application, or stored in the cloud platform. When the correspondence between the speech text and the recognition result sample is stored in the cloud platform, it can be downloaded from the cloud platform before the execution subject of the present application executes step S31. Among them, in the correspondence between the speech text and the recognition result sample, one recognition result sample corresponds to at least one speech text, that is, multiple speech texts can correspond to the same recognition result sample. Specifically, the recognition result sample can be the text semantics corresponding to at least one speech text, or it can be a voice instruction sample corresponding to at least one speech text. The target recognition result sample refers to the recognition result sample corresponding to the speech text speech.
[0036] S32: Obtain recognition results corresponding to each of the voice texts according to each of the target recognition result samples and each of the voice recognition results.
[0037] Wherein, step S32 may include:
[0038] S321: Compare each target recognition result sample with the corresponding speech recognition result. If they are the same, determine that the recognition result is correct; if they are different, determine that the recognition result is abnormal.
[0039] Specifically, comparing the target recognition result sample with the corresponding speech recognition result means comparing the target recognition result sample with the speech recognition result obtained by speech recognition of the same speech text. In step S321, recognition anomaly refers to the abnormal situation that the speech recognition result output by the speech recognition module to be detected is different from the target recognition result sample, and the abnormal situation is that the speech recognition module to be detected recognizes misalignment.
[0040] S33: Obtaining a speech recognition detection result according to each of the recognition results.
[0041] It can be understood that the recognition result is the recognition result corresponding to each voice text, that is, the result of comparing the target recognition result sample corresponding to each voice text with the voice recognition result, and the voice recognition detection result is a comprehensive reflection of multiple recognition results. For example, the voice recognition detection result of the voice recognition module to be detected can be obtained based on the number of correct recognitions and / or abnormal recognitions in the recognition results. Specifically, the voice recognition detection result can be a parameter for evaluating the voice recognition performance of the voice recognition module to be detected. For example, the accuracy of voice recognition can be obtained based on the proportion of correct recognition results in the total recognition results.
[0042] It should be noted that in the present application, the voices of several voice texts can be played according to a preset order, and after the voices of all voice texts are played, it is considered that a round of playback is completed, and the voice texts played in this round are executed as steps S1-S3, and the voice recognition detection results of the voice texts played in the round by the voice recognition module to be detected can be obtained, and the voices of several voice texts can be played repeatedly for multiple rounds, and the playback parameters of each round can be adjusted by the user, and the playback parameters can be parameters such as the playback order and the playback volume. That is, the order and playback volume of each round of playback can be different.
[0043] Compared with the related art, the present application can obtain several voice texts and voice recognition results output by the voice recognition module to be detected; wherein, the voice recognition result is the voice recognition result obtained by the voice recognition module to be detected recognizing the voice corresponding to each voice text played by the voice playback module; then, according to each voice text and each voice recognition result, the voice recognition detection result of each voice recognition module to be detected is obtained. In the detection of voice recognition, the operations that need to be manually processed can be reduced, the number of manpower required for voice recognition detection can be reduced, and the manpower and time consumed by voice recognition detection can be saved. Moreover, the voice recognition detection results of each voice recognition module to be detected can be obtained according to each voice text and each voice recognition result of at least one voice recognition module to be detected at the same time, and the voice recognition detection of at least one voice recognition module to be detected can be performed at the same time, which is conducive to improving the detection efficiency.
[0044] In a feasible embodiment, after the step of determining that the recognition result is a recognition abnormality in S321, the following steps are included:
[0045] According to the correspondence between the speech text and the recognition result sample, the speech text corresponding to the speech recognition result with recognition abnormality is determined as the reduced sensitivity text, so that the speech recognition module to be detected reduces the speech recognition sensitivity to the reduced sensitivity text.
[0046] Among them, recognition abnormality may be caused by the excessive sensitivity of the voice recognition module to be detected to some voice recognition. Take the clothes drying machine as an example. The clothes drying machine includes a liftable clothes drying rod and a light group with adjustable brightness, and the currently played voice text is "increase the brightness". However, the voice recognition module to be detected has too high sensitivity to the voice recognition of "increase the height of the clothes drying rod", resulting in the voice recognition module to be detected may recognize "increase the brightness" as "increase the brightness of the light group" or "increase the height of the clothes drying rod", or obtain the control instruction corresponding to the recognized text. That is, the voice recognition result output by the voice recognition module to be detected may be wrong at this time, so it is best to perform multiple rounds of detection. When the voice recognition module to be detected recognizes "increase brightness" as "increase height of clothesline" because of its high sensitivity to the voice recognition of "increase height of clothesline", and obtains an erroneous voice recognition result, since the comparison result of the target recognition result sample and the voice recognition result is different, it is necessary to obtain a candidate recognition result sample that is the same as the voice recognition result from the correspondence between the voice text and the recognition result sample, and then obtain the voice text corresponding to the voice recognition result based on the candidate recognition result sample, so as to obtain a text with reduced sensitivity, so that the voice recognition module to be detected can reduce the voice recognition sensitivity to the text with reduced sensitivity. Among them, reducing the voice recognition sensitivity of the voice recognition module to be detected to the text with reduced sensitivity can be achieved by a preset code program, or by user operation modification.
[0047] In this embodiment, by determining the voice text as a text with reduced sensitivity, it is helpful to obtain the voice text for which the voice recognition sensitivity of the voice recognition module to be detected needs to be reduced.
[0048] See also Figure 2 In a feasible embodiment, after the step of obtaining a plurality of voice texts in S1, the method further includes:
[0049] S101: Acquire the voice playing time of the voice text.
[0050] The voice playing time may be the instant time when the voice text starts to be played, or the instant time when the voice text is finished being played. Specifically, the voice playing time may be specific to seconds.
[0051] S102: Obtaining a time limit for obtaining an instruction according to the voice playing time and a preset time delay value.
[0052] The time delay value can be set by the user, or can be set according to the time required for the speech recognition module to be detected to perform speech recognition and obtain the speech recognition result. The time delay value can be a few seconds, a dozen seconds or tens of seconds, etc. At this time, the corresponding instruction acquisition time limit refers to a few seconds, a dozen seconds or tens of seconds from the start of the speech playback time.
[0053] S103: If the speech recognition result output by the speech recognition module to be detected is not obtained within the instruction acquisition time limit, it is determined that the recognition result of the entry sample speech is a recognition abnormality.
[0054] It should be noted that the recognition anomaly in step S103 refers to an abnormal situation in which the speech recognition module to be detected fails to recognize the speech corresponding to the speech text and therefore fails to output a speech recognition result.
[0055] In this embodiment, according to the voice playback time of the voice text, the instruction acquisition time limit can be determined to determine whether there is a situation in which the voice recognition result output by the voice recognition module to be detected is not obtained within the instruction acquisition time limit, which is a recognition abnormality, thereby obtaining voice recognition results for more situations.
[0056] In a feasible embodiment, after the step S103, the following steps are included:
[0057] The speech text is determined as a sensitivity-enhanced text, so that the speech recognition module to be detected improves the speech recognition sensitivity to the sensitivity-enhanced text.
[0058] Among them, if the voice recognition result output by the voice recognition module to be detected is not obtained within the instruction acquisition time limit, and the recognition result of the entry sample voice is determined to be a recognition abnormality, it means that the voice recognition degree of the voice recognition module to be detected for the played voice text is low, resulting in the inability to recognize voice results. Therefore, it is necessary to improve the voice recognition sensitivity of the voice recognition module to be detected to the sensitivity-enhanced text, for example, determine the parameters that need to be modified for the voice recognition module to be detected according to the sensitivity-enhanced text, so as to improve the voice recognition sensitivity to the sensitivity-enhanced text.
[0059] In this embodiment, by determining the voice text as the sensitivity-enhanced text, it is helpful to obtain the voice text whose voice recognition sensitivity needs to be improved in the voice recognition module to be detected.
[0060] In a feasible embodiment, the step of S1: obtaining the voice text corresponding to each played voice sample of each entry includes:
[0061] S11: Input the voice text into a trained entry editing network model to obtain an extended voice text; the entry editing network model is used to edit the text structure of the voice text to obtain an extended voice text that is synonymous with the voice text; the extended voice text corresponds to the same control instruction sample as the voice text.
[0062] The entry editing network model can be obtained by training the neural network with the first speech text sample in the initial state and the second speech text sample after editing. The trained entry editing network model can be used to add, subtract and replace words in the speech text to obtain an extended speech text that is synonymous with the speech text. Since the extended speech text is synonymous with the speech text, the extended speech text corresponds to the same recognition result sample as the speech text, for example, the semantics or control instructions corresponding to the extended speech text are the same as those of the speech text.
[0063] S12: when playing the entry sample voice according to the voice text and / or the extended voice text, obtaining the voice text and / or the extended voice text corresponding to the entry sample voice.
[0064] In this embodiment, the text diversity of the played speech can be increased by editing the network model of the word entry, so as to more accurately detect the speech recognition accuracy of the speech recognition module to be detected.
[0065] Referring to 3, the second embodiment of the present application provides a speech recognition detection device 100, including:
[0066] Playing module 101, used to obtain a number of voice texts;
[0067] The speech recognition result acquisition module 102 is used to acquire the speech recognition result output by at least one speech recognition module to be detected; the speech recognition result is the speech recognition result obtained by the speech recognition module to be detected recognizing the speech corresponding to each speech text played by the speech playing module 101;
[0068] The speech recognition detection result acquisition module 103 is used to obtain the speech recognition detection results of each speech recognition module to be detected according to each of the speech texts and each of the speech recognition results.
[0069] It should be noted that when the speech recognition detection device 100 provided in the second embodiment of the present application executes the speech recognition detection method, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the speech recognition detection device 100 provided in the second embodiment of the present application and the speech recognition detection method of the first embodiment of the present application belong to the same concept. The implementation process is detailed in the method embodiment, which will not be repeated here.
[0070] A third embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the speech recognition detection method described above are implemented.
[0071] The fourth embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the steps of the speech recognition detection method as described above when executing the computer program.
[0072] The device embodiments described above are merely illustrative, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present application. Ordinary technicians in this field can understand and implement it without creative work.
[0073] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0074] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0075] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0076] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0077] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0078] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0079] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0080] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0081] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.
Claims
1. A speech recognition detection method, characterized in that: include: Get several voice texts; Obtaining a speech recognition result output by at least one speech recognition module to be detected; the speech recognition result is a speech recognition result obtained by the speech recognition module to be detected recognizing the speech corresponding to each speech text played by the speech playing module; According to each of the voice texts and each of the voice recognition results, the voice recognition detection results of each voice recognition module to be detected are obtained.
2. The speech recognition detection method according to claim 1, characterized in that: The step of obtaining the speech recognition detection result according to each of the speech texts and the corresponding speech recognition result comprises: According to the preset correspondence between the speech text and the recognition result sample, obtain the target recognition result sample corresponding to the speech text speech; Obtaining recognition results corresponding to each of the voice texts according to each of the target recognition result samples and each of the voice recognition results; According to each of the recognition results, a speech recognition detection result is obtained.
3. The speech recognition detection method according to claim 2, characterized in that: The step of obtaining the recognition results corresponding to the voice texts according to the target recognition result samples and the voice recognition results comprises: Each target recognition result sample is compared with the corresponding speech recognition result. If they are the same, the recognition result is determined to be correct; if they are different, the recognition result is determined to be abnormal.
4. The speech recognition detection method according to claim 3, characterized in that: After the step of determining that the recognition result is a recognition abnormality, the method further comprises: According to the correspondence between the speech text and the recognition result sample, the speech text corresponding to the speech recognition result with recognition abnormality is determined as the reduced sensitivity text, so that the speech recognition module to be detected reduces the speech recognition sensitivity to the reduced sensitivity text.
5. The speech recognition detection method according to claim 2, characterized in that: After the step of obtaining a plurality of voice texts, the method further includes: Obtaining the voice playing time of the voice text; Obtaining a time limit for obtaining a command according to the voice playing time and a preset time delay value; If the speech recognition result output by the speech recognition module to be detected is not obtained within the instruction acquisition time limit, it is determined that the recognition result of the entry sample speech is a recognition abnormality.
6. The speech recognition detection method according to claim 5, characterized in that: If the speech recognition result output by the speech recognition module to be detected is not obtained within the instruction acquisition time limit, after the step of determining that the recognition result of the entry sample speech is a recognition abnormality, the method includes: The speech text is determined as a sensitivity-enhanced text, so that the speech recognition module to be detected improves the speech recognition sensitivity to the sensitivity-enhanced text.
7. The speech recognition detection method according to claim 1, characterized in that: The step of obtaining the voice text corresponding to each played entry sample voice includes: The speech text is input into a trained entry editing network model to obtain an extended speech text; the entry editing network model is used to edit the text structure of the speech text to obtain an extended speech text that is synonymous with the speech text; the extended speech text corresponds to the same control instruction sample as the speech text; When playing the entry sample voice according to the voice text and / or the extended voice text, the voice text and / or the extended voice text corresponding to the entry sample voice is obtained.
8. A speech recognition detection device, characterized in that: include: A playback module is used to obtain a number of voice texts; A speech recognition result acquisition module, used to acquire a speech recognition result output by at least one speech recognition module to be detected; the speech recognition result is a speech recognition result obtained by the speech recognition module to be detected recognizing the speech corresponding to each speech text played by the speech playing module; The speech recognition detection result acquisition module is used to obtain the speech recognition detection results of each speech recognition module to be detected according to each of the speech texts and each of the speech recognition results.
9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the speech recognition detection method according to any one of claims 1 to 7 are implemented.
10. A computer device, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and executable by the processor, wherein the processor implements the steps of the speech recognition detection method according to any one of claims 1 to 7 when executing the computer program.
