CN115376509A - Method and device for implementing voice recognition interaction - Google Patents

Abstract

The invention relates to a method and a device for realizing voice recognition interaction, wherein the method comprises the steps of obtaining voice to be recognized, sending the voice to be recognized to a voice awakening engine for analysis and processing, and outputting an awakening result and awakening delay information; and sending the awakening delay information to a voice recognition engine, and outputting a recognition result. According to the invention, the problem of multiple words or few words can be recognized through boundary processing and awakening time delay set in the voice awakening engine, so that the voice identification accuracy is higher, the reaction speed and accuracy of the intelligent voice assistant are improved, and the intelligent voice becomes more intelligent.

Description

Method and device for implementing voice recognition interaction

technical field

The invention belongs to the technical field of artificial intelligence, and in particular relates to a method and device for realizing voice recognition interaction.

Background technique

With the continuous development of artificial intelligence and speech recognition technology, the demand for improving the efficiency of human-computer interaction is increasing, and there are higher requirements for the response speed and accuracy of the machine. Speech recognition is to convert "speech" into corresponding "text", and voice wake-up is to detect the characteristic segment of the speaker in real time in the continuous voice stream.

Oneshot is an interaction method that wakes up and recognizes a sentence. For example, if the wake-up word is "Hi Xiaoyi", you can say "Hi Xiaoyi, turn on the radio" at this time. This is a voice recognition interaction. The interactive mode of voice recognition requires a voice wake-up engine and a voice recognition engine. The initial speech recognition is to send a complete sentence to the recognition, which is processed by the speech recognition engine, and then the recognition results are cut. There are two problems in this method. One of the problems is that the recognition rate of wake-up words is not as strong as the ability to check wake-up keywords. If misrecognition occurs, the wrongly recognized wake-up words will not be cut, resulting in wrong recognition results. Another problem is that the wake-up word can be changed in design. If the wake-up word is changed, the recognition model is not compatible with the new wake-up word, which will lead to recognition errors.

In the related technology, the wake-up part of the sound data is not sent to the speech recognition engine by means of wake-up boundary processing. Although the above problems can be solved, a new problem has arisen. Because there is a time delay in the wake-up of the voice, it enters the recognition engine. Part of the speech will be lost, resulting in the inability to recognize correctly, this phenomenon is called word loss. If there is no wake-up delay, the missing posterior will lead to a serious increase in the false wake-up rate, which leads to the fact that the wake-up delay must exist objectively. If the wake-up delay is sent to the recognition engine together, this can solve the problem of missing words in recognition, but a new problem arises, the wake-up delay may include the end of the wake-up word, which will lead to recognition of possible multi-word problem, this phenomenon is called multi-word problem.

Contents of the invention

In view of this, the purpose of the present invention is to overcome the deficiencies of the prior art, and provide a method and device for implementing speech recognition interaction, so as to solve the problems of missing and multiple characters in speech recognition in the prior art.

In order to achieve the above object, the present invention adopts the following technical solutions: a method for implementing voice recognition interaction, comprising:

Obtain the voice to be recognized and send it to the voice wake-up engine for analysis and processing, output the wake-up result, and output the wake-up delay information after confirming the wake-up result;

Send the wake-up delay information to the speech recognition engine, and output the recognition result.

Further, the voice wake-up engine includes: a data processing unit and a boundary processing unit; the output wake-up result and wake-up delay information include:

The data processing unit analyzes the speech to be recognized, judges whether the wake-up condition is satisfied in the speech to be recognized, and outputs a wake-up result according to the judgment result;

The boundary processing unit processes the speech to be recognized to obtain wake-up delay information.

Further, judging whether the wake-up condition is satisfied in the speech to be recognized includes:

Judging whether there is a wake-up word in the voice to be recognized, if yes, the voice to be recognized meets the wake-up condition, otherwise the voice to be recognized does not meet the wake-up condition.

Further, the processing of the speech to be recognized to obtain wake-up delay information includes:

Analyzing the speech to be recognized, calculating an average value according to the values of interval sampling points in a preset time period, and determining the average value as a speech energy value;

It is judged according to the voice energy value whether to retain the voice data in the interval as the wake-up delay information; wherein, the wake-up delay information includes the ending sound of the last word of the wake-up word.

Further, the judging whether to retain the voice data in the interval as wake-up delay information according to the voice energy value includes:

If the voice energy value of the wake-up delay information is less than the energy value of the first frame in the wake-up delay information, discard the wake-up delay information;

If the voice energy value of the wake-up delay information is greater than the energy value of the first frame in the wake-up delay information, the line change delay information is retained.

Further, the preset time period is 10s.

An embodiment of the present application provides a device for realizing voice recognition interaction, including:

The wake-up module is used to obtain the voice to be recognized and send it to the voice wake-up engine for analysis and processing, output the wake-up result, and output wake-up delay information after confirming the wake-up result;

A recognition module, configured to send the wake-up delay information to a speech recognition engine, and output a recognition result.

Further, the wake-up module includes:

Data processing unit and boundary processing unit;

The data processing unit is used to analyze the speech to be recognized, judge whether the wake-up condition is satisfied in the speech to be recognized, and output a wake-up result according to the judgment result;

The boundary processing unit is configured to process the speech to be recognized to obtain wake-up delay information.

An embodiment of the present application provides a computer device, including: a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor performs any of the above speech recognition interaction steps of the implementation method.

The embodiment of the present application also provides a computer storage medium storing a computer program, and when the computer program is executed by a processor, the processor is made to perform the steps of any one of the methods for implementing voice recognition interaction described above.

The present invention adopts the above technical scheme, and the beneficial effects that can be achieved include:

The present invention provides a method and device for implementing voice recognition interaction. The method includes acquiring the voice to be recognized and sending it to the voice wake-up engine for analysis and processing, outputting the wake-up result and wake-up delay information; sending the wake-up delay information To the speech recognition engine to output the recognition result. The present invention can recognize the problem of many words or few words through the boundary processing and wake-up delay set in the voice wake-up engine, so that the accuracy of voice recognition is higher, the response speed and accuracy of the intelligent voice assistant are improved, and the intelligent voice changes more intelligent.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is a schematic diagram of the steps of the implementation method of voice recognition interaction of the present invention;

FIG. 2 is a schematic flow diagram of a method for implementing voice recognition interaction in the present invention;

FIG. 3 is a schematic structural diagram of a device for implementing voice recognition interaction according to the present invention;

FIG. 4 is a schematic diagram of the hardware structure of the operating environment of the implementation method of voice recognition interaction in the present invention.

Detailed ways

In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described in detail below. Apparently, the described embodiments are only some of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other implementations obtained by persons of ordinary skill in the art without making creative efforts fall within the protection scope of the present invention.

A specific method and device for implementing voice recognition interaction provided in the embodiments of the present application will be described below with reference to the accompanying drawings.

As shown in Figure 1, the implementation method of the speech recognition interaction provided in the embodiment of the present application includes:

S101. Obtain the voice to be recognized and send it to the voice wake-up engine for analysis and processing, output the wake-up result, and output wake-up delay information after confirming the wake-up result;

In some embodiments, the voice wake-up engine includes: a data processing unit and a boundary processing unit; the output wake-up result and wake-up delay information include:

The data processing unit analyzes the speech to be recognized, judges whether the wake-up condition is satisfied in the speech to be recognized, and outputs a wake-up result according to the judgment result;

The boundary processing unit processes the speech to be recognized to obtain wake-up delay information.

In some embodiments, judging whether the wake-up condition is satisfied in the speech to be recognized includes:

Judging whether there is a wake-up word in the voice to be recognized, if yes, the voice to be recognized meets the wake-up condition, otherwise the voice to be recognized does not meet the wake-up condition.

Specifically, as shown in FIG. 2 , in this application, the voice to be recognized is first obtained. For example, the voice to be recognized is: "Hi Hao Xiaoyi, open XXAPP", wherein "Hi Hao Xiaoyi" is a wake-up word, which is understandable Yes, the wake-up word is stored in the voice wake-up engine through the wake-up algorithm, and the wake-up word can include multiple, which is not limited in this application. After the data processing unit analyzes the voice to be recognized and recognizes the wake-up word, the wake-up result is wake-up, and the boundary processing unit processes the voice to be recognized to obtain wake-up delay information. If the data processing unit fails to recognize the wake-up word, the wake-up result is a failure, and the wake-up cannot be performed, and the wake-up delay information will not be output.

In some embodiments, the processing the speech to be recognized to obtain wake-up delay information includes:

Analyzing the speech to be recognized, calculating an average value according to the values of interval sampling points in a preset time period, and determining the average value as a speech energy value;

It is judged according to the voice energy value whether to retain the voice data in the interval as the wake-up delay information; wherein, the wake-up delay information includes the ending sound of the last word of the wake-up word.

In a specific embodiment, the judging whether to retain the voice data in the interval as the wake-up delay information according to the voice energy value includes:

If the voice energy value of the wake-up delay information is less than the energy value of the first frame in the wake-up delay information, discard the wake-up delay information;

If the voice energy value of the wake-up delay information is greater than the energy value of the first frame in the wake-up delay information, the line change delay information is retained.

In some embodiments, the preset time period is 10s.

Specifically, the present application processes the voice data, and selectively discards or reserves it for the recognition engine according to the data energy value of the wake-up delay segment. The energy value is calculated based on the average value of the interval sampling point, for example, the interval can be 10ms.

Specifically, for example, select a speech interval of the speech to be recognized as the wake-up delay segment, wherein the wake-up delay segment includes the end sound of the last word of the speech to be recognized, calculate the energy value of a plurality of sampling points in the wake-up delay segment, and then Taking the average value of multiple energy values, using the average value as the voice energy value, comparing the voice energy value with the energy value of the first frame in the wake-up delay information, if the voice energy value is less than the first frame in the wake-up delay information The energy value of the frame, then discard the wake-up delay information to solve the problem of multiple words. If the voice energy value is greater than the energy value of the first frame in the wake-up delay information, the wake-up segment will be retained and sent to the recognition engine for processing to solve the problem of word loss.

S102. Send the wake-up delay information to a speech recognition engine, and output a recognition result.

Finally, the wake-up delay information is sent to the speech recognition engine for recognition, and the recognition result is output.

The working principle of the implementation method of voice recognition interaction is as follows: Referring to Fig. 2, first use the data processing unit in the voice wake-up engine to analyze the voice to be recognized and recognize the wake-up word, the wake-up result is wake-up, and the boundary processing unit now Recognize the voice for processing, and get the wake-up delay information. If the data processing unit fails to recognize the wake-up word, the wake-up result is a failure, and the wake-up cannot be performed, and the wake-up delay information will not be output. After the wake-up delay information is obtained, it is sent to the speech recognition engine for recognition, and the recognition result is obtained. The method for implementing voice recognition interaction provided by the present application can solve the problem of multiple characters or missing characters during voice recognition, and improve the recognition accuracy.

As shown in Figure 3, the embodiment of the present application provides a device for implementing speech recognition interaction, including:

The wake-up module 301 is used to obtain the voice to be recognized and send it to the voice wake-up engine for analysis and processing, output the wake-up result, and output wake-up delay information after determining the wake-up result;

The recognition module 302 is configured to send the wake-up delay information to a voice recognition engine, and output a recognition result.

The working principle of the implementation device for voice recognition interaction provided by this application is that the wake-up module 301 acquires the voice to be recognized and sends it to the voice wake-up engine for analysis and processing, outputs the wake-up result, and outputs wake-up delay information after confirming the wake-up result; the recognition module 302 Send the wake-up delay information to the speech recognition engine, and output the recognition result.

The present application provides a computer device, including: a memory and a processor, and may also include a network interface, the memory stores a computer program, and the memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and/or non-volatile memory such as read-only memory (ROM) or flash memory (flashRAM). The computer device stores an operating system and the memory is an example of a computer readable medium. When the computer program is executed by the processor, the processor executes the implementation method of speech recognition interaction. The structure shown in FIG. The limitation of the computer device to which the application scheme is applied, the specific computer device may include more or less components than those shown in the figure, or combine certain components, or have a different arrangement of components.

In one embodiment, the method for implementing speech recognition interaction provided in this application may be implemented in the form of a computer program, and the computer program may run on the computer device as shown in FIG. 4 .

In some embodiments, when the computer program is executed by the processor, the processor performs the following steps: acquire the voice to be recognized and send it to the voice wake-up engine for analysis and processing, output the wake-up result, and output the wake-up result after determining the wake-up result Wake-up delay information; sending the wake-up delay information to a speech recognition engine, and outputting a recognition result.

The present application also provides a computer storage medium. 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 technologies, compact disc read-only memory (CD-ROM), digital multiplayer Functional Optical Disc (DVD) or other optical storage, magnetic cassette storage or other magnetic storage device, or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

In some embodiments, the present invention also proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by the processor, the speech to be recognized is acquired and sent to the voice wake-up engine for analysis and processing, and the wake-up call is output. As a result, the wake-up delay information is output after the wake-up result is determined; the wake-up delay information is sent to the speech recognition engine, and the recognition result is output.

To sum up, the present invention provides a method and device for realizing voice recognition interaction. The method includes acquiring the voice to be recognized and sending it to the voice wake-up engine for analysis and processing, outputting the wake-up result and wake-up delay information; The wake-up delay information is sent to the speech recognition engine, and the recognition result is output. The present invention can recognize the problem of many words or few words through the boundary processing and wake-up delay set in the voice wake-up engine, so that the accuracy of voice recognition is higher, the response speed and accuracy of the intelligent voice assistant are improved, and the intelligent voice changes more intelligent.

It can be understood that the method embodiments provided above correspond to the above device embodiments, and the corresponding specific contents can be referred to each other, and will not be repeated here.

Those skilled in the art should understand that the embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) having computer-usable program code embodied therein.

The present application is described with reference to flowcharts and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

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, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising the instructions, the instructions The method implements the function specified in the procedure or procedures of the flowchart and/or the block or blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

The above is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Anyone skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present invention. Should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims (10)

1. A method for implementing voice recognition interaction is characterized by comprising the following steps:

acquiring a voice to be recognized, sending the voice to be recognized to a voice awakening engine for analysis and processing, outputting an awakening result, and outputting awakening delay information after the awakening result is determined;

and sending the awakening delay information to a voice recognition engine, and outputting a recognition result.

2. The method of claim 1, wherein the voice wake engine comprises: a data processing unit and a boundary processing unit; the outputting of the wake-up result and the wake-up delay information includes:

the data processing unit analyzes the voice to be recognized, judges whether the voice to be recognized meets an awakening condition or not, and outputs an awakening result according to a judgment result;

and the boundary processing unit processes the voice to be recognized to obtain awakening delay information.

3. The method according to claim 2, wherein determining whether the voice to be recognized satisfies a wake-up condition comprises:

judging whether a wake-up word exists in the voice to be recognized, if so, judging that the voice to be recognized meets a wake-up condition, otherwise, judging that the voice to be recognized does not meet the wake-up condition.

4. The method according to claim 2, wherein the processing the speech to be recognized to obtain wake-up delay information comprises:

analyzing the voice to be recognized, calculating an average value according to values of interval sampling points of a preset time period, and determining the average value as a voice energy value;

judging whether voice data of the interval is reserved as awakening delay information or not according to the voice energy value; and the wake-up delay information contains the tail tone of the last word of the wake-up word.

5. The method according to claim 4, wherein the determining whether to reserve the voice data of the interval as the wake-up delay information according to the voice energy value comprises:

if the voice energy value of the awakening delay information is smaller than the energy value of the first frame in the awakening delay information, discarding the awakening delay information;

and if the voice energy value of the awakening delay information is larger than the energy value of the first frame in the awakening delay information, reserving the line-changing delay information.

6. The method of claim 4,

the preset time period is 10s.

7. An apparatus for implementing voice recognition interaction, comprising:

the awakening module is used for acquiring the voice to be recognized, sending the voice to be recognized to the voice awakening engine for analysis and processing, outputting an awakening result, and outputting awakening delay information after the awakening result is determined;

and the recognition module is used for sending the awakening delay information to a voice recognition engine and outputting a recognition result.

8. The apparatus of claim 7, wherein the wake-up module comprises:

a data processing unit and a boundary processing unit;

the data processing unit is used for analyzing the voice to be recognized, judging whether the voice to be recognized meets an awakening condition or not, and outputting an awakening result according to a judgment result;

and the boundary processing unit is used for processing the voice to be recognized to obtain awakening delay information.

9. A computer device, comprising: a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform a method of implementing a speech recognition interaction according to any one of claims 1 to 6.

10. A computer storage medium, characterized in that a computer program is stored which, when executed by a processor, causes the processor to carry out an implementation method of a speech recognition interaction according to any one of claims 1 to 6.

Images