KR20210052563A - Method and apparatus for providing context-based voice recognition service - Google Patents

Abstract

The present invention relates to a method and apparatus for recognizing voice. More specifically, the voice recognition apparatus according to the present invention may obtain voice information from a user and convert the obtained voice information into voice data.
Thereafter, the voice recognition model generates a first voice recognition result by recognizing the converted voice data with a first voice recognition model, and generates a second voice recognition result by recognizing the converted voice data with a second voice recognition model. And, a specific voice recognition result may be selected from among the first voice recognition result and the second voice recognition result through a specific determination procedure.

Description

Method and apparatus for providing context-based voice recognition service

The present invention relates to a method and apparatus for recognizing a user's voice. More specifically, it relates to a method and apparatus for improving speech recognition accuracy based on context in a method for recognizing a speech acquired from a user.

Automatic voice recognition (hereinafter referred to as voice recognition) is a technology that converts voice into text using a computer. Such speech recognition has achieved a rapid improvement in recognition rate in recent years.

However, although the recognition rate has improved as a whole, performance differences occur depending on the structure of the data or the structure of the data used when learning a language model or an acoustic model.

An object of the present invention is to provide a method for selecting a speech recognition result with high accuracy from among a plurality of speech recognition results when speech is recognized using a plurality of speech recognition models.

Another object is to provide a method for selecting a speech recognition model for speech recognition using context information.

The technical problems to be achieved in the present invention are not limited to the technical problems mentioned above, and other technical problems that are not mentioned will be clearly understood by those of ordinary skill in the technical field to which the present invention belongs from the following description. I will be able to.

A method of recognizing a voice according to the present invention comprises: obtaining voice information from a user; Converting the obtained voice information into voice data; Generating a first speech recognition result by recognizing the converted speech data with a first speech recognition model; Generating a second speech recognition result by recognizing the converted speech data with a second speech recognition model; And selecting a specific speech recognition result from among the first speech recognition result and the second speech recognition result through a specific determination procedure.

In addition, in the present invention, the specific determination procedure includes: extracting context information from the first voice recognition result and the second voice recognition result; Comparing the context information with a preset first characteristic of the first speech recognition model and a second characteristic of the second speech recognition model, respectively; And selecting one of the first voice recognition result and the second voice recognition result based on the comparison result.

In addition, in the present invention, the context information may include at least one of a part of the voice information, information that can be obtained from the first voice recognition result and the second voice recognition result, or information related to the user who uttered the voice. have.

In addition, in the present invention, the first voice recognition model and the second voice recognition model are one of a plurality of voice recognition models for recognizing the voice information obtained from the user.

In addition, the present invention further comprises the step of generating a plurality of speech recognition results by recognizing the converted speech data with the plurality of speech recognition models, wherein the specific speech recognition result is the first speech recognition result, the The second voice recognition result and the plurality of voice recognition results are selected.

In addition, in the present invention, the specific determination procedure is a procedure for determining a speech recognition result based on the context included in the context information.

In addition, the present invention, obtaining voice information from a user; Converting the obtained voice information into voice data; Generating a first speech recognition result by recognizing the speech data with the first speech recognition model; Selecting a second voice recognition model for recognizing the voice data from among a plurality of voice recognition models based on the first voice recognition result; And generating a second speech recognition result by recognizing the speech data with the second speech recognition model.

In addition, the present invention includes the steps of extracting context information from the first speech recognition result; And comparing the context information with preset specifics of the plurality of speech recognition models, wherein the second speech recognition model is selected based on the comparison result.

In addition, in the present invention, the first speech recognition model is a speech recognition model for extracting the context information.

In addition, the present invention, obtaining voice information from a user; Converting the obtained voice information into voice data; And generating a speech recognition result by recognizing the speech data with a specific speech recognition model selected from among a plurality of speech recognition models.

In addition, the present invention, the step of setting context information for speech recognition; And selecting the specific speech recognition model from among the plurality of speech recognition models whose characteristics are most suitable for the context information.

According to an embodiment of the present invention, when a plurality of results are generated using a plurality of voice recognition models when recognizing a voice input, the accuracy of speech recognition is selected by selecting a recognition result of a food recognition model with high accuracy among them. Can increase.

In addition, by selecting a speech recognition model according to context information, each of a plurality of speech recognition models can be used according to a purpose.

In addition, it is possible to select an appropriate voice recognition model even in a service for a large-scale user or in an environment where the physical and contextual environment in which the user is located changes from time to time.

In addition, since it is possible to select an appropriate speech recognition model, it is possible to reduce misrecognition due to similar vocabulary that may occur while applying a large language model, and reduce misrecognition due to unregistered vocabulary that may occur while applying a small language model. .

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are included as part of the detailed description to aid in understanding of the present invention, provide embodiments of the present invention, and together with the detailed description, the technical features of the present invention are described.
1 is a block diagram of a voice recognition apparatus according to an embodiment of the present invention.
2 and 3 are diagrams illustrating an example of a speech recognition apparatus according to an embodiment of the present invention.
4 and 5 are diagrams illustrating still another example of a speech recognition apparatus according to an embodiment of the present invention.
6 is a diagram illustrating another example of a speech recognition apparatus according to an embodiment of the present invention.
7 is a flowchart illustrating an example of a speech recognition method according to an embodiment of the present invention.
8 is a flow chart illustrating another example of a voice recognition method according to an embodiment of the present invention.
9 is a flowchart illustrating still another example of a voice recognition method according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. The detailed description to be disclosed below together with the accompanying drawings is intended to describe exemplary embodiments of the present invention, and is not intended to represent the only embodiments in which the present invention may be practiced. The following detailed description includes specific details to provide a thorough understanding of the present invention. However, those skilled in the art know that the invention may be practiced without these specific details.

In some cases, in order to avoid obscuring the concept of the present invention, well-known structures and devices may be omitted or illustrated in a block diagram form centering on core functions of each structure and device.

1 is a block diagram of a voice recognition apparatus according to an embodiment of the present invention.

Referring to FIG. 1, a voice recognition apparatus 100 for recognizing a user's voice may include an input unit 110, a storage unit 120, a control unit 130, and/or an output unit 140.

Since the components shown in FIG. 1 are not essential, an electronic device having more components or fewer components may be implemented.

Hereinafter, the above components will be described in order.

The input unit 110 may receive an audio signal, a video signal, or audio information (or audio signal) and data from a user.

The input unit 110 may include a camera and a microphone to receive an audio signal or a video signal. The camera processes image frames such as still images or moving pictures obtained by an image sensor in a video call mode or a photographing mode.

The image frames processed by the camera may be stored in the storage unit 120.

The microphone receives an external sound signal by a microphone in a call mode, recording mode, or voice recognition mode, and processes it as electrical voice data. Various noise removal algorithms may be implemented in the microphone to remove noise generated in the process of receiving an external sound signal.

When a user's uttered voice is input through a microphone or a microphone, the input unit 110 converts the voice into an electrical signal and transmits it to the controller 130.

The control unit 130 may obtain the user's voice data by applying a speech recognition algorithm or a speech recognition engine to the signal received from the input unit 110.

At this time, the signal input to the control unit 130 may be converted into a more useful form for speech recognition, and the control unit 130 converts the input signal from an analog form to a digital form, and detects the start and end points of the voice. Thus, the actual voice section/data included in the voice data can be detected. This is called EPD (End Point Detection).

And, the control unit 130 within the detected section Cepstrum (Cepstrum), linear predictive coding (Linear Predictive Coefficient: LPC), Mel Frequency Cepstral Coefficient (MFCC) or filter bank energy (Filter Bank). Energy), etc., can be applied to extract a feature vector of a signal.

The memory 120 may store a program for the operation of the controller 130 and may temporarily store input/output data. A sample file for a symbol-based malicious code detection model can be saved from a user, and the analysis result of malicious code can be saved.

The memory 120 may store various data related to the recognized voice, and in particular, may store information and feature vectors related to an end point of the voice data processed by the controller 130.

The memory 120 includes flash memory, hard disc, memory card, ROM (Read-Only Memory), RAM (Random Access Memory), memory card, EEPROM (Electrically Erasable Programmable Read). -Only Memory), Programmable Read-Only Memory (PROM), magnetic memory, magnetic disk, and optical disk.

In addition, the controller 130 may obtain a recognition result by comparing the extracted feature vector with the trained reference pattern. To this end, a speech recognition model for modeling and comparing signal characteristics of speech and a language model for modeling a linguistic order relationship such as words or syllables corresponding to the recognized vocabulary may be used.

The speech recognition model can be divided into a direct comparison method that sets the recognition object as a feature vector model and compares it with the feature vector of speech data, and a statistical method that statistically processes and uses the feature vector of the recognition object.

The direct comparison method is a method of setting units such as words and phonemes to be recognized as a feature vector model and comparing how similar the input speech is to this, and representatively, there is a vector quantization method. According to the vector quantization method, a feature vector of input speech data is mapped with a codebook, which is a reference model, and encoded as a representative value, thereby comparing the code values with each other.

The statistical model method is a method of configuring the unit for the recognition object as a state sequence and using the relationship between the state sequences. The state column may be composed of a plurality of nodes. Methods of using the relationship between state sequences include Dynamic Time Warping (DTW), Hidden Markov Model (HMM), and neural networks.

Dynamic time warping is a method of compensating for differences in the time axis when compared with the reference model by considering the dynamic characteristics of the voice whose signal length varies with time even if the same person speaks the same pronunciation. And after assuming the Markov process with the observation probability of the node (output symbol) in each state, the state transition probability and the observation probability of the node are estimated through the learning data, and the probability of generating the input voice from the estimated model is calculated. It is a recognition technology.

On the other hand, a language model for modeling linguistic order relations such as words and syllables can reduce acoustic ambiguity and reduce recognition errors by applying order relations between units constituting a language to units obtained from speech recognition. Linguistic models include statistical language models and models based on Finite State Automata (FSA), and statistical language models use chain probabilities of words such as Unigram, Bigram, and Trigram.

The control unit 130 may use any of the above-described methods in recognizing the voice. For example, a voice recognition model to which the Hidden Markov model is applied may be used, or an N-best search method in which a voice recognition model and a language model are integrated may be used. The N-best search method can improve recognition performance by selecting up to N recognition result candidates using a speech recognition model and a language model, and then re-evaluating the ranking of these candidates.

The controller 130 may calculate a confidence score (or may be abbreviated as “confidence”) to secure the reliability of the recognition result.

Reliability score is a measure of how reliable the result is for a speech recognition result, and can be defined as a relative value for the probability of the speech being uttered from a phoneme or word that is a recognized result, or from another phoneme or word. have. Therefore, the reliability score may be expressed as a value between 0 and 1 or a value between 0 and 100. If the reliability score is greater than a preset threshold, the recognition result may be recognized, and if the reliability score is small, the recognition result may be rejected.

In addition to this, the reliability score may be obtained according to various conventional reliability score acquisition algorithms.

The control unit 130 may be implemented in a computer-readable recording medium using software, hardware, or a combination thereof. According to the hardware implementation, Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), processors, and microcontrollers It may be implemented using at least one of electrical units such as (microcontrollers) and micro-processors.

According to the software implementation, it may be implemented together with a separate software module that performs at least one function or operation, and the software code may be implemented by a software application written in an appropriate programming language.

The control unit 130 implements the functions, processes, and/or methods proposed in FIGS. 2 to 6 to be described later, and hereinafter, for convenience of explanation, the control unit 130 is identical with the voice recognition device 100 and described. do.

The output unit 140 is for generating output related to vision, hearing, and the like, and outputs information processed by the device 100.

For example, the output unit 140 may output a recognition result of the voice signal processed by the control unit 130 so that the user can recognize it through visual or audio.

The voice recognition model described below may recognize voice information input from a user through the same method as the voice recognition model described in FIG. 1.

2 and 3 are diagrams illustrating an example of a speech recognition apparatus according to an embodiment of the present invention.

2 and 3, the speech recognition apparatus recognizes speech data acquired from a user as a plurality of speech recognition models, and selects one of the recognized results from the plurality of speech recognition models based on context information to recognize speech. Service can be provided.

Specifically, the voice recognition apparatus may generate voice data by converting voice information input from a user into an electrical signal, and converting an analog signal, which is a converted electrical signal, into a digital signal.

Thereafter, the voice recognition model may recognize voice data using the first voice recognition model 2010 and the second voice recognition model 2020, respectively.

The voice recognition device uses two voice recognition results (voice recognition result 1(2030), voice recognition result 2(2040)) from the voice data converted from the voice signal input from the user using each of the basic voice recognition model and the user voice recognition model. ) Can be obtained.

The speech recognition apparatus applies the first speech recognition result and the second speech recognition result to a first specific determination procedure (for example, a first context-based appropriate speech recognition model determination technique) to provide a first speech recognition result and a second speech recognition result. Among the speech recognition results, a more suitable speech recognition result 2050 may be selected and output.

That is, the voice recognition apparatus may select a voice recognition result more suitable for the purpose of voice recognition among the first voice recognition result and the second voice recognition result through the first specific determination procedure, and may output the selected voice recognition result. .

For example, when the context information extracted from the first speech recognition result and the second speech recognition result is related to an address search, a speech recognition model more suitable for address search is selected from among the first speech recognition model and the second speech recognition model, The speech recognition result of the selected speech recognition model can be provided as a speech recognition service.

Hereinafter, a specific determination procedure will be described with reference to FIG. 3.

3 is a flowchart illustrating an example of a first specific determination procedure for determining an appropriate speech recognition model based on a context based on a first speech recognition result and a second speech recognition result.

As shown in FIG. 3, the first specific determination procedure is performed when the first voice recognition result 3010 and the second voice recognition result 3020 are generated from the first voice recognition model and the second voice recognition model, respectively. Based on the context 3032 extracted from the 1 voice recognition result 3010 and the second voice recognition result 3020, among the first voice recognition result 3010 and the second voice recognition result 3020, the purpose of speech recognition A suitable speech recognition model can be selected (3034).

Thereafter, the speech recognition apparatus may select 3036 and output 3040 a speech recognition result generated from the selected speech recognition model.

For example, in FIG. 3, in the first voice recognition result of'tell me the address of Lee Ki-tong' and the second voice recognition result of'tell me the address of Lee Gil-dong', the voice recognition apparatus determines'tell me the address' as context information.

Specifically, the speech recognition device may extract context information'Tell me an address' (3032) from'Tell me the address of Gi-Tong Lee' and'Tell me the address of Gil-dong Lee'.

Thereafter, the speech recognition apparatus compares the extracted context information with the characteristic of the first speech recognition model (first characteristic) and the characteristic of the second speech recognition model (second characteristic) to obtain a speech recognition model more suitable for the purpose of speech recognition. A first speech recognition model may be selected (3034).

Thereafter, the voice recognition apparatus may select a first voice recognition result of the selected first voice recognition model (3036), and may output a “tell me an address of this machine” which is the selected first voice recognition result.

In this case, in addition to the part of the recognition sentence recognized from the voice data as context information, all information that can be inferred through the recognition result or the like may be used as the context information.

For example, at least one of information related to the user, such as the user's location, the user's weather, the user's habits, the user's previous utterance context, the user's career, the user's job title, the user's financial status, the current time, and the user's language. Can be used as contextual information.

4 and 5 are diagrams illustrating still another example of a speech recognition apparatus according to an embodiment of the present invention.

4 and 5, the speech recognition apparatus recognizes speech data obtained from a user as a plurality of speech recognition models, and selects one of the recognized results from the plurality of speech recognition models based on context information to recognize speech. Service can be provided.

Specifically, the voice recognition apparatus may generate a first voice recognition result 4020 by recognizing voice information input from a user as the first voice recognition model 4010.

In this case, the first voice recognition model is a voice information model for extracting a context from voice information obtained from a user, and may be configured to use only a small resource according to the purpose of the voice information model according to the purpose of voice recognition.

The speech recognition device uses a second specific determination procedure (e.g., a second context-based appropriate speech recognition model determination technique), among a plurality of preset speech recognition models, which are most suitable for recognizing speech information input from a user. A recognition model can be selected (4030).

That is, the speech recognition apparatus may select a specific speech recognition model from among a plurality of speech recognition models based on the first speech recognition result according to the purpose and use of speech recognition.

For example, when context information extracted from the first speech recognition result is related to address search, a speech recognition model most suitable for address search among a plurality of candidate speech recognition models may be selected as a specific speech recognition model.

In this case, the second specific determination procedure includes a procedure of extracting context information from the first speech recognition result in order to select a specific speech recognition model, and selecting a specific speech recognition model using the extracted context information.

Thereafter, the voice recognition procedure may re-recognize voice data from which voice information input from the user is converted using a selected specific voice recognition model, and finally generate a voice recognition result 4040.

Hereinafter, a second specific determination procedure will be described with reference to FIG. 5.

5 is a flowchart illustrating an example of a second specific determination procedure for determining an appropriate speech recognition model based on a context based on a first speech recognition result and a second speech recognition result.

Specifically, the speech recognition apparatus generates (or receives input) the first speech recognition result 5010 in which the user's speech information is recognized by the first speech recognition model described in FIG. 4, and the generated first speech recognition result is Based on the second specific determination procedure, a specific speech recognition model most suitable for the purpose of speech recognition may be selected from among a plurality of (for example, N) speech recognition models (5020).

In this case, the second specific determination procedure includes a procedure of extracting context information from the first speech recognition result in order to select a specific speech recognition model, and selecting a specific speech recognition model using the extracted context information.

For example, from the first voice recognition result recognized through the first voice recognition model, “tell me the address of Gil-dong Lee”, “tell me the address” may be extracted as context information.

In this case, the first voice recognition model is a voice information model for extracting context from voice information obtained from a user as described above, and may be configured to use only small resources according to the purpose of the voice information model according to the purpose of voice recognition. have.

As for the context information, in addition to a part of the sentence recognized through the speech recognition model, all information that can be inferred through the recognition result can be used as the context information.

For example, at least one of information related to the user, such as the user's location, the user's weather, the user's habits, the user's previous utterance context, the user's career, the user's job title, the user's financial status, the current time, and the user's language. Can be used as contextual information.

Thereafter, the voice recognition apparatus may select a specific voice recognition model most suitable for the purpose of voice recognition from among a plurality of voice recognition models using the extracted context information,'Tell me an address,' as shown in FIG. 4.

Through this method, the speech recognition apparatus can select a specific speech recognition model most suitable for the purpose of speech recognition by extracting context information through a speech recognition model for acquiring context information.

6 is a diagram illustrating another example of a speech recognition apparatus according to an embodiment of the present invention.

Referring to FIG. 6, the speech recognition apparatus may select a specific speech recognition model from among a plurality of speech recognition models in advance by setting context information for speech recognition, and use the speech recognition result recognized through the selected speech recognition model. Thus, a voice recognition service can be provided.

Specifically, the speech recognition apparatus may select a specific speech recognition model determined to be most suitable for speech recognition from among a plurality of speech recognition models according to preset context information (6010).

For example, when the purpose and purpose of the voice recognition service is address search, the voice recognition apparatus may select a voice recognition model preset for address search among a plurality of voice recognition models as a specific voice recognition model.

Thereafter, the voice recognition model may generate a voice recognition result 6020 by recognizing voice data obtained from the user through the selected specific food model.

In this case, the voice data may mean data in which voice information obtained from a user is changed into an electrical signal, and an analog signal, which is a changed electrical signal, is changed into a digital signal.

7 is a flowchart illustrating an example of a speech recognition method according to an embodiment of the present invention.

Referring to FIG. 7, as shown in FIGS. 2 and 3, a speech recognition device generates a speech recognition result through a plurality of speech recognition devices, and selects the most suitable speech recognition result from the generated speech recognition results to recognize speech. Service can be provided.

Specifically, the voice recognition apparatus may obtain voice information from a user and convert the obtained voice information into voice data (S7010).

For example, the voice recognition apparatus may convert voice information acquired from a user into an electric signal, and convert an analog signal, which is a changed electric signal, into voice data which is a digital signal.

Thereafter, the voice recognition apparatus may recognize the voice data as a first voice recognition model and a second voice recognition model, respectively, and generate a first voice recognition result and a second voice recognition result (S7020 and S7030).

Thereafter, the voice recognition apparatus provides a voice recognition service by selecting a voice recognition result more suitable for the purpose of voice recognition from among the first voice recognition result and the second voice recognition result through the first specific determination procedure described in FIGS. 2 and 3. It can be done (S7040).

For example, the speech recognition apparatus extracts context information from the first speech recognition result and the second speech recognition result, and uses the extracted context information as a first characteristic of a preset first speech recognition model and a second speech recognition model. 2 characteristics and can be compared respectively.

Thereafter, the speech recognition apparatus may select a speech recognition model suitable for the purpose and/or purpose of speech recognition from among the first speech recognition model and the second speech recognition model based on the comparison result.

Thereafter, the voice recognition apparatus may select a second voice recognition result generated by the selected second voice recognition model as the voice recognition result, and provide a voice recognition service based on the selected second voice recognition result.

8 is a flow chart illustrating another example of a voice recognition method according to an embodiment of the present invention.

Referring to FIG. 8, the voice recognition model may extract context information through voice data and provide a voice recognition service based on the extracted context information.

First, step S8010 is the same as step S7010 of FIG. 7, so a description thereof will be omitted.

Thereafter, the speech recognition apparatus generates a first speech recognition result by recognizing speech data with the first speech recognition model (S8020).

At this time, the first voice recognition model is a voice information model for extracting context from voice information obtained from a user as described in FIG. 4, and is configured to use only small resources according to the purpose of the voice information model according to the purpose of voice recognition. Can be.

The speech recognition apparatus may extract context information from the first speech recognition result (S8030).

Context information may mean all information that can be inferred through recognition results, etc. in addition to a part of a sentence recognized through a speech recognition model.

For example, at least one of information related to the user, such as the user's location, the user's weather, the user's habits, the user's previous utterance context, the user's career, the user's job title, the user's financial status, the current time, and the user's language. Can be used as contextual information.

Thereafter, the voice recognition apparatus may select a specific voice recognition model most suitable for recognizing voice information input from a user from among a plurality of preset voice recognition models using the second specific determination procedure described in FIGS. 4 and 5 (S8040). ).

That is, the speech recognition apparatus may select a specific speech recognition model from among a plurality of speech recognition models based on the first speech recognition result according to the purpose and use of speech recognition.

For example, when context information extracted from the first speech recognition result is related to address search, a speech recognition model most suitable for address search among a plurality of candidate speech recognition models may be selected as a specific food recognition model.

In this case, the second specific determination procedure includes a procedure of extracting context information from the first speech recognition result in order to select a specific speech recognition model, and selecting a specific speech recognition model using the extracted context information.

Thereafter, the speech recognition procedure may re-recognize the voice data from which voice information input from the user is converted using the selected specific voice recognition model, and finally generate a voice recognition result (S8040).

Thereafter, the voice recognition apparatus may provide a voice recognition service based on a voice recognition result of recognizing voice data through a specific voice recognition model.

9 is a flowchart illustrating still another example of a voice recognition method according to an embodiment of the present invention.

Referring to FIG. 9, before receiving voice information from a user, the voice recognition device may select a specific voice recognition model from among a plurality of voice recognition models based on context information, and voice information input from a user through the selected voice recognition model. Can be recognized.

Specifically, the speech recognition apparatus may preset context information for speech recognition.

Context information may mean all information that can be inferred through recognition results, etc. in addition to a part of a sentence recognized through a speech recognition model.

For example, at least one of information related to the user, such as the user's location, the user's weather, the user's habits, the user's previous utterance context, the user's career, the user's job title, the user's financial status, the current time, and the user's language. Can be used as contextual information.

Thereafter, the speech recognition apparatus selects a specific speech recognition model according to the purpose/use of speech recognition from among the plurality of speech recognition models based on the context information (S9020).

For example, in the case of address search, the speech recognition apparatus may select a speech recognition model preset for address search among a plurality of speech recognition models as a specific speech recognition model.

Thereafter, when voice information is obtained from the user, the voice recognition apparatus may convert the obtained voice information into voice data (S9010).

For example, the voice recognition apparatus may convert voice information acquired from a user into an electric signal, and convert an analog signal, which is a changed electric signal, into voice data which is a digital signal.

Thereafter, the speech recognition apparatus may generate a speech recognition result by recognizing speech data using the selected specific speech recognition model (S9050).

Thereafter, the voice recognition apparatus may provide a voice recognition service based on a voice recognition result of recognizing voice data through a specific voice recognition model.

The embodiment according to the present invention may be implemented by various means, for example, hardware, firmware, software, or a combination thereof. In the case of implementation by hardware, an embodiment of the present invention is one or more ASICs (application specific integrated circuits), DSPs (digital signal processors), DSPDs (digital signal processing devices), PLDs (programmable logic devices), FPGAs ( field programmable gate arrays), processors, controllers, microcontrollers, microprocessors, etc.

In the case of implementation by firmware or software, an embodiment of the present invention may be implemented in the form of a module, procedure, or function that performs the functions or operations described above. The software code can be stored in a memory and driven by a processor. The memory may be located inside or outside the processor, and may exchange data with the processor through various known means.

It is obvious to those skilled in the art that the present invention can be embodied in other specific forms without departing from the essential features of the present invention. Therefore, the above detailed description should not be construed as restrictive in all respects and should be considered as illustrative. The scope of the present invention should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the present invention are included in the scope of the present invention.

The present invention can be applied to various fields of speech recognition technology, and the present invention can provide a method for selecting an optimal speech recognition model based on context.

Due to this characteristic, it is possible to derive the best speech recognition result when an unspecified voice input is received from a service using a number of voice recognition models with different strengths for each field.

This feature can be applied not only to voice recognition, but also to other AI services.

Claims (11)

In the method of recognizing speech,
Obtaining voice information from a user;
Converting the obtained voice information into voice data;
Generating a first speech recognition result by recognizing the converted speech data with a first speech recognition model;
Generating a second speech recognition result by recognizing the converted speech data with a second speech recognition model; And
And selecting a specific speech recognition result from among the first speech recognition result and the second speech recognition result through a specific determination procedure. The method of claim 1, wherein the specific determination procedure,
Extracting context information from the first speech recognition result and the second speech recognition result;
Comparing the context information with a preset first characteristic of the first speech recognition model and a second characteristic of the second speech recognition model, respectively; And
And selecting one of the first speech recognition result and the second speech recognition result based on the comparison result. The method of claim 2,
The context information includes at least one of a part of the voice information, information that can be obtained from the first voice recognition result and the second voice recognition result, or information related to a user who uttered the voice. The method of claim 1,
The first voice recognition model and the second voice recognition model are one of a plurality of voice recognition models for recognizing the voice information obtained from the user. The method of claim 1,
Further comprising the step of generating a plurality of speech recognition results by recognizing the converted speech data with the plurality of speech recognition models,
The specific voice recognition result is selected from among the first voice recognition result, the second voice recognition result, and the plurality of voice recognition results. The method of claim 1,
The specific determination procedure is a procedure for determining a speech recognition result based on a context included in context information. In the method of recognizing speech,
Obtaining voice information from a user;
Converting the obtained voice information into voice data;
Generating a first speech recognition result by recognizing the speech data with the first speech recognition model;
Selecting a second voice recognition model for recognizing the voice data from among a plurality of voice recognition models based on the first voice recognition result; And
And generating a second speech recognition result by recognizing the speech data with the second speech recognition model. The method of claim 7,
Extracting context information from the first speech recognition result; And
Comprising the step of comparing the context information and a predetermined specificity of the plurality of speech recognition models,
The second speech recognition model is selected based on the comparison result. The method of claim 8,
The first speech recognition model is a speech recognition model for extracting the context information. In the method of recognizing speech,
Obtaining voice information from a user;
Converting the obtained voice information into voice data; And
And generating a speech recognition result by recognizing the speech data with a specific speech recognition model selected from among a plurality of speech recognition models. The method of claim 10,
Setting context information for speech recognition; And
And selecting the specific speech recognition model from among the plurality of speech recognition models in which characteristics of the speech recognition model are most suitable for the context information.

Images