CN106128462A - Speech Recognition Method and System - Google Patents

Abstract

The present invention provides a speech recognition method, which is applied to an electronic device, and the method includes: obtaining speech information input by a user; using a first speech recognition method to recognize the speech information to obtain a first speech recognition result, and using a second speech recognition method to recognize the speech information to obtain a second speech recognition result, wherein the first speech recognition method and the second speech recognition method are run in parallel; and displaying the first speech recognition result and the second speech recognition result according to a preset rule. The present invention also provides a speech recognition system. Using the present invention, the second speech recognition method can be used to assist the first speech recognition method in recognizing the user's speech information, thereby improving the speech recognition rate.

Description

Speech Recognition Method and System

【Technical field】

The invention relates to the technical field of voice recognition, in particular to a voice recognition method and system based on geographic location assistance.

【Background technique】

In recent years, speech recognition technology has made remarkable progress and has moved from the laboratory to the market. In practical applications, such as an automatic telephone answering system, the interaction with the user is automatically completed by recognizing the user's voice input information.

At present, Putonghua has been basically popularized as a communication language, but people in different regions still have different dialects. Therefore, due to the influence of dialects, the Mandarin of people in different places has different characteristics. Phonetics and semantics are similar.

Since the Putonghua of users in different regions is not completely standard and has regional characteristics, the existing voice input records the user's voice data and recognizes the output results according to the voice recognition algorithm, without combining the user's geographic location and other data as auxiliary references. The recognition rate of speech recognition with some local accents and some local dialects is not high.

【Content of invention】

In view of the above, it is necessary to provide a voice recognition method and system, which can call auxiliary voice data packets to recognize the user's voice according to the user's current geographic location, thereby improving the accuracy of voice recognition.

A speech recognition method applied to electronic equipment, the method comprising:

Obtain the voice information input by the user;

Using a first voice recognition method to recognize the voice information to obtain a first voice recognition result, using a second voice recognition method to recognize the voice information to obtain a second voice recognition result; and

The first speech recognition result and the second speech recognition result are displayed according to preset rules.

According to a preferred embodiment of the present invention, the first speech recognition method is a large vocabulary speech recognition method based on a preset model, and the second speech recognition method is a speech recognition method based on auxiliary speech data packets.

According to a preferred embodiment of the present invention, the voice recognition method based on auxiliary voice data packets includes:

When receiving the voice information, obtain the user's current geographic location information;

Invoking the corresponding auxiliary voice data package according to the geographic location information; and

Recognizing the voice information according to the auxiliary voice data packet to obtain the second voice recognition result.

According to a preferred embodiment of the present invention, the method also includes:

A plurality of voice data packets based on geographic location are preset, and the voice data packets are stored in the electronic device or in a server connected to the electronic device.

According to a preferred embodiment of the present invention, before invoking the corresponding auxiliary voice data package according to the geographic location information, the method further includes:

determining the voice type of the user according to the voice information, the voice type including accent and dialect; and

A corresponding auxiliary voice data packet is jointly determined based on the voice type and the geographic location information.

According to a preferred embodiment of the present invention, the method also includes:

When receiving the voice information, obtain the user's current geographic location information and historical geographic location information; and

The auxiliary voice data package to be called is determined according to the historical geographic location information and the current geographic location information.

According to a preferred embodiment of the present invention, the method also includes:

The preset rules are updated in combination with the obtained user feedback information, and the preset rules include:

Pre-assigning a first weight to the first speech recognition result, pre-allocating a second weight to the second speech recognition result, and determining the display mode of the speech recognition result corresponding to the weight value according to the weight value; or

Presetting a first recognition score for the first speech recognition result, presetting a second recognition score for the second speech recognition result, determining the display mode of the speech recognition result corresponding to the recognition score according to the size of the recognition score,

Wherein, the display manner includes display time or display position.

According to a preferred embodiment of the present invention, the updating of the preset rules includes:

According to the speech recognition result selected by the user, the weight value or recognition score corresponding to the speech recognition result is increased, and/or the weight value or recognition score corresponding to the speech recognition result not selected by the user is decreased.

A speech recognition system running in an electronic device, the system includes:

An acquisition module, configured to acquire voice information input by the user;

A first recognition module, configured to recognize the voice information to obtain a first voice recognition result;

A second recognition module, configured to recognize the voice information to obtain a second voice recognition result; and

A display module, configured to display the first speech recognition result and the second speech recognition result according to preset rules.

According to a preferred embodiment of the present invention, the first speech recognition module is a large vocabulary speech recognition module based on a preset model, and the second speech recognition module is a speech recognition module based on auxiliary speech data packets.

According to a preferred embodiment of the present invention,

The obtaining module also obtains the user's current geographic location information when the user receives the voice information;

The second identification module includes:

Calling a submodule, used to call the corresponding auxiliary voice data package according to the geographic location information; and

The second recognition module is configured to recognize the voice information according to the auxiliary voice data packet to obtain the second voice recognition result.

According to a preferred embodiment of the present invention, the system also includes:

The setting module is used to pre-set a plurality of voice data packages based on geographic location, and store the voice data packages in the electronic device or in a server connected to the electronic device.

According to a preferred embodiment of the present invention, the system also includes a determination submodule:

for determining the voice type of the user according to the voice information, the voice type including accent and dialect; and

A corresponding auxiliary voice data packet is jointly determined based on the voice type and the geographic location information.

According to a preferred embodiment of the present invention, it is characterized in that,

The acquiring module is further configured to acquire the user's current geographic location information and historical geographic location information when receiving the voice information; and

The invoking sub-module is also used to determine the auxiliary voice data package to be invoked according to historical geographic location information and current geographic location information.

According to a preferred embodiment of the present invention, the system also includes:

An update module, configured to update the preset rules in combination with the acquired user feedback information, the preset rules are set by the setting module, including:

Pre-assigning a first weight to the first speech recognition result, pre-allocating a second weight to the second speech recognition result, and determining the display mode of the speech recognition result corresponding to the weight value according to the weight value; or

Presetting a first recognition score for the first speech recognition result, presetting a second recognition score for the second speech recognition result, determining the display mode of the speech recognition result corresponding to the recognition score according to the size of the recognition score,

Wherein, the display manner includes display time or display position.

According to a preferred embodiment of the present invention, updating the preset rules by the update module includes:

According to the speech recognition result selected by the user, the weight value or recognition score corresponding to the speech recognition result is increased, and/or the weight value or recognition score corresponding to the speech recognition result not selected by the user is decreased.

As can be seen from the above technical solutions, the speech recognition method and system of the present invention can establish multiple auxiliary voice data packets according to the characteristics of Mandarin in different regions, and call different auxiliary voice data packets for users in different geographic locations, which can effectively Reduce the types of speech recognition libraries and improve the speech recognition rate.

【Description of drawings】

FIG. 1 is a schematic diagram of the hardware architecture of an electronic device for implementing a voice recognition system according to the present invention.

Fig. 2 is a flowchart of a preferred embodiment of the speech recognition method of the present invention.

Fig. 3 is a flow chart of a preferred embodiment of the voice recognition method based on auxiliary voice data packets in the present invention.

Fig. 4 is a functional block diagram of the first embodiment of the speech recognition system of the present invention.

Fig. 5 is a functional block diagram of the second embodiment of the speech recognition system of the present invention.

[Description of main component symbols]

【detailed description】

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. Apparently, the described embodiments are only some of the embodiments of the present invention, not all of them. In addition, it should be understood that the specific embodiments described herein are only used to explain the present invention, not to limit the present invention.

As shown in FIG. 1 , it is a schematic diagram of the hardware architecture of a preferred embodiment of an electronic device for implementing a speech recognition system according to the present invention. As shown in the schematic diagram of the hardware architecture, the electronic device 1 includes a speech recognition system 10 . The electronic device 1 further includes a storage unit 20 , a display unit 30 , a processing unit 40 and a voice receiving unit 50 .

Preferably, the speech recognition method of the present invention is implemented by the speech recognition system 10 in the electronic device 1 .

The electronic device 1 includes an electronic device that can automatically perform numerical calculation and/or information processing according to preset or stored instructions, and its hardware includes but is not limited to a microprocessor, an application specific integrated circuit (Application Specific Integrated Circuit, ASIC) ), programmable gate array (Field Programmable Gate Array, FPGA), digital processor (Digital Signal Processor, DSP), embedded devices, etc. The electronic device 1 may also include user equipment. The user equipment includes but is not limited to any electronic product that can interact with the user through keyboards, mice, remote controls, touch pads, or voice-activated devices, such as personal computers, tablet computers, smart phones, personal digital Assistant (Personal Digital Assistant, PDA), game console, interactive Internet TV (Internet Protocol Television, IPTV), smart wearable devices, etc. Wherein, the network where the user equipment is located includes but is not limited to the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (Virtual Private Network, VPN) and the like.

It should be noted that the user equipment described above is only an example, and other existing or future user equipment that can be adapted to the present invention should also be included in the protection scope of the present invention and included here by reference.

In one embodiment, the speech recognition system 10 is configured to acquire the speech information input by the user when the user inputs the speech information, and utilize a large vocabulary speech recognition method based on a preset model (for example, based on a hidden Markov model Large vocabulary speech recognition method) to recognize the input speech information to obtain the first speech recognition result, using the speech recognition method based on auxiliary speech data packets (for example, according to the user's current geographic location information call and the geographic location information corresponding auxiliary voice data packet) to obtain the second voice recognition result. The voice recognition system 10 obtains an optimal recognition result by comparing the first voice recognition result and the second voice recognition result, which not only improves the voice recognition rate, but also improves user experience.

In one embodiment, the storage unit 20 is used to store software programs and data installed in the electronic device 1 , such as the speech recognition system 10 . The storage unit 20 may be an internal storage unit of the electronic device 1 , such as a hard disk or memory of the electronic device 1 . The storage unit 20 can also be an external storage device of the electronic device 1, such as a plug-in hard disk on the electronic device 1, a smart media card (Smart Media Card, SMC), a secure digital card (Secure Digital Card, SD ), flash memory card (flash card) and other storage units. Further, the storage unit 20 may include both an internal storage unit of the electronic device 1 and an external storage device.

In this embodiment, the storage unit 20 pre-stores a plurality of auxiliary voice data packets and voice information corresponding to the plurality of auxiliary voice data packets. The auxiliary voice data package may be a voice data package based on a geographic location, and correspondingly, the storage unit 20 stores voice information having a voice feature of the geographic location.

In this embodiment, the geographical location is divided in units of prefectures and cities. In other embodiments, geographical locations with complex dialects can also be subdivided into areas below prefectures and cities, for example, divided in units of county-level cities or in units of set areas.

Due to the same geographical location, there will be differences in accents and dialects in the Mandarin spoken. Or even if they are not in the same geographic location, the dialect or accent may be the same, therefore, the voice data packets based on the geographic location stored in the storage unit 20 further include voice data packets based on the dialect and geographic location and Voice data packages based on accent and geographic location.

For example, the voice data packets based on dialect and geographic location may include: Cantonese_Hong Kong, Cantonese_Guangzhou, Hokkien_Quanzhou, Hokkien_Xiamen. The voice data packets based on the accent and geographic location may include: Accent_Fujian, Accent_Guangzhou. It should be noted that the voice data packets based on the accent and geographic location include, but are not limited to, the way of pronouncing initials and finals.

In one embodiment, the display unit 30 is used to display a graphical user interface (Graphic User Interface, GUI), which may include a plurality of application program icons and/or a plurality of virtual keys, the application program icons and virtual keys Buttons can represent various functions that the electronic device 1 can provide. For example, the voice input icon represents the function that the electronic device 1 can provide for voice input, and the geographic location selection list button represents that the electronic device 1 can provide geographical selection. The function of location and the button of the text input box represent that the electronic device 1 can provide a function of inputting a geographic location.

The display unit 30 may be, but not limited to, a display unit with touch function such as a touch screen. Therefore, in addition to viewing the application program icons and/or virtual keys displayed on the electronic device 1 through the display unit 30, the user can also input functional instructions through the display unit 30, for example, to run the corresponding application program icon. instructions of the application program, or instructions for activating the virtual key to start the corresponding function.

In one embodiment, the processing unit 40 is one or more central processing units (Central Processing unit, CPU), microprocessors or other digital processing chips. The processing unit 40 is used for executing software program codes or computing data, such as executing the speech recognition system 10 described above. In this embodiment, the processing unit 40 receives the voice information input by the user, and at the same time obtains the user's current geographical location information, and when performing voice recognition, it combines the large vocabulary voice recognition based on the preset model (for example, based on Hidden Mark Large vocabulary speech recognition method of Cove model, or speech recognition method based on artificial neural network model) and speech recognition based on auxiliary speech data packet (for example, speech recognition based on auxiliary speech data packet of geographic location) respectively output the first The recognition result and the second recognition result, according to the selection made by the user comparing the first recognition result and the second recognition result, dynamically adjust the weight of the speech recognition based on the large vocabulary of the preset model and the speech recognition based on the auxiliary voice data packet, to Improve the accuracy of speech recognition.

The processing unit 40 is communicatively connected with the voice recognition system 10 , the storage unit 20 , the display unit 30 and the voice input unit 50 . The communication may be implemented through a serial peripheral device interface bus (Universal Serial Bus, USB) or other communication paths or protocols.

The voice input unit 50 is used to input voice information of the user. The display unit 30 includes, but is not limited to, a microphone.

As shown in FIG. 2 , it is a flowchart of a preferred embodiment of the speech recognition method of the present invention. According to different requirements, the order of the steps in the flowchart can be changed, and some steps can be omitted.

S100. Acquire voice information input by a user.

In this embodiment, the user can directly input voice through the voice receiving unit 50 of the electronic device 1 , and the voice recognition system 10 acquires voice information according to the content of the voice input by the user.

In other embodiments, the display unit 30 of the electronic device 1 provides a graphical user interface, the graphical user interface includes a voice input icon, when the user clicks the voice input icon, the voice recognition system 10 The voice information input by the user is acquired through the voice receiving unit 50 .

S102. Use a first voice recognition method to recognize the voice information to obtain a first recognition result, and use a second voice recognition method to recognize the voice information to obtain a second recognition result.

In this embodiment, the first speech recognition method may be a large vocabulary speech recognition method based on a preset model, and the second speech recognition method may be a speech recognition method based on auxiliary speech data packets. That is, the speech recognition method based on the auxiliary speech data packet is used to assist the speech recognition method based on a preset model with a large vocabulary. The voice recognition method based on auxiliary voice data packets may be a voice recognition method based on auxiliary voice data packets established by geographic location. In some embodiments, the speech recognition system 10 may first execute the first speech recognition method to recognize the speech information, and then execute the second speech recognition method to recognize the second speech information.

In some embodiments, in order to improve recognition efficiency, the speech recognition system 10 may execute the first speech recognition method and the second speech recognition method in parallel to recognize the speech information respectively. When using the preset model-based large vocabulary speech recognition method to recognize the speech information, at the same time utilize the auxiliary speech packet-based speech recognition method to recognize the speech information, that is, the speech recognition system 10 uses the first A thread runs the preset model-based large-vocabulary speech recognition method to recognize the speech information, and a second thread runs the speech recognition based on auxiliary speech data packets to recognize the speech information in parallel.

In this embodiment, the large-vocabulary speech recognition method based on a preset model refers to a speech recognition library established according to standard Mandarin, and any user can call the speech recognition library to perform recognition according to standard Mandarin. Large-vocabulary speech recognition based on preset models does not take dialect and geographic location and/or accent and geographic location into account. The large-vocabulary speech recognition method based on the preset model can adopt the speech recognition method in the prior art, learn and train through multiple pre-established models to recognize the user's speech, and convert the speech information into text information.

The speech recognition method based on auxiliary voice data packets (for convenience of description, hereinafter referred to as "auxiliary speech recognition method") considers the impact of dialect and geographical location and/or accent and geographical location, and needs to establish a geographically-based speech recognition method through training and learning in advance. Voice packets for the location. Please refer to FIG. 3 and corresponding description for the speech recognition method based on geographic location.

S104. Display the first speech recognition result and the second speech recognition result according to a preset rule.

In this embodiment, the preset rule may be that the speech recognition system 10 pre-assigns a first weight to the first speech recognition result, pre-allocates a second weight to the second speech recognition result, and according to the weight The magnitude of the value determines how the speech recognition results for that weight value are displayed. The sum of the first weight value and the second weight value may be a fixed number, for example, an integer 1. Preferably, the first weight value preset by the speech recognition system 10 is greater than the second weight value, that is to say, the speech recognition system 10 assigns a weight value larger to the first speech recognition method than to the second speech recognition method. Weights.

In other embodiments, the preset rule may also be that the speech recognition system 10 presets a first recognition score for the first speech recognition result, and presets a second recognition score for the second speech recognition result. Score, the display mode of the speech recognition result corresponding to the recognition score is determined according to the size of the recognition score. Preferably, the speech recognition system 10 presets a first recognition score value greater than a second recognition score value.

The display manner of the voice recognition result includes, but not limited to: display time and/or display location.

For example, the preset rule of the speech recognition system 10 is to assign weights to the speech recognition results, then when the preset first weight value is greater than the preset second weight value, the display unit 30 of the electronic device 1 Display the first speech recognition result corresponding to a large weight value in the first position, such as the upper part of the user interface provided by the display unit 30; when the preset first weight value is smaller than the preset second weight value , displaying the first speech recognition result corresponding to a smaller weight value in a second position, such as the lower half of the user interface provided by the display unit 30 .

In addition, when the preset first weight value is greater than the preset second weight value, the first speech recognition result is displayed on the display unit 30 of the electronic device 1, after a preset time (for example, after 2 seconds) The second voice recognition result is displayed on the display unit 30 of the electronic device 1 .

In this embodiment, the speech recognition method further includes: updating the preset rules in combination with the acquired user feedback information.

The user feedback information can be obtained according to user operations. For example, if the user selects the first speech recognition result, the user feedback information obtained by the speech recognition system 10 indicates that the best speech recognition result is obtained by using the first speech recognition method. If the user selects the second speech recognition result, the user feedback information obtained by the speech recognition system 10 indicates that the best speech recognition result is obtained by using the second speech recognition method.

The updating of the preset rule may be adjusting a preset weight value or adjusting a preset recognition score value.

Specifically, the speech recognition system 10 increases the weight value or recognition score corresponding to the speech recognition result according to the speech recognition result selected by the user, and/or increases the weight value or recognition score corresponding to the speech recognition result not selected by the user. The score value decreases. For example, when the obtained user feedback information is that the first speech recognition result is selected, the first weight value or the first recognition score value corresponding to the first speech recognition result will be increased, and/or the corresponding second speech recognition result will be The second weight value or the second recognition score value of is decreased. When the obtained user feedback information is that the second speech recognition result is selected, the second weight value or the second recognition score value corresponding to the second speech recognition result will be increased, and/or the second weight value corresponding to the first speech recognition result will be increased. A weight value or first recognition score value decreases.

Wherein, the above-mentioned increase or decrease of the weight value or score value may be performed according to a preset ratio or value.

Please also refer to FIG. 3 , which is a flow chart of a preferred embodiment of the voice recognition method based on auxiliary voice data packets. According to different requirements, the order of the steps in the flowchart can be changed, and some steps can be omitted.

S1020, when receiving the voice information of the user, acquire the current geographic location information of the user.

In this embodiment, the voice recognition system 10 obtains the current geographic location information of the electronic device 1 through the built-in positioning module and/or network connection module of the electronic device 1 . The positioning module includes, but is not limited to: Global Positioning System (Global Positioning System, GPS). The network connection module includes, but is not limited to: 3rd generation mobile communication technology (The 3rd Generation Telecommunication, 3G), general packet radio service (General Packet Radio Service, GPRS) and wireless fidelity technology (wireless fidelity, Wi -Fi). The current geographic location information of the electronic device 1 is considered as the current geographic location information of the user.

In some embodiments, the speech recognition system 10 may also receive an instruction set by the user, and determine the current geographic location information of the user according to the instruction set by the user.

For example, the electronic device 1 is provided with a location selection list, and the location selection list includes names of all cities in China. The user selects geographic location information corresponding to the voice information input by the user by triggering the location selection list.

As another example, the electronic device 1 is provided with a text input box, and the user inputs current geographical location information in a corresponding interface by activating the function of the text input box.

S1022. Invoke a corresponding auxiliary voice data package according to the geographic location information.

In this embodiment, the electronic device 1 invokes the corresponding auxiliary voice data package from the storage unit 20 according to the geographic location information.

The storage unit 20 pre-stores the auxiliary voice data package and the voice information with geographical location voice features included in the auxiliary voice data package.

For example, if the geographic location information is Guangdong, then the voice recognition system 10 invokes an auxiliary voice data package for identifying voice features of Guangdong.

In some embodiments, if the storage unit 20 of the electronic device 1 is not pre-stored with auxiliary voice data packets corresponding to the geographic location information, the voice recognition system 10 will obtain the current geographic location information of the user , downloading the auxiliary voice data package from a server communicatively connected with the electronic device 1 . The communication link may be a wireless communication link. The auxiliary voice data packet is obtained through training and learning by the user in advance and deployed on the server, and the voice recognition system 10 may request the server to send the auxiliary voice data packet corresponding to the geographic location information through the network.

S1024. Recognize the voice information according to the auxiliary voice data packet to obtain a second voice recognition result.

In this embodiment, the speech recognition system 10 uses the second speech recognition method to recognize the speech information to obtain the second speech recognition result.

Further, in order to solve the problem of low speech recognition rate caused by differences in dialects or accents even in the same geographic location, before the speech recognition system 10 invokes the corresponding auxiliary voice data package according to the geographic location information, The S1022 may further include: determining the voice type of the user according to the voice information, and jointly determining a corresponding auxiliary voice data packet based on the voice type and the geographic location information.

The user's voice type is determined by the pronunciation and intonation of the user's language, which can include dialects and accents.

For example, the user's current geographic location is Guangzhou, and the user's voice type may be an accent (for example, Cantonese), then the voice recognition system 10 calls the auxiliary voice data packet of "accent_Guangzhou" to recognize the voice information. In some embodiments, the voice recognition system 10 can also acquire the user's voice type by acquiring information input on an interface provided by the display unit 30 that includes a text input box.

Furthermore, in order to prevent the user from temporarily going on a business trip or traveling in a certain place, the electronic device 1 obtains the user's current geographic location information, and calls the corresponding auxiliary voice data package according to the current geographic location information, resulting in a low recognition rate. , the S1022 may further include: acquiring the user's current geographic location information and historical geographic location information, and determining the called auxiliary voice data package according to the historical geographic location information and the current geographic location information.

In this embodiment, the historical geographic location information refers to the geographic location information of the user's usual residence.

For example, if the user's current geographic location is Guangzhou, but the user's usual place of residence is in Fujian, then the electronic device 1 invokes an auxiliary voice data package for identifying voice features of Fujian to identify the voice information.

To sum up, in the speech recognition method disclosed in the embodiment of the present invention, a plurality of auxiliary speech data packets are obtained through training and learning in advance, and the auxiliary speech data packets are speech databases divided by geographic location. At the same time, based on the voice type of the user, the auxiliary voice data packets are further subdivided into auxiliary voice data packets based on dialect and geographical location, and auxiliary voice data packets based on accent and geographical location. When using the large vocabulary speech recognition method based on the preset model to recognize the user's voice information, the auxiliary voice data packet is also used to recognize the user's voice information to assist the large vocabulary speech recognition method based on the preset model, which not only improves The user's voice recognition rate is improved, and the user experience is also improved.

As shown in FIG. 4 , it is a functional block diagram of the first embodiment of the speech recognition system of the present invention. The speech recognition system 10 includes an acquisition module 100 , a first recognition module 102 , a second recognition module 104 , a display module 106 , a setting module 108 and an update module 110 . The module referred to in the present invention refers to a series of computer program segments that can be executed by the processing unit 40 and can complete fixed functions, which are stored in the storage unit 20 . In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

The acquiring module 100 is configured to acquire voice information input by a user.

In this embodiment, the user can directly input voice through the voice receiving unit 50 of the electronic device 1 , and the acquisition module 100 acquires voice information according to the content of the voice input by the user.

In other embodiments, the display unit 30 of the electronic device 1 provides a graphical user interface, the graphical user interface includes a voice input icon, and when the user clicks the voice input icon, the acquisition module 100 passes The voice receiving unit 50 acquires voice information input by the user.

The first recognition module 102 is configured to recognize the voice information to obtain a first recognition result.

The second recognition module 102 is configured to recognize the voice information to obtain a second recognition result.

In this embodiment, the first recognition module 102 may be a large-vocabulary speech recognition module based on a preset model, and the second recognition module 102 may be a speech recognition module based on auxiliary voice data packets. That is, the speech recognition module based on the auxiliary speech data package is used to assist the speech recognition module based on the large vocabulary of the preset model to perform speech recognition. The voice recognition module based on the auxiliary voice data package may be a voice recognition module based on the auxiliary voice data package established based on the geographic location. In some embodiments, the speech recognition system 10 may first execute the first speech recognition module 102 to recognize the speech information, and then execute the second speech recognition module 102 to recognize the second speech information.

In some embodiments, in order to improve recognition efficiency, the speech recognition system 10 may execute the first speech recognition module 102 and the second speech recognition module 102 in parallel to recognize the speech information respectively. When the speech information is recognized by the large vocabulary speech recognition module based on the preset model, the speech information is recognized by the speech recognition module based on the auxiliary speech data packet at the same time, that is, the speech recognition system 10 runs with the first thread The first recognition module 102 recognizes the voice information, and a second thread runs the second recognition module 102 in parallel to recognize the voice information.

In this embodiment, the large-vocabulary speech recognition module based on the preset model refers to a speech recognition library established according to standard Mandarin, and any user can call the speech recognition library to perform recognition according to standard Mandarin. Large-vocabulary speech recognition based on preset models does not take dialect and geographic location and/or accent and geographic location into account. The large vocabulary speech recognition module based on the preset model is the same as that in the prior art.

The speech recognition module based on auxiliary voice data packets (for convenience of description, hereinafter referred to as "auxiliary speech recognition module") considers the influence of dialect and geographical location and/or accent and geographical location, and needs to establish a geographical basis through training and learning in advance. Voice packets for the location. For the location-based speech recognition module, please refer to FIG. 5 and the corresponding description.

The display module 106 is configured to display the first speech recognition result and the second speech recognition result according to preset rules.

In this embodiment, the preset rule is preset by the setting module 108 . The setting module 108 may pre-allocate a first weight for the first speech recognition result, pre-allocate a second weight for the second speech recognition result, and determine the display of the speech recognition result corresponding to the weight value according to the weight value Way. The sum of the first weight value and the second weight value may be a fixed number, for example, an integer 1. Preferably, the first weight value preset by the setting module 108 is greater than the second weight value, that is to say, the weight value assigned by the setting module 108 to the first speech recognition method is greater than the weight value assigned to the second speech recognition method .

In other embodiments, the preset rule set by the setting module 108 may also be: preset a first recognition score for the first speech recognition result, and preset a second recognition score for the second speech recognition result, according to The size of the recognition score determines the display mode of the speech recognition result corresponding to the recognition score. Preferably, the first recognition score value preset by the setting module 108 is greater than the second recognition score value.

The display manner of the voice recognition result includes, but not limited to: display time and/or display location. But not limited to the displayed time and displayed location.

For example, the preset rule set by the setting module 108 is to assign weights to speech recognition results, then when the preset first weight value is greater than the preset second weight value, the display module 106 can On the display unit 30 of the display unit 30, the first speech recognition result corresponding to a large weight value is displayed in the first position, such as the upper part of the user interface provided by the display unit 30; when the preset first weight value is smaller than the preset first weight value When the weight value is two, the display module 106 displays the first speech recognition result corresponding to a smaller weight value in a second position, such as the lower part of the user interface provided by the display unit 30 .

In addition, when the preset first weight value is greater than the preset second weight value, the display module 106 displays the first voice recognition result on the display unit 30 of the electronic device 1, after a preset time (eg , after 2 seconds) display the second speech recognition result on the display unit 30 of the electronic device 1 .

In this embodiment, the speech recognition system 10 further includes the update module 110, configured to update the preset rules in combination with the acquired user feedback information.

In this embodiment, the user feedback information may be obtained according to user operations. For example, if the user selects the first speech recognition result, the user feedback information obtained by the acquisition module 100 indicates that the best speech recognition result is obtained by using the first speech recognition method. If the user selects the second speech recognition result, the user feedback information obtained by the acquisition module 100 indicates that the best speech recognition result is obtained by using the second speech recognition method.

The updating module 110 may update the preset rule by adjusting a preset weight value or adjusting a preset recognition score value.

Specifically, according to the speech recognition result selected by the user, the update module 110 increases the weight value or recognition score corresponding to the speech recognition result, and/or increases the weight value or recognition score corresponding to the speech recognition result not selected by the user. The value decreases. For example, when the obtained user feedback information is that the first speech recognition result is selected, the update module 110 increases the first weight value or the first recognition score corresponding to the first speech recognition result, and/or increases the corresponding The second weight value or the second recognition score value of the second speech recognition result is decreased. When the obtained user feedback information is that the second speech recognition result is selected, the update module 110 increases the second weight value or the second recognition score corresponding to the second speech recognition result, and/or increases the second weight value corresponding to the first speech recognition result. The first weight value or the first recognition score value of the speech recognition result decreases.

Wherein, the above-mentioned increase or decrease of the weight value or score value may be performed according to a preset ratio or value.

Please also refer to FIG. 5 , which is a functional block diagram of the second embodiment of the speech recognition system of the present invention. Wherein, the second identifying module 104 includes a calling submodule 1040 , a downloading submodule 1042 and a determining submodule 1044 . The module referred to in the present invention refers to a series of computer program segments that can be executed by the processing unit 40 and can complete fixed functions, which are stored in the storage unit 20 . In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

The acquiring module 100 is further configured to acquire the current geographic location information of the user when receiving the user's voice information.

In this embodiment, the obtaining module 100 obtains the current geographic location information of the electronic device 1 through the built-in positioning module and/or network connection module of the electronic device 1 . The positioning module includes, but is not limited to: Global Positioning System (Global Positioning System, GPS). The network connection module includes, but is not limited to: 3rd generation mobile communication technology (The 3rd Generation Telecommunication, 3G), general packet radio service (General Packet Radio Service, GPRS) and wireless fidelity technology (wireless fidelity, Wi -Fi). The current geographic location information of the electronic device 1 is considered as the current geographic location information of the user.

In some embodiments, the acquiring module 100 may also receive an instruction set by the user, and determine the current geographic location information of the user according to the instruction set by the user.

For example, the electronic device 1 is provided with a location selection list, and the location selection list includes names of all cities in China. The user selects geographic location information corresponding to the voice information input by the user by triggering the location selection list.

As another example, the electronic device 1 is provided with a text input box, and the user inputs current geographical location information in a corresponding interface by activating the function of the text input box.

The calling sub-module 1040 is configured to call the corresponding auxiliary voice data package according to the geographic location information.

In this embodiment, the calling submodule 1040 calls the corresponding auxiliary voice data package from the storage unit 20 according to the geographic location information.

The storage unit 20 pre-stores the auxiliary voice data package and the voice information with geographical location voice features included in the auxiliary voice data package.

For example, if the geographic location information is Guangdong, then the invoking submodule 1040 invokes the auxiliary voice data package for identifying the voice characteristics of Cantonese.

In some embodiments, if the storage unit 20 of the electronic device 1 does not pre-store the auxiliary voice data package corresponding to the geographic location information, when the acquisition module 100 acquires the current geographic location information of the user, Execute the download submodule 102. The downloading sub-module 1042 downloads the auxiliary voice data package from a server communicated with the electronic device 1 . The communication link may be a wireless communication link. The auxiliary voice data package is obtained through training and learning by the user in advance and deployed on the server, and the download sub-module 1042 may request the server to send the auxiliary voice data package corresponding to the geographic location information through the network.

The second recognition module 104 is configured to recognize the voice information according to the auxiliary voice data packet to obtain a second voice recognition result.

In this embodiment, the second recognition module 104 uses the second speech recognition method to recognize the speech information to obtain the second speech recognition result.

Further, in order to solve the problem of low speech recognition rate caused by differences in dialects or accents even in the same geographical location, the second recognition module 104 may also include a determination submodule 1044: used to information to determine the user's voice type. The invoking sub-module 1040 jointly determines the corresponding auxiliary voice data package based on the voice type and the geographic location information.

The user's voice type is determined by the pronunciation and intonation of the user's language, which can include dialects and accents.

For example, the user's current geographic location is Guangzhou, and the user's voice type is accent (for example, Cantonese), then the calling sub-module 1040 calls the auxiliary voice data package of "accent_Guangzhou" to identify the voice information.

In some embodiments, the obtaining module 100 may also obtain the voice type of the user by obtaining information input on an interface provided by the display unit 30 that includes a text input box.

Furthermore, in order to prevent the user from temporarily going to a certain place for business or tourism, the acquisition module 100 acquires the user's current geographic location information, and the calling sub-module 1040 invokes the corresponding auxiliary voice data package according to the current geographic location information When the recognition rate is low, the acquiring module 100 is also used to acquire the user's current geographic location information and historical geographic location information, and the calling sub-module 1040 determines the auxiliary voice data package to call according to the historical geographic location information and the current geographic location information .

In this embodiment, the historical geographic location information refers to the geographic location information of the user's usual residence.

For example, if the user's current geographic location is Guangzhou, but the user's usual residence is in Fujian, then the invoking sub-module 1040 calls the auxiliary voice data package for identifying the voice characteristics of Fujian to identify the voice information.

To sum up, in the speech recognition system disclosed in the embodiment of the present invention, a plurality of auxiliary speech data packets are obtained through training and learning in advance, and the auxiliary speech data packets are speech databases divided by geographic location. At the same time, based on the voice type of the user, the auxiliary voice data packets are further subdivided into auxiliary voice data packets based on dialect and geographical location, and auxiliary voice data packets based on accent and geographical location. When using the large vocabulary speech recognition module based on the preset model to recognize the user's voice information, the auxiliary voice data packet is also used to identify the user's voice information to assist the large vocabulary speech recognition method based on the preset model, not only Improve the user's voice recognition rate, but also improve the user experience.

In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, or each unit may physically exist separately, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware, or in the form of hardware plus software function modules.

The above-mentioned integrated units implemented in the form of software function modules can be stored in a computer-readable storage medium. The above-mentioned software function modules are stored in a storage medium, and include several instructions to make a computer unit (which may be a personal computer, server, or network unit, etc.) or a processor (processor) execute the methods described in various embodiments of the present invention. partial steps.

It will be apparent to those skilled in the art that the invention is not limited to the details of the above-described exemplary embodiments, but that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Accordingly, the embodiments should be regarded in all points of view as exemplary and not restrictive, the scope of the invention being defined by the appended claims rather than the foregoing description, and it is therefore intended that the scope of the invention be defined by the appended claims rather than by the foregoing description. All changes within the meaning and range of equivalents of the elements are embraced in the present invention. Any reference sign in a claim should not be construed as limiting the claim concerned. In addition, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of units or devices stated in the system claims may also be realized by one unit or device through software or hardware. The words first, second, etc. are used to denote names and do not imply any particular order.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention without limitation. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be Modifications or equivalent replacements can be made without departing from the spirit and scope of the technical solutions of the present invention.

Claims (16)

1. A speech recognition method, applied in electronic equipment, is characterized in that, the method comprises: Obtain the voice information input by the user; Using a first voice recognition method to recognize the voice information to obtain a first voice recognition result, using a second voice recognition method to recognize the voice information to obtain a second voice recognition result; and The first speech recognition result and the second speech recognition result are displayed according to preset rules. 2. The speech recognition method according to claim 1, wherein the first speech recognition method is a large vocabulary speech recognition method based on a preset model, and the second speech recognition method is based on auxiliary voice data packets voice recognition method. 3. the voice recognition method as claimed in claim 2, is characterized in that, the voice recognition method based on auxiliary voice data packet comprises: When receiving the voice information, obtain the user's current geographic location information; Invoking the corresponding auxiliary voice data package according to the geographic location information; and Recognizing the voice information according to the auxiliary voice data packet to obtain the second voice recognition result. 4. speech recognition method as claimed in claim 3, is characterized in that, described method also comprises: A plurality of voice data packets based on geographic location are preset, and the voice data packets are stored in the electronic device or in a server connected to the electronic device. 5. The speech recognition method according to claim 3, wherein, before invoking the corresponding auxiliary voice data package according to the geographic location information, the method further comprises: determining the voice type of the user according to the voice information, the voice type including accent and dialect; and A corresponding auxiliary voice data packet is jointly determined based on the voice type and the geographic location information. 6. The speech recognition method according to any one of claims 3 to 5, wherein the method further comprises: When receiving the voice information, obtain the user's current geographic location information and historical geographic location information; and The auxiliary voice data package to be called is determined according to the historical geographic location information and the current geographic location information. 7. speech recognition method as claimed in claim 6, is characterized in that, described method also comprises: The preset rules are updated in combination with the obtained user feedback information, and the preset rules include: Pre-assigning a first weight to the first speech recognition result, pre-allocating a second weight to the second speech recognition result, and determining the display mode of the speech recognition result corresponding to the weight value according to the weight value; or Presetting a first recognition score for the first speech recognition result, presetting a second recognition score for the second speech recognition result, determining the display mode of the speech recognition result corresponding to the recognition score according to the size of the recognition score, Wherein, the display manner includes display time or display position. 8. The speech recognition method according to claim 7, wherein said updating said preset rule comprises: According to the speech recognition result selected by the user, the weight value or recognition score corresponding to the speech recognition result is increased, and/or the weight value or recognition score corresponding to the speech recognition result not selected by the user is decreased. 9. A speech recognition system running in an electronic device, characterized in that the system comprises: An acquisition module, configured to acquire voice information input by the user; A first recognition module, configured to recognize the voice information to obtain a first voice recognition result; A second recognition module, configured to recognize the voice information to obtain a second voice recognition result; and A display module, configured to display the first speech recognition result and the second speech recognition result according to preset rules. 10. The speech recognition system according to claim 9, wherein the first speech recognition module is a large vocabulary speech recognition module based on a preset model, and the second speech recognition module is based on auxiliary voice data packets speech recognition module. 11. The speech recognition system as claimed in claim 10, wherein, The acquiring module is further configured to acquire the current geographic location information of the user when the voice information is received; The second identification module includes: Calling a submodule, used to call the corresponding auxiliary voice data package according to the geographic location information; and The second recognition module is configured to recognize the voice information according to the auxiliary voice data packet to obtain the second voice recognition result. 12. speech recognition system as claimed in claim 11, is characterized in that, described system also comprises: The setting module is used to pre-set a plurality of voice data packages based on geographic location, and store the voice data packages in the electronic device or in a server connected to the electronic device. 13. speech recognition system as claimed in claim 11, is characterized in that, described system also comprises determining submodule: for determining the voice type of the user according to the voice information, the voice type including accent and dialect; and A corresponding auxiliary voice data packet is jointly determined based on the voice type and the geographic location information. 14. The speech recognition system according to any one of claims 11 to 13, wherein: The acquiring module is further configured to acquire the user's current geographic location information and historical geographic location information when receiving the voice information; and The invoking sub-module is also used to determine the auxiliary voice data package to be invoked according to historical geographic location information and current geographic location information. 15. speech recognition system as claimed in claim 14, is characterized in that, described system also comprises: An update module, configured to update the preset rules in combination with the acquired user feedback information, the preset rules are set by the setting module, including: Pre-assigning a first weight to the first speech recognition result, pre-allocating a second weight to the second speech recognition result, and determining the display mode of the speech recognition result corresponding to the weight value according to the weight value; or Presetting a first recognition score for the first speech recognition result, presetting a second recognition score for the second speech recognition result, determining the display mode of the speech recognition result corresponding to the recognition score according to the size of the recognition score, Wherein, the display manner includes display time or display position. 16. speech recognition system as claimed in claim 15 is characterized in that, described updating module updates described preset rule and comprises: According to the speech recognition result selected by the user, the weight value or recognition score corresponding to the speech recognition result is increased, and/or the weight value or recognition score corresponding to the speech recognition result not selected by the user is decreased.