CN114582333A - Voice recognition method and device, electronic equipment and storage medium - Google Patents

Abstract

The embodiment of the invention discloses a voice recognition method, a voice recognition device, electronic equipment and a storage medium, wherein the method comprises the following steps: receiving a first voice stream input by a user and used for voice interaction, and performing semantic recognition on the first voice stream; when the first voice stream is detected to be stopped, determining the information integrity state of the first voice stream according to the semantic recognition result, and determining the pause duration of the first voice stream; and determining an interactive response strategy corresponding to the first voice stream according to the information integrity state and the pause duration. According to the technical scheme of the embodiment of the invention, different processing strategies can be adopted according to the integrity of the voice stream information before pause, so that the accuracy of semantic recognition is effectively improved, the requirements of a user are quickly responded, and the user experience is improved.

Description

Speech recognition method, device, electronic device and storage medium

technical field

Embodiments of the present invention relate to the technical field of speech recognition, and in particular, to a speech recognition method, apparatus, electronic device, and storage medium.

Background technique

At present, with the development of network technology and mobile communication technology, the application of electronic products closely related to people's daily life is becoming more and more common, for example, smart phones, car voice systems and so on. When using electronic products, there are various limitations whether handwriting input or keyboard input. For the convenience of use, the user is often required to input speech, and after the device recognizes the speech, it outputs the text content of the speech or executes corresponding operation instructions.

However, the recognition accuracy of voice information in the prior art is not high, and the problem of recognition errors often occurs, which brings inconvenience to users, low processing efficiency, and poor user experience.

SUMMARY OF THE INVENTION

Embodiments of the present invention provide a speech recognition method, apparatus, electronic device, and storage medium, so as to determine a corresponding response strategy according to semantic analysis.

In a first aspect, an embodiment of the present invention provides a speech recognition method, the method includes:

receiving a first voice stream input by a user for performing voice interaction, and performing semantic recognition on the first voice stream;

When it is detected that the first voice stream is paused, the information integrity state of the first voice stream is determined according to the result of semantic recognition, and the pause duration of the first voice stream is determined;

An interaction response strategy corresponding to the first voice stream is determined according to the information integrity state and the pause duration.

In a second aspect, an embodiment of the present invention further provides a voice recognition device, the device comprising:

a semantic recognition module, configured to receive a first voice stream input by a user for performing voice interaction, and perform semantic recognition on the first voice stream;

a state determination module, configured to determine the information integrity state of the first voice stream according to the result of semantic recognition when it is detected that the first voice stream is paused, and determine the pause duration of the first voice stream;

A response determination module, configured to determine an interactive response strategy corresponding to the first voice stream according to the information integrity state and the pause duration.

In a third aspect, an embodiment of the present invention further provides an electronic device, the electronic device comprising:

one or more processors;

storage means for storing one or more programs,

When the one or more programs are executed by the one or more processors, the one or more processors implement the speech recognition method provided by any embodiment of the present invention.

In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements the speech recognition method provided by any embodiment of the present invention.

According to the technical solution of the embodiment of the present invention, by receiving the voice stream information input by the user and performing semantic recognition on it, when a pause occurs in the voice stream information input by the user is detected, the information integrity status of the voice stream and The pause duration, and further, the corresponding interactive response strategy can be determined according to the complete state of the information and the pause duration, which solves the problem in the prior art that the pause in the voice information input by the user affects the device's understanding of the user's semantic intention, and realizes the Different response strategies are determined according to the complete state of the voice information and the pause time, which effectively improves the accuracy of semantic understanding and the speed of response to user needs, and improves user experience.

Description of drawings

In order to illustrate the technical solutions of the exemplary embodiments of the present invention more clearly, the following briefly introduces the accompanying drawings used in describing the embodiments. Obviously, the introduced drawings are only a part of the drawings of the embodiments to be described in the present invention, rather than all drawings. For those of ordinary skill in the art, without creative work, they can also obtain the drawings according to these drawings. Additional drawings.

1 is a schematic flowchart of a speech recognition method according to Embodiment 1 of the present invention;

2 is a schematic flowchart of a speech recognition method according to Embodiment 2 of the present invention;

3 is a schematic flowchart of a speech recognition method in an application scenario provided by Embodiment 3 of the present invention;

FIG. 4 is a schematic structural diagram of a speech recognition device according to Embodiment 4 of the present invention;

FIG. 5 is a schematic structural diagram of an electronic device according to Embodiment 5 of the present invention.

Detailed ways

The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention. In addition, it should be noted that, for the convenience of description, the drawings only show some but not all structures related to the present invention.

In addition, it should be noted that, for the convenience of description, the drawings only show some but not all of the contents related to the present invention. Before discussing the exemplary embodiments in greater detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although a flowchart depicts various operations (or steps) as a sequential process, many of the operations may be performed in parallel, concurrently, or concurrently. Additionally, the order of operations can be rearranged. The process may be terminated when its operation is complete, but may also have additional steps not included in the figures. The processes may correspond to methods, functions, procedures, subroutines, subroutines, and the like.

Example 1

FIG. 1 is a schematic flowchart of a speech recognition method provided in Embodiment 1 of the present invention. This embodiment is applicable to a situation where a user interacts with a device by speech. The method can be executed by a speech recognition device, and the device can use software and/or hardware, and may be configured in a terminal and/or a server to implement the speech recognition method in this embodiment of the present invention.

As shown in FIG. 1 , the method of this embodiment may specifically include:

S110. Receive a first voice stream input by a user for performing voice interaction, and perform semantic recognition on the first voice stream.

The first voice stream may be understood as a piece of voice information. The first voice stream can be used for voice interaction between the user and the smart terminal device. Typically, the smart terminal may be a mobile phone, a tablet computer, a vehicle-mounted computer, or a personal computer, which is not limited in this embodiment. Exemplarily, the first voice stream may include, but is not limited to. For example, the first voice stream may be "turn on the air conditioner", "turn on", or "turn up the temperature", and so on. Semantic recognition can be understood as the recognition and conversion of lexical content in the user's speech into input that can be recognized by the computer.

In a specific implementation, when the user is about to perform voice interaction with the terminal device, a piece of voice information will be input to the terminal device, and after the terminal device receives the first voice stream input by the user, it will recognize the acquired first voice stream, thereby Convert the lexical content in the voice stream input by the user into a language that the terminal can recognize.

It should be noted that when the terminal device receives the first voice stream input by the user and performs semantic recognition on it, it may be that when the user inputs the first voice stream information, the terminal device performs semantic recognition on the voice information in real time. When a pause occurs in the first voice stream, the recognition ends, and subsequent operations continue to be performed; it may also be that when a pause occurs in the first voice stream information input by the user is detected, the semantics of the first voice stream information before the pause can be started. Identification, etc., are not limited in this embodiment.

S120. When it is detected that the first voice stream is paused, determine the information integrity state of the first voice stream according to the result of semantic recognition, and determine the pause duration of the first voice stream.

The information integrity state can be understood as whether the information in the first speech stream has complete semantic intent. Alternatively, the information complete status may include information complete and incomplete. The pause duration may be understood as the time interval from the occurrence of the pause to the end of the pause in the first voice stream.

In a specific implementation, the first voice stream input by the user is received, and semantic recognition is performed on it, and when a pause is detected in the first voice stream, the first voice stream is determined according to the semantic recognition result of the first voice stream. Whether the contained information is complete, and determine the pause duration in the first voice stream.

Optionally, determining the information integrity state of the first voice stream according to the result of the semantic recognition, including: according to the matching degree between the identified semantic information of the first voice stream and each response function in the preset function whitelist , and determine the information integrity status of the first voice stream.

The pre-set function whitelist may be understood as a pre-stored list of existing functions in the terminal device. Specifically, the advantages of setting the function whitelist may be that the voice information input by the user can be quickly screened, and the functions in the function whitelist can be quickly responded, which can greatly improve the response speed and user experience. The response function can be understood as various functions that the terminal device can perform according to the semantic recognition result. For example, the response function can be "turn on the air conditioner", "play music", or "turn the temperature up to 26 degrees", etc. It should be noted that each response function in the function whitelist is not fixed, and different response functions may be set according to different terminal devices, which is not limited in this embodiment.

Specifically, the matching degree between the semantic information of the first voice stream and each response function in the preset function whitelist can be used for description by comparing the semantic information of the first voice stream with various functions included in the function whitelist. The information of each response function is determined. Exemplarily, the function whitelist may include, but is not limited to, at least one item of information such as vertical information, action information, and slot information. Among them, the vertical information can be used to reflect the type of the response function, for example, it can be an air conditioner, a music, or a navigation system. The action information can be used to reflect the action to be performed by the response function, for example, it can be open, close, or adjust. The slot information can be used to reflect the key parameter information in the response function. For example, "26 degrees" in "Turn up the air conditioner to 26 degrees" is the slot information.

In a specific implementation, by comparing the semantic information in the first voice stream with the information used to describe each response function contained in the preset function whitelist, it is determined that the first voice stream input by the user and each response in the function whitelist The degree of matching between the functions can further determine whether the information in the first voice stream input by the user is complete.

It should be noted that if there is at least one match between the information in the first voice stream and the information in the function whitelist, the first voice stream can be considered to be in a complete information state; the information in the first voice stream and the information in the function whitelist If they do not match, it can be considered that the first voice stream is in a state of incomplete information.

S130. Determine an interaction response strategy corresponding to the first voice stream according to the information integrity state and the pause duration.

The interactive response strategy may be understood as a feedback strategy determined by the terminal device according to the information input by the user. Specifically, the terminal device may have different interactive response strategies according to the complete information state of the first voice stream and the pause duration.

It should be noted that the corresponding interactive response strategy can be determined according to the complete information status and the pause duration. Different interactive responses can be determined by the information complete status being complete and incomplete, and the pause duration being the same pause duration and different pause durations. Strategy.

Optionally, when the information integrity state is that the information is complete or the information is incomplete, the same pause duration is used, and the corresponding interaction response strategy is determined according to the information integrity state and the same pause duration.

Specifically, when the user inputs the first voice stream information, the terminal device performs semantic recognition on it, regardless of whether the semantic intent of the first voice stream information is complete, the terminal device continues to wait according to the same pause duration, and judges that within the same pause duration Whether the voice stream information newly input by the user is received, the corresponding interactive response strategy is determined according to the judgment result.

Optionally, according to the complete state of the information, different pause durations are used to determine the corresponding interactive response strategy. The first pause duration may be set when the information is complete, and the second pause duration may be set when the information is complete.

Specifically, when the user inputs the first voice stream information, the terminal device performs semantic recognition on it to determine the complete state of the first voice stream information, and when the complete state of the information is complete information, determines the corresponding interaction response according to the first pause duration Strategy. Exemplarily, the interactive response strategy corresponding to the first pause duration may include, but is not limited to, directly executing the response function corresponding to the first voice stream, and combining all the voice stream information input by the user according to the voice stream information subsequently input by the user and executing them together. The response function corresponding to the total voice flow information, etc., is not limited in this embodiment. When the information integrity state is that the information is incomplete, a corresponding interaction response strategy is determined according to the second pause duration. Exemplarily, the interactive response strategy corresponding to the second pause duration may include, but is not limited to, sending prompt information to the user to remind the user whether the input has not been completed and whether to continue inputting; The information infers related or similar response functions, and is pushed to the user in the form of a list for the user to select, etc., which is not limited in this embodiment.

It should be noted that this embodiment does not limit the lengths and specific values of the first pause duration and the second pause duration. For example, the first pause duration corresponds to when the information of the first voice stream is complete, and corresponds to when the information is incomplete. The second pause duration can be set to be shorter than the second pause duration. The advantage of this setting is that it can ensure a quick response to the first voice information when the semantics are complete, and at the same time, the considered voice input appears when the To solve the temporary pause problem, prolong the pause time to determine whether the voice input is over, improve the intelligence of the response method, and improve the user experience.

It should also be noted that the "first" and "second" mentioned in this embodiment do not indicate any order, quantity or importance, but only that the corresponding target subjects are different.

According to the technical solution of the embodiment of the present invention, by receiving the voice stream information input by the user and performing semantic recognition on it, when a pause occurs in the voice stream information input by the user is detected, the information integrity status of the voice stream and The pause duration, and further, the corresponding interactive response strategy can be determined according to the complete state of the information and the pause duration, which solves the problem in the prior art that the pause in the voice information input by the user affects the device's understanding of the user's semantic intention, and realizes the Different response strategies are determined according to the complete state of the voice information and the pause time, which effectively improves the accuracy of semantic understanding and the speed of response to user needs, and improves user experience.

Embodiment 2

FIG. 2 is a schematic flowchart of a speech recognition method according to Embodiment 2 of the present invention. On the basis of the above technical solution, this embodiment further refines the technical solution. This embodiment is based on any optional technical solution in the embodiment of the present invention, optionally, the interactive response strategy corresponding to the first voice stream is determined according to the information integrity state and the pause duration, Including: if the semantics of the voice stream is complete, determining an interaction response strategy corresponding to the voice interaction information according to the pause duration and the first duration threshold; if the semantics of the voice stream is incomplete, according to The pause duration and the second duration threshold determine an interaction response strategy corresponding to the voice interaction information; wherein the first duration threshold is less than or equal to the second duration threshold.

Wherein, the technical terms that are the same as or corresponding to the above-mentioned embodiments will not be repeated here. As shown in Figure 2, the method of this embodiment specifically includes the following steps:

S210. Receive a first voice stream input by a user for performing voice interaction, and perform semantic recognition on the first voice stream.

S220. When it is detected that the first voice stream is paused, determine the information integrity state of the first voice stream according to the result of semantic recognition, and determine the pause duration of the first voice stream.

S230. If the semantics of the speech stream is complete, determine an interaction response strategy corresponding to the first speech stream according to the pause duration and the first duration threshold.

Among them, the semantics of the speech stream is complete and can be understood as the speech stream has complete semantic intent. The first duration threshold may be preset, for example, the first duration threshold may be 0.4 seconds, 0.6 seconds, or 1 second.

Specifically, when a pause is detected in the first speech stream, and it is determined according to the semantic recognition result that the semantics of the first speech stream is complete, it is necessary to compare the pause duration with the preset first duration threshold, so that it can be determined that Different interactive response strategies.

Optionally, if the semantics of the voice stream is complete, it is judged whether a newly input second voice stream is received within the first duration threshold; Interactive function. If so, continue to perform semantic recognition on the second speech stream.

In a specific implementation, if it is determined according to the semantic recognition result that the semantics of the first speech stream is complete, the pause duration can be compared with the first duration threshold to determine whether the second speech input by the user is received within the first duration threshold. flow, according to the judgment result, execute the corresponding interactive response strategy, specifically, if the terminal device does not receive other voice information input by the user within the first duration threshold, it can respond to the semantic recognition result of the first voice flow, Execute the interactive response function corresponding to the semantic intent contained in the first voice stream; accordingly, if the terminal device receives the second voice stream information newly input by the user within the first duration threshold, it can be considered that the received second voice stream The stream information is valid speech information, and the semantic recognition of the second speech stream information is continued to determine the result of the semantic recognition.

In order to clearly introduce the specific implementation of this step, the terminal device is used as a vehicle-mounted voice interaction system to describe it. For example, if the user pauses after inputting "turn on the air conditioner", the terminal device responds to the voice information of "turn on the air conditioner" before the pause. Perform semantic recognition, and the result obtained from the analysis is that the semantic intention of "turn on the air conditioner" is complete. Further, continue to wait according to the first duration threshold. If the system does not receive any more voice information input by the user within the first duration threshold, then After the first duration threshold, the device stops receiving voice information and semantic recognition, and directly executes the response function corresponding to "turn on the air conditioner", for example, the system replies "the air conditioner has been turned on".

S240. If the semantics of the speech stream is incomplete, determine an interaction response strategy corresponding to the first speech stream according to the pause duration and the second duration threshold.

Among them, the incomplete semantics of the speech stream can be understood as the received speech stream information does not have complete semantic intent. Similarly, the second duration threshold may also be preset, for example, the second duration threshold may be 3 seconds, 4 seconds, or 5 seconds.

Specifically, when a pause in the first speech stream is detected, and it is determined that the semantics of the first speech stream is incomplete according to the semantic recognition result, the pause duration needs to be compared with the second duration threshold, and different interactive responses are determined according to the comparison result. Strategy.

Optionally, if the semantics of the speech stream is not complete, then determine whether a newly input second speech stream is received within the second duration threshold; if so, continue to perform semantic recognition on the second speech stream; Set timeout response policy.

Among them, the timeout response strategy can be understood as the response method and strategy when the user does not continue to input within a certain period of time, and the previously input information does not have complete semantic intentions. Exemplarily, the timeout response policy may be to send a prompt message to the user, prompting the user that the input voice stream is incomplete and whether to continue inputting, and may also provide the user with a list of possible related functions so that the user can choose from them, or It may be to keep the device in the current state, without a response, etc., which is not limited in this embodiment. Wherein, the judgment basis of the related function list may include, but is not limited to, semantic inference based on the input voice information or searching for the functions most frequently used by the user from historical behavior data.

In a specific implementation, if it is determined according to the semantic recognition result that the first speech stream input by the user does not have complete semantic intent, it is necessary to compare the pause duration with the second duration threshold to determine whether the user continues to be received within the second duration threshold. Input second voice stream. If the second voice stream information newly input by the user within the second duration threshold is received, continue to perform semantic recognition on the second voice stream, and if no voice stream information newly input by the user is received, a timeout response policy needs to be implemented.

In order to clearly introduce the specific implementation of this step, the terminal device is used as the vehicle-mounted voice interaction system to describe it. The "open" before the pause is semantically recognized, and the result of the recognition is that the semantic intent of "open" is incomplete, and then, continue to press the time of the second duration threshold (for example, the second duration threshold can be 5 seconds), if the user Continue to input the word "air conditioner" within 5 seconds, the system will perform semantic recognition and analysis on the words "turn on the air conditioner", determine the correct semantic intent, and execute the work of turning on the air conditioner. If the user does not continue to input within 5 seconds For subsequent voice messages, the timeout response strategy will be executed after 5 seconds (for example, the system will ask "what do you want to open?", etc.).

It should be noted that, in this embodiment, optionally, the first duration threshold may be less than or equal to the second duration threshold.

It should also be noted that the "first" and "second" mentioned in this embodiment do not indicate any order, quantity or importance, but only that the corresponding target subjects are different.

According to the technical solution of the embodiment of the present invention, by receiving the first voice stream information input by the user and performing semantic recognition on it, when a pause occurs in the voice stream information input by the user is detected, the first voice stream information is determined according to the semantic recognition result. The complete state and pause duration of the system, and then, according to the complete information and incomplete information, different duration thresholds are set, and the corresponding interactive response strategy is further determined, which solves the problem in the prior art due to the setting of the pause duration threshold is too long and the response speed. Slow affects the user experience, or, if the pause duration threshold is set too short, the user will not have time to continue to input subsequent voice information after unintentional pause, resulting in the terminal failing to recognize the user's complete semantic intent and unable to execute the command correctly, which affects the user experience even more. It realizes that when a pause in user input is detected, different pause duration thresholds can be set according to the semantic integrity state before the pause, and then the corresponding interactive response strategy can be determined, so that it can respond to user needs more quickly and improve user input. Beneficial effect of completeness of user intent recognition when pauses occur.

Embodiment 3

To make the technical solutions of the embodiments of the present invention clear to those skilled in the art, a specific application scenario example is given below by taking the terminal device as the vehicle-mounted voice system.

FIG. 3 is a schematic flowchart of a speech recognition method in an application scenario provided by Embodiment 3 of the present invention; the embodiment of the present invention is a preferred embodiment of the above-mentioned embodiments of the present invention. Referring to FIG. 3 , the method of the embodiment of the present invention is shown in FIG. Including the following steps;

Step 1: The user inputs the first voice stream;

Step 2: Perform speech recognition (Automatic Speech Recognition, ASR) on the first speech stream input by the user;

The third step: perform semantic understanding (Natural Language Understanding, NLU) on the first voice stream input by the user;

Step 4: Detect the pause in the first voice stream input by the user through the Voice Activity Detection (VAD) technology;

Step 5: when it is detected that the first voice stream pauses, determine whether the information state of the first voice stream is complete according to the semantic recognition result, and determine the pause duration;

Step 6: If the information integrity status of the first voice stream is complete, continue to wait for the time t1, and determine whether there is a second voice stream newly input by the user within the time t1;

Step 7: If there is a second voice stream that the user continues to input within the time t1, continue to perform processes such as voice recognition and semantic understanding on the second voice stream;

Step 8: If there is no voice information newly input by the user within the time t1, execute the response function corresponding to the first voice stream information before the pause;

Step 9: If the information integrity status of the first voice stream is incomplete, continue to wait for the time t2, and determine whether there is a second voice stream newly input by the user within the time t2;

Step 10: If the user has a second voice stream that the user continues to input within the time t2, continue to perform processes such as voice recognition and semantic understanding on the second voice stream;

Step 11: If there is no second voice stream for new input within t2 time, execute a timeout response policy.

According to the technical solution of the embodiment of the present invention, by receiving the first voice stream information input by the user and performing semantic recognition on it, when a pause occurs in the voice stream information input by the user is detected, the first voice stream information is determined according to the semantic recognition result. The complete state and pause duration of the system, and then, according to the complete information and incomplete information, different duration thresholds are set, and the corresponding interactive response strategy is further determined, which solves the problem in the prior art due to the setting of the pause duration threshold is too long and the response speed. Slow affects the user experience, or, if the pause duration threshold is set too short, the user will not have time to continue to input subsequent voice information after unintentional pause, resulting in the terminal failing to recognize the user's complete semantic intent and unable to execute the command correctly, which affects the user experience even more. It realizes that when a pause in user input is detected, different pause duration thresholds can be set according to the semantic integrity state before the pause, and then the corresponding interactive response strategy can be determined, so that it can respond to user needs more quickly and improve user input. Beneficial effect of completeness of user intent recognition when pauses occur.

Embodiment 4

FIG. 4 is a schematic structural diagram of a speech recognition apparatus provided in Embodiment 4 of the present invention. The speech recognition apparatus provided in this embodiment may be implemented by software and/or hardware, and may be configured in a terminal and/or a server to implement The speech recognition method in the embodiment of the present invention. The apparatus may specifically include: a semantic recognition module 410 , a state determination module 420 and a response policy determination module 430 .

Wherein, the semantic recognition module 410 is configured to receive a first voice stream input by a user for performing voice interaction, and perform semantic recognition on the first voice stream;

The state determination module 420 is used to determine the information integrity state of the first voice stream according to the result of semantic recognition when it is detected that the first voice stream pauses, and determine the pause duration of the first voice stream;

The response strategy determination module 430 is configured to determine an interaction response strategy corresponding to the first voice stream according to the complete state of the information and the pause duration.

According to the technical solution of the embodiment of the present invention, by receiving the voice stream information input by the user and performing semantic recognition on it, when a pause occurs in the voice stream information input by the user is detected, the information integrity status of the voice stream and The pause duration, and further, the corresponding interactive response strategy can be determined according to the complete state of the information and the pause duration, which solves the problem in the prior art that the pause in the voice information input by the user affects the device's understanding of the user's semantic intention, and realizes the Different response strategies are determined according to the complete state of the voice information and the pause time, which effectively improves the accuracy of semantic understanding and the speed of response to user needs, and improves user experience.

Optionally, the information complete state includes complete information and incomplete information.

Optionally, the response strategy determination module 430 includes a first response strategy determination unit and a second response strategy determination unit.

Wherein, the first response strategy determination unit is used to determine the interactive response strategy corresponding to the first speech stream according to the pause duration and the first duration threshold if the semantics of the speech stream is complete; the second response strategy determination unit is used to determine if If the semantics of the speech stream is not complete, the interaction response strategy corresponding to the first speech stream is determined according to the pause duration and the second duration threshold, wherein the first duration threshold is less than or equal to the second duration threshold.

Optionally, the first response strategy determining unit is further configured to, if the semantics of the speech stream is complete, determine whether a newly input second speech stream is received within the first duration threshold; if not, respond to the first speech stream, and execute the interactive function corresponding to the first voice stream.

Optionally, the first response strategy determination unit is further configured to, if the semantics of the voice stream is complete, determine whether a newly input second voice stream is received within the first duration threshold; if so, continue to process the second voice stream. perform semantic recognition.

Optionally, the second response strategy determining unit is further configured to, if the semantics of the voice stream is complete, determine whether a newly input second voice stream is received within the second duration threshold; if so, continue to process the second voice stream. perform semantic recognition.

Optionally, the second response strategy determination unit is further configured to, if the semantics of the voice stream is complete, determine whether a newly input second voice stream is received within the second duration threshold; if not, execute a preset timeout response strategy.

Optionally, the state determination module 420 is further configured to determine that the information of the first voice stream is complete according to the matching degree between the identified semantic information of the first voice stream and each response function in the preset function whitelist. state.

The above voice recognition apparatus can execute the voice recognition method provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects for executing the voice recognition method.

Embodiment 5

FIG. 5 is a schematic structural diagram of an electronic device according to Embodiment 5 of the present invention. Figure 5 shows a block diagram of an exemplary electronic device 50 suitable for use in implementing embodiments of the present invention. The electronic device 50 shown in FIG. 5 is only an example, and should not impose any limitation on the function and scope of use of the embodiments of the present invention.

As shown in FIG. 5, electronic device 50 takes the form of a general-purpose computing device. Components of electronic device 50 may include, but are not limited to, one or more processors or processing units 501, system memory 502, and a bus 503 connecting different system components (including system memory 502 and processing unit 501).

Bus 503 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of a variety of bus structures. By way of example, these architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect ( PCI) bus.

Electronic device 50 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by electronic device 50, including volatile and non-volatile media, removable and non-removable media.

System memory 502 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 504 and/or cache memory 505 . Electronic device 50 may further include other removable/non-removable, volatile/non-volatile computer system storage media. For example only, storage system 506 may be used to read and write to non-removable, non-volatile magnetic media (not shown in FIG. 5, commonly referred to as a "hard drive"). Although not shown in Figure 5, a disk drive may be provided for reading and writing to removable non-volatile magnetic disks (eg "floppy disks"), as well as removable non-volatile optical disks (eg CD-ROM, DVD-ROM) or other optical media) to read and write optical drives. In these cases, each drive may be connected to bus 503 through one or more data media interfaces. Memory 502 may include at least one program product having a set (eg, at least one) of program modules configured to perform the functions of various embodiments of the present invention.

A program/utility 508 having a set (at least one) of program modules 507, which may be stored, for example, in memory 502, such program modules 507 including, but not limited to, an operating system, one or more application programs, other program modules, and program data , each or some combination of these examples may include an implementation of a network environment. Program modules 507 generally perform the functions and/or methods of the described embodiments of the present invention.

The electronic device 50 may also communicate with one or more external devices 509 (eg, keyboards, pointing devices, display 510, etc.), with one or more devices that enable a user to interact with the electronic device 50, and/or with Any device (eg, network card, modem, etc.) that enables the electronic device 50 to communicate with one or more other computing devices. Such communication may take place through input/output (I/O) interface 511 . Also, the electronic device 50 may communicate with one or more networks (eg, a local area network (LAN), a wide area network (WAN), and/or a public network such as the Internet) through a network adapter 512 . As shown, network adapter 512 communicates with other modules of electronic device 50 via bus 503 . It should be understood that, although not shown in FIG. 5, other hardware and/or software modules may be used in conjunction with electronic device 50, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tapes drives and data backup storage systems, etc.

The processing unit 501 executes various functional applications and data processing by running the programs stored in the system memory 502, for example, to implement the speech recognition method provided by the embodiment of the present invention.

Embodiment 6

Embodiment 6 of the present invention also provides a storage medium containing computer-executable instructions, where the computer-executable instructions are used to execute a speech recognition method when executed by a computer processor, and the method includes:

receiving a first voice stream input by a user for performing voice interaction, and performing semantic recognition on the first voice stream;

When it is detected that the first voice stream is paused, the information integrity state of the first voice stream is determined according to the result of semantic recognition, and the pause duration of the first voice stream is determined;

An interaction response strategy corresponding to the first voice stream is determined according to the information integrity state and the pause duration.

The computer storage medium of the embodiments of the present invention may adopt any combination of one or more computer-readable mediums. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or a combination of any of the above. More specific examples (a non-exhaustive list) of computer readable storage media include: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), Erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

A computer-readable signal medium may include a propagated data signal in baseband or as part of a carrier wave with computer-readable program code embodied thereon. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device .

Program code embodied on a computer readable medium may be transmitted using any suitable medium, including - but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for carrying out the operations of embodiments of the present invention may be written in one or more programming languages, including object-oriented programming languages—such as Java, Smalltalk, C++, and A conventional procedural programming language - such as the "C" language or similar programming language. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider through Internet connection).

Note that the above are only preferred embodiments of the present invention and applied technical principles. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the protection scope of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and can also include more other equivalent embodiments without departing from the concept of the present invention. The scope is determined by the scope of the appended claims.

Claims (10)

1. A speech recognition method, comprising:

receiving a first voice stream input by a user and used for voice interaction, and performing semantic recognition on the first voice stream;

when the first voice stream is detected to be stopped, determining the information integrity state of the first voice stream according to the semantic recognition result, and determining the pause duration of the first voice stream;

and determining an interactive response strategy corresponding to the first voice stream according to the information integrity state and the pause duration.

2. The method of claim 1, wherein the information integrity status comprises information is complete and information is not complete;

the determining the interactive response strategy corresponding to the first voice stream according to the information integrity state and the pause duration includes:

if the semantics of the voice stream are complete, determining an interactive response strategy corresponding to the first voice stream according to the pause duration and the first time threshold;

if the semantics of the voice stream are not complete, determining an interactive response strategy corresponding to the first voice stream according to the pause duration and the second duration threshold;

wherein the first duration threshold is less than or equal to a second duration threshold.

3. The method according to claim 2, wherein determining an interactive response policy corresponding to the first voice stream according to the pause duration and the first time threshold if the semantics of the voice stream are already complete comprises:

if the semanteme of the voice stream is complete, judging whether a second newly input voice stream is received within a first time length threshold value;

and if not, responding to the first voice stream, and executing an interactive function corresponding to the first voice stream.

4. The method of claim 3, further comprising:

and if so, continuing to perform semantic recognition on the second voice stream.

5. The method according to claim 2, wherein determining the interactive response policy corresponding to the first voice stream according to the pause duration and the second duration threshold if the semantics of the voice stream are incomplete comprises:

if the semantics of the voice stream are not complete, judging whether a newly input second voice stream is received within a second duration threshold;

and if so, continuing to perform semantic recognition on the second voice stream.

6. The method of claim 4, further comprising:

if not, executing a preset timeout response strategy.

7. The method according to claim 1, wherein the determining the information integrity status of the first speech stream according to the result of semantic recognition comprises:

and determining the information integrity state of the first voice stream according to the matching degree between the identified semantic information of the first voice stream and each response function in a preset function white list.

8. A speech recognition apparatus, comprising:

the semantic recognition module is used for receiving a first voice stream which is input by a user and used for voice interaction and performing semantic recognition on the first voice stream;

the state determining module is used for determining the information integrity state of the first voice stream according to the semantic recognition result and determining the pause duration of the first voice stream when the first voice stream is detected to be stopped;

and the response determining module is used for determining an interactive response strategy corresponding to the first voice stream according to the information integrity state and the pause duration.

9. An electronic device, characterized in that the electronic device comprises:

one or more processors;

a storage device for storing one or more programs,

when executed by the one or more processors, cause the one or more processors to implement a speech recognition method as recited in any of claims 1-7.

10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the speech recognition method according to any one of claims 1 to 7.

Images