Method, device, equipment and medium for determining speech semantics based on wake-up word speed

Abstract

The present application relates to a method, device, electronic device and storage medium for determining speech semantics based on the speed of a wake-up word, the method comprising: determining that the device is in a waiting state to be awakened, obtaining the wake-up word, and judging whether to wake up the device according to the wake-up word; when the judgment result is that the device is awakened by the wake-up word, determining the total duration of the wake-up word; determining the back-end duration of the current speech endpoint detection according to the total duration of the wake-up word; obtaining the speech, setting the cloud back-end duration equal to the current speech endpoint duration; sending the audio corresponding to the speech to the cloud, and the cloud determines the semantics corresponding to the speech according to the cloud back-end duration. This application dynamically adjusts the cloud segmentation node according to the speech speed of the wake-up word based on the user's habit of speaking at almost the same speed when reciting the wake-up word and the specific voice command content, thereby achieving the effect of dynamically determining the speech endpoint, determining where to segment according to the speech endpoint, further determining the semantics, and improving the accuracy.

Description

Technical Field

[0001] The present application relates to the field of speech semantics technology, and in particular to a method, device, equipment and medium for determining speech semantics based on wake-up word speed. Background Art

[0002] Although the problem solved by patent WO2020024885A1 is similar to that of this patent application, they mainly analyze the speech speed and dynamically set the sentence segmentation nodes through the previous voice command. The biggest drawback of this solution is that in actual application scenarios, it cannot be guaranteed that the previous voice command and the current voice command are issued by the same user, and the two sentences spoken by the same person may not have the same speech speed. Therefore, the speech speed of the previous sentence cannot be directly used as the basis for judging the current voice interaction.

[0003] In the voice interaction scenario, each user has different speaking habits, and even the same user has different speaking speeds under different emotions. Therefore, the traditional method of setting a pause time threshold, such as determining whether the pause length in a long speech is greater than a preset threshold, can segment the speech if the pause length is greater than the preset threshold. The effect of segmenting the speech using the above method is not good. Since the speaking speed of a person is not fixed when speaking, using whether the pause length is greater than the preset threshold as the basis for segmentation may lead to frequent segmentation or long-term uninterrupted sentence problems, thereby reducing the accuracy of segmentation. Summary of the invention

[0004] Based on the above problems, the present application provides a method, apparatus, device and medium for determining speech semantics based on wake-up word speed.

[0005] In a first aspect, an embodiment of the present application provides a method for determining speech semantics based on a wake-up word speed, including:

[0006] Determine that the device is in a waiting state, obtain a wake-up word, and determine whether to wake up the device based on the wake-up word;

[0007] When the result of the determination is that the device is awakened by the wake-up word, determining the total duration of the wake-up word;

[0008] Determine the back endpoint duration of the current voice endpoint detection according to the total duration of the wake-up word;

[0009] Get the voice and set the cloud back-end detection duration to be equal to the back-end detection duration of the current voice endpoint detection;

[0010] The audio corresponding to the speech is sent to the cloud, and the cloud determines the semantics corresponding to the speech according to the detection duration of the back-end point of the cloud.

[0011] Furthermore, in the above method for determining speech semantics based on the wake-up word speed, before determining that the device is in the waiting-to-wake-up state, the method further includes:

[0012] Start the application corresponding to the device and initialize the voice-related engine.

[0013] Furthermore, in the above method for determining speech semantics based on wake-up word speed, after the cloud determines the semantics corresponding to the speech according to the cloud back-end detection duration, it also includes:

[0014] Get the semantics of speech from the cloud.

[0015] Furthermore, in the above method for determining speech semantics based on the wake-up word speed, determining whether to wake up the device according to the wake-up word includes:

[0016] Determine whether the wake-up word is consistent with the preset wake-up word, and when the determination result is that the wake-up word is consistent with the preset wake-up word, determine to wake up the device;

[0017] When the judgment result is that the wake-up word is inconsistent with the preset wake-up word, the device will not be woken up.

[0018] Furthermore, in the above method for determining speech semantics based on the wake-up word speed, determining the total duration of the wake-up word includes:

[0019] Determine the start time and end time of the wake-up word according to the voice wake-up engine;

[0020] The total duration of the wake-up word is determined according to the start time of the wake-up word and the end time of the wake-up word.

[0021] Furthermore, in the above method for determining speech semantics based on the wake-up word speed, determining the rear endpoint duration of the current speech endpoint detection according to the total duration of the wake-up word includes:

[0022] Obtain the wake-up record of the wake-up word, and determine the average duration of the wake-up word and the back endpoint duration of the speech endpoint detection corresponding to the average duration according to the wake-up record;

[0023] The back endpoint duration of the current voice endpoint detection is determined based on the average duration of the wake-up word, the back endpoint duration of the voice endpoint detection corresponding to the average duration, and the total duration of the wake-up word.

[0024] Furthermore, in the above method for determining speech semantics based on the wake-up word speed, the cloud determines the speech semantics according to the detection duration of the cloud back-end endpoint, including:

[0025] After determining the endpoint of the voice according to the endpoint detection duration in the cloud;

[0026] The semantics of speech is determined based on automatic speech recognition technology and natural language understanding technology.

[0027] In a second aspect, an embodiment of the present application further provides a device for determining speech semantics based on a wake-up word speed, including:

[0028] The first determination module is used to determine whether the device is in a waiting state, obtain a wake-up word, and determine whether to wake up the device according to the wake-up word;

[0029] The second determination module is used to determine the total duration of the wake-up word when the judgment result is that the device is awakened by the wake-up word;

[0030] The third determination module is used to determine the back endpoint duration of the current voice endpoint detection according to the total duration of the wake-up word;

[0031] Acquisition module and setting module: used to acquire voice and set the cloud back-end detection duration to be equal to the back-end detection duration of the current voice endpoint detection;

[0032] The sending module and the fourth determining module are used to send the audio corresponding to the speech to the cloud, and the cloud determines the semantics corresponding to the speech according to the detection duration of the cloud back-end.

[0033] In a third aspect, an embodiment of the present application further provides an electronic device, including: a processor and a memory;

[0034] The processor is used to execute any one of the above methods for determining speech semantics based on wake-up word speed by calling the program or instruction stored in the memory.

[0035] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program or instruction, wherein the program or instruction enables a computer to execute any one of the above-mentioned methods for determining speech semantics based on the wake-up word speed.

[0036] The advantages of the embodiments of the present application are: the present application relates to a method, device, electronic device and storage medium for determining speech semantics based on the speed of a wake-up word, the method comprising: determining that the device is in a waiting state to be awakened, obtaining a wake-up word, and judging whether to wake up the device according to the wake-up word; when the judgment result is that the device is awakened by the wake-up word, determining the total duration of the wake-up word; determining the back-end duration of the current speech endpoint detection according to the total duration of the wake-up word; obtaining the speech, setting the cloud back-end duration equal to the back-end duration of the current speech endpoint detection; sending the audio corresponding to the speech to the cloud, and the cloud determines the semantics corresponding to the speech according to the cloud back-end duration. The present application dynamically adjusts the cloud segmentation node according to the speech speed of the wake-up word based on the user's habit of speaking at almost the same speed when reciting the wake-up word and the specific voice instruction content, thereby achieving the effect of dynamically determining the speech endpoint, determining where to segment according to the speech endpoint, further determining the semantics, and improving the accuracy of segmentation. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the conventional technology, the drawings required for use in the embodiments or the conventional technology descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0038] Figure 1 A method for determining speech semantics based on wake-up words provided in an embodiment of the present application is shown in FIG. Figure 1 ;

[0039] Figure 2 A method for determining speech semantics based on wake-up words provided in an embodiment of the present application is shown in FIG. Figure 2 ;

[0040] Figure 3 A method for determining speech semantics based on wake-up words provided in an embodiment of the present application is shown in FIG. Figure 3 ;

[0041] Figure 4 A schematic diagram of a device for determining speech semantics based on wake-up word speed provided in an embodiment of the present application;

[0042] Figure 5 It is a schematic block diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present application, so the present application is not limited by the specific implementation disclosed below.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0045] Figure 1 A method for determining speech semantics based on wake-up words provided in an embodiment of the present application is shown in FIG. Figure 1 .

[0046] In the first aspect, the embodiment of the present application provides a method for determining speech semantics based on the wake-up word speed, combining Figure 1 ,include:

[0047] S101: Determine that the device is in a waiting-to-wake-up state, obtain a wake-up word, and determine whether to wake up the device according to the wake-up word.

[0048] Specifically, in the embodiment of the present application, the device can be any smart terminal device, such as Tmall Genie. The wake-up state can be that the smart terminal device is turned on. When the corresponding wake-up word is received, the smart terminal device can be awakened. The wake-up word can be set by the user according to his or her preferences, or it can be set by the manufacturer, such as "Xiaodu Xiaodu". Whether to wake up the device is determined based on the wake-up word. For example, if the voice wake-up word received from the user is "Xiaodu Xiaodu", whether it is a smart terminal device is determined based on the wake-up word "Xiaodu Xiaodu".

[0049] S102: When the result of the determination is that the device is awakened by the wake-up word, determine the total duration of the wake-up word.

[0050] Specifically, in an embodiment of the present application, when the judgment result is that the smart terminal device is awakened by the wake-up word "Xiaodu Xiaodu", the total duration of the wake-up word is determined. Exemplarily, the total duration of the wake-up word from the beginning of the word "Xiao" to the end of the word "du". The following article introduces the total duration of the wake-up word in combination with specific steps.

[0051] S103: Determine the back endpoint duration of the current voice endpoint detection according to the total duration of the wake-up word.

[0052] Specifically, in an embodiment of the present application, for example, the back endpoint duration of "degree" is determined based on the total duration of the wake-up word from the beginning of the word "small" to the end of the word "degree". The following article introduces the specific steps for determining the back endpoint duration of the current voice endpoint detection.

[0053] S104: Acquire the voice and set the cloud back-end point detection duration to be equal to the back-end point duration of the current voice endpoint detection.

[0054] Specifically, in an embodiment of the present application, after the smart terminal device is awakened by the wake-up word, the smart terminal device obtains the voice spoken by the user, such as "Beijing weather forecast", "Open the curtains", "Play Faye Wong's song Legend", etc. After obtaining the voice spoken by the user, the cloud back-end point detection time is set equal to the back-end point time of the current voice endpoint detection.

[0055] It should be understood that users have almost the same speaking speed habits when reading the wake-up word and the specific voice command content. Therefore, the cloud segmentation node is dynamically adjusted using the speaking speed of the wake-up word, which effectively solves the problem of inconsistent speaking speeds of different users.

[0056] S105: The audio corresponding to the speech is sent to the cloud, and the cloud determines the semantics corresponding to the speech according to the detection duration of the back-end point of the cloud.

[0057] Specifically, in an embodiment of the present application, the audio corresponding to speech such as "Beijing weather forecast", "Open the curtains", and "Play Faye Wong's song Legend" is sent to the cloud. The cloud first determines the node of the sentence segmentation, and then determines the semantics corresponding to the speech based on the speech after the sentence segmentation.

[0058] Furthermore, in the above method for determining speech semantics based on the wake-up word speed, before determining that the device is in the waiting-to-wake-up state, the method further includes:

[0059] Start the application corresponding to the device and initialize the voice-related engine.

[0060] Specifically, in the embodiment of the present application, before the smart terminal device is turned on, it also initializes the voice-related engine by starting the application corresponding to the smart terminal device, thereby preparing for waking up the device through the wake-up word.

[0061] Furthermore, in the above method for determining speech semantics based on wake-up word speed, after the cloud determines the semantics corresponding to the speech according to the cloud back-end detection duration, it also includes:

[0062] Get the semantics of speech from the cloud.

[0063] Specifically, in the embodiment of the present application, the cloud determines the node of the sentence segmentation, and then determines the semantics corresponding to the speech based on the speech after the sentence segmentation. The smart terminal device can directly obtain the semantics corresponding to the speech from the cloud.

[0064] Furthermore, in the above method for determining speech semantics based on the wake-up word speed, determining whether to wake up the device according to the wake-up word includes:

[0065] Determine whether the wake-up word is consistent with the preset wake-up word, and when the determination result is that the wake-up word is consistent with the preset wake-up word, determine to wake up the device.

[0066] When the judgment result is that the wake-up word is inconsistent with the preset wake-up word, the device will not be woken up.

[0067] Specifically, in an embodiment of the present application, if the wake-up word "Xiaodu Xiaodu" is used to determine whether it is a smart terminal device, if the preset wake-up word is also "Xiaodu Xiaodu", the smart terminal device will be woken up; if the preset wake-up word is "Hello Xiaodi", the smart terminal device will not be woken up.

[0068] Figure 2 A method for determining speech semantics based on wake-up words provided in an embodiment of the present application is shown in FIG. Figure 2 .

[0069] Furthermore, in the above method for determining speech semantics based on the wake-up word speed, the total duration of the wake-up word is determined, combined with Figure 2 , including two steps S201 and S202:

[0070] S201: Determine the start time and the end time of the wake-up word according to the voice wake-up engine.

[0071] S202: Determine the total duration of the wake-up word according to the start time and the end time of the wake-up word.

[0072] Specifically, in an embodiment of the present application, after the wake-up word "Xiaodu Xiaodu" wakes up the smart terminal device, the voice wake-up engine determines the start time of "Xiaodu Xiaodu" and the end time of "Xiaodu Xiaodu", and further determines the total duration of the wake-up word "Xiaodu Xiaodu" according to the start time of "Xiaodu Xiaodu" and the end time of "Xiaodu Xiaodu".

[0073] Figure 3 A method for determining speech semantics based on wake-up words provided in an embodiment of the present application is shown in FIG. Figure 3 .

[0074] Furthermore, in the above method for determining speech semantics based on the wake-up word speed, determining the rear endpoint duration of the current speech endpoint detection according to the total duration of the wake-up word includes:

[0075] Get the wake-up record of the wake-up word, and determine the average duration of the wake-up word and the back endpoint duration of the speech endpoint detection corresponding to the average duration based on the wake-up record.

[0076] The back endpoint duration of the current voice endpoint detection is determined based on the average duration of the wake-up word, the back endpoint duration of the voice endpoint detection corresponding to the average duration, and the total duration of the wake-up word.

[0077] Specifically, in an embodiment of the present application, the wake-up record may be a historical record of the corresponding smart terminal device being awakened multiple times by the wake-up word. The average duration of the wake-up word and the back endpoint duration of the voice endpoint detection corresponding to the average duration are determined based on the historical record of being awakened multiple times by the wake-up word. For example, the average duration of the wake-up word "Xiaodu Xiaodu" is equal to 1S, and the back endpoint duration of the voice endpoint detection corresponding to the average duration is 0.3S. The total duration of the current wake-up word "Xiaodu Xiaodu" is 0.8S. The back endpoint duration of the current voice endpoint detection is calculated based on the average duration of the wake-up word, the back endpoint duration of the voice endpoint detection corresponding to the average duration, and the total duration of the wake-up word.

[0078] Furthermore, in the above method for determining speech semantics based on wake-up word speed, the cloud determines the speech semantics according to the detection time of the cloud back-end endpoint, combined with Figure 3 , including two steps S301 and S302:

[0079] S301: Determine the endpoint of the voice according to the cloud endpoint detection duration.

[0080] S302: Determine the semantics of the speech based on automatic speech recognition technology and natural language understanding technology.

[0081] Specifically, in the embodiments of the present application, dynamically adjusting the cloud segmentation node according to the wake-up word speed effectively solves the problem of inconsistent speaking speeds of different users, realizes dynamic segmentation, and determines the semantics of the voice based on the automatic speech recognition technology ASR and the natural language understanding technology NLU. The smart terminal device can directly obtain the semantics corresponding to the voice from the cloud and make corresponding processing, such as executing actions such as "opening the curtains" and "playing Faye Wong's song Legend".

[0082] Figure 4 A schematic diagram of a device for determining speech semantics based on a wake-up word speed provided in an embodiment of the present application.

[0083] Second, combining Figure 4 The embodiment of the present application also provides a device for determining speech semantics based on the wake-up word speed, including:

[0084] The first determination module 401 is used to determine whether the device is in a waiting-to-wake-up state, obtain a wake-up word, and determine whether to wake up the device according to the wake-up word.

[0085] Specifically, in the embodiment of the present application, the device can be any smart terminal device, such as Tmall Genie. The first determination module 401 determines that the device is in a waiting state. The waiting state can be to determine that the smart terminal device is turned on. When the corresponding wake-up word is received, the smart terminal device can be woken up. The wake-up word can be set by the user according to his or her own preferences, or it can be set by the manufacturer, such as "Xiaodu Xiaodu". Whether to wake up the device is determined based on the wake-up word. For example, if the voice wake-up word received from the user is "Xiaodu Xiaodu", whether it is a smart terminal device is determined based on the wake-up word "Xiaodu Xiaodu".

[0086] The second determination module 402 is used to determine the total duration of the wake-up word when the judgment result is that the device is awakened by the wake-up word.

[0087] Specifically, in an embodiment of the present application, when the judgment result is that the smart terminal device is awakened by the wake-up word "Xiaodu Xiaodu", the second determination module 402 determines the total duration of the wake-up word. Exemplarily, the total duration of the wake-up word from the beginning of the word "Xiao" to the end of the word "du". The following article introduces the total duration of the wake-up word in combination with specific steps.

[0088] The third determination module 403 is used to determine the rear endpoint duration of the current voice endpoint detection according to the total duration of the wake-up word.

[0089] Specifically, in an embodiment of the present application, exemplarily, the third determination module 403 determines the back endpoint duration of "degree" based on the total duration of the wake-up word from the beginning of the word "small" to the end of the word "degree". The following is an introduction to determining the back endpoint duration of the current voice endpoint detection in conjunction with specific steps.

[0090] The acquisition module 404 and the setting module 405 are used to acquire voice and set the cloud end-point detection duration to be equal to the end-point duration of the current voice end-point detection.

[0091] Specifically, in an embodiment of the present application, after the wake-up word wakes up the smart terminal device, the acquisition module 404 of the smart terminal device acquires the voice spoken by the user, such as "Beijing weather forecast", "Open the curtains", "Play Faye Wong's song Legend", etc. After acquiring the voice spoken by the user, the setting module 405 sets the cloud back-end point detection duration equal to the back-end point duration of the current voice endpoint detection.

[0092] It should be understood that users have almost the same speaking speed habits when reading the wake-up word and the specific voice command content. Therefore, the cloud segmentation node is dynamically adjusted using the speaking speed of the wake-up word, which effectively solves the problem of inconsistent speaking speeds of different users.

[0093] The sending module 406 and the fourth determining module 407 are used to send the audio corresponding to the speech to the cloud, and the cloud determines the semantics corresponding to the speech according to the detection duration of the cloud back-end.

[0094] Specifically, in the embodiment of the present application, the sending module 406 sends the audio corresponding to the speech such as "Beijing weather forecast", "Open the curtains", "Play Faye Wong's song Legend" to the cloud, and the fourth determination module 407 in the cloud first determines the node of the sentence segmentation, and then determines the semantics corresponding to the speech based on the speech after the sentence segmentation.

[0095] In a third aspect, an embodiment of the present application further provides an electronic device, including: a processor and a memory;

[0096] The processor is used to execute any one of the above methods for determining speech semantics based on wake-up word speed by calling the program or instruction stored in the memory.

[0097] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program or instruction, wherein the program or instruction enables a computer to execute any one of the above-mentioned methods for determining speech semantics based on the wake-up word speed.

[0098] Figure 5 It is a schematic block diagram of an electronic device provided by an embodiment of the present disclosure.

[0099] like Figure 5As shown, the electronic device includes: at least one processor 501, at least one memory 502 and at least one communication interface 503. The various components in the electronic device are coupled together through a bus system 504. The communication interface 503 is used to transmit information between external devices. It can be understood that the bus system 504 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 504 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, Figure 5 Various buses are labeled as bus system 504 .

[0100] It can be understood that the memory 502 in this embodiment can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories.

[0101] In some implementations, the memory 502 stores the following elements, executable units or data structures, or a subset thereof, or an extended set thereof: an operating system and application programs.

[0102] Among them, the operating system includes various system programs, such as the framework layer, the core library layer, the driver layer, etc., which are used to implement various basic services and process hardware-based tasks. The application program includes various application programs, such as a media player (Media Player), a browser (Browser), etc., which are used to implement various application services. A program that implements any method of the method for determining speech semantics based on the wake-up word speed provided in an embodiment of the present application can be included in the application program.

[0103] In an embodiment of the present application, the processor 501 calls a program or instruction stored in the memory 502, specifically, a program or instruction stored in an application, and the processor 501 is used to execute the steps of each embodiment of a method for determining speech semantics based on wake-up word speed provided in an embodiment of the present application.

[0104] Determine that the device is in a waiting state, obtain a wake-up word, and determine whether to wake up the device based on the wake-up word;

[0105] When the result of the determination is that the device is awakened by the wake-up word, determining the total duration of the wake-up word;

[0106] Determine the back endpoint duration of the current voice endpoint detection according to the total duration of the wake-up word;

[0107] Get the voice and set the cloud back-end detection duration to be equal to the back-end detection duration of the current voice endpoint detection;

[0108] The audio corresponding to the speech is sent to the cloud, and the cloud determines the semantics corresponding to the speech according to the detection duration of the back-end point of the cloud.

[0109] Any of the methods for determining speech semantics based on wake-up word speed provided in the embodiments of the present application can be applied to the processor 501, or implemented by the processor 501. The processor 501 can be an integrated circuit chip with signal processing capabilities. In the implementation process, the steps of the above method can be completed by hardware integrated logic circuits in the processor 501 or instructions in software form. The above-mentioned processor 501 can be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0110] The steps of any method in the method for quickly determining speech semantics based on a wake-up word provided in an embodiment of the present application can be directly embodied as a hardware decoding processor, or a combination of hardware and software units in the decoding processor. The software unit can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, or other mature storage media in the art. The storage medium is located in the memory 502, and the processor 501 reads the information in the memory 502, and combines its hardware to complete the steps of a method for quickly determining speech semantics based on a wake-up word.

[0111] Those skilled in the art will appreciate that although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features from different embodiments is meant to be within the scope of the present application and to form different embodiments.

[0112] Those skilled in the art will appreciate that the description of each embodiment has its own emphasis, and for parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0113] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for determining speech semantics based on wake-up words, characterized in that: include: Determine that the device is in a waiting-to-wake-up state, obtain a wake-up word, and determine whether to wake up the device according to the wake-up word; When the judgment result is that the device is awakened by the wake-up word, determining the total duration of the wake-up word; Obtain a wake-up record of the wake-up word, and determine an average duration of the wake-up word and a rear endpoint duration of speech endpoint detection corresponding to the average duration according to the wake-up record; Determine the back endpoint duration of the current voice endpoint detection according to the average duration of the wake-up word, the back endpoint duration of the voice endpoint detection corresponding to the average duration, and the total duration of the wake-up word; obtain the voice, and set the cloud back endpoint detection duration to be equal to the back endpoint duration of the current voice endpoint detection; The audio corresponding to the speech is sent to the cloud, and the cloud determines the semantics corresponding to the speech according to the detection duration of the back-end point of the cloud.

2. The method for determining speech semantics based on wake-up word speed according to claim 1, characterized in that: Before determining that the device is in the wake-up state, the method further includes: The application corresponding to the device is started to initialize the voice-related engine.

3. The method for determining speech semantics based on wake-up word speed according to claim 1, characterized in that: After the cloud determines the semantics corresponding to the speech according to the detection duration of the cloud back-end, the method further includes: The semantics of the speech is obtained from the cloud.

4. The method for determining speech semantics based on wake-up word speed according to claim 1, characterized in that: The determining whether to wake up the device according to the wake-up word includes: Determine whether the wake-up word is consistent with a preset wake-up word, and when the determination result is that the wake-up word is consistent with the preset wake-up word, determine to wake up the device; When the judgment result is that the wake-up word is inconsistent with the preset wake-up word, the device is not woken up.

5. The method for determining speech semantics based on wake-up word speed according to claim 1, characterized in that: The total duration of determining the wake-up word includes: Determine the start time and end time of the wake-up word according to the voice wake-up engine; The total duration of the wake-up word is determined according to the start time of the wake-up word and the end time of the wake-up word.

6. The method for determining speech semantics based on wake-up word speed according to claim 1, characterized in that: The cloud determines the semantics of the speech according to the detection duration of the cloud back-end, including: After determining the endpoint of the voice according to the cloud-based endpoint detection duration; The semantics of the speech is determined based on automatic speech recognition technology and natural language understanding technology.

7. A device for determining speech semantics based on wake-up word speed, characterized in that: include: The first determination module is used to determine whether the device is in a waiting state, obtain a wake-up word, and determine whether to wake up the device according to the wake-up word; A second determination module: configured to determine the total duration of the wake-up word when the judgment result is that the device is awakened by the wake-up word; The third determination module is used to obtain the wake-up record of the wake-up word, and determine the average duration of the wake-up word and the back endpoint duration of the voice endpoint detection corresponding to the average duration according to the wake-up record; Determine the back endpoint duration of the current voice endpoint detection according to the average duration of the wake-up word, the back endpoint duration of the voice endpoint detection corresponding to the average duration, and the total duration of the wake-up word; an acquisition module and a setting module: used to acquire voice and set the cloud back endpoint detection duration to be equal to the back endpoint duration of the current voice endpoint detection; The sending module and the fourth determining module are used to send the audio corresponding to the voice to the cloud, and the cloud determines the semantics corresponding to the voice according to the detection duration of the back-end point of the cloud.

8. An electronic device, characterized in that: include: Processor and memory; The processor is used to execute the method for determining speech semantics based on wake-up word speed as described in any one of claims 1 to 6 by calling the program or instruction stored in the memory.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program or instruction, and the program or instruction enables a computer to execute a method for determining speech semantics based on wake-up word speed as described in any one of claims 1 to 6.

Images