Wake-up method and device, electronic equipment and storage medium

Abstract

The invention provides an awakening method and device, electronic equipment and a storage medium, and relates to the technical field of computers, in particular to the technical field of artificial intelligence such as voice interaction and deep learning. According to the specific implementation scheme, the method comprises the following steps: storing acquired audio data into a preset cache region, and inputting the acquired audio data into a preset wake-up model to obtain a first recognition result output by the wake-up model; under the condition that the first recognition result comprises the wake-up word, calling a voice recognition service to recognize the audio data to obtain a second recognition result; and determining whether to wake up the system or not according to the matching degree of the second recognition result and the wake-up word.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, in particular to the field of artificial intelligence technology such as voice interaction and deep learning, and specifically to a wake-up method, device, electronic device and storage medium. Background Art

[0002] With breakthroughs in natural language processing, speech recognition, and speech synthesis, in-vehicle voice systems have been able to achieve more intelligent and humanized interactions. These technologies provide strong support for in-vehicle voice systems, enabling them to accurately recognize users' voice commands, understand users' intentions, and respond accordingly. In-vehicle voice systems have gradually become an important competitive point in the automotive market, allowing users to control various functions of the car through voice commands and improve driving convenience. Summary of the invention

[0003] The present disclosure aims to solve one of the technical problems in the related art at least to some extent.

[0004] To this end, the purpose of the present disclosure is to propose a wake-up method, device, electronic device and storage medium, which can verify the results of the wake-up model to avoid the problem of false wake-up of the wake-up model during voice interaction, and upload the audio data to the server in an automated manner for the algorithm to locate the problem and subsequent model optimization.

[0005] According to a first aspect of the present disclosure, a wake-up method is provided, comprising:

[0006] The acquired audio data is stored in a preset buffer area, and input into a preset wake-up model to obtain a first recognition result output by the wake-up model;

[0007] In a case where the first recognition result includes a wake-up word, calling a speech recognition service to recognize the audio data to obtain a second recognition result;

[0008] Determine whether to wake up the system based on the degree of match between the second recognition result and the wake-up word.

[0009] According to a second aspect of the present disclosure, a wake-up device is provided, including:

[0010] A first recognition module, used to store the acquired audio data into a preset buffer area, and input it into a preset wake-up model to obtain a first recognition result output by the wake-up model;

[0011] A second recognition module, configured to, when the first recognition result includes a wake-up word, call a speech recognition service to recognize the audio data to obtain a second recognition result;

[0012] A determination module is used to determine whether to wake up the system according to the matching degree between the second recognition result and the wake-up word.

[0013] According to a third aspect of the present disclosure, there is provided an electronic device, including:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the wake-up method as described in the first aspect.

[0017] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the wake-up method as described in the first aspect.

[0018] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising computer instructions, which, when executed by a processor, implement the steps of the wake-up method as described in the first aspect.

[0019] The wake-up method, device, electronic device, and storage medium provided by the present disclosure have the following beneficial effects:

[0020] First, the acquired audio data is stored in a preset buffer area and input into a preset wake-up model to obtain a first recognition result output by the wake-up model. Then, when the wake-up word is included in the first recognition result, the speech recognition service is called to recognize the audio data to obtain a second recognition result. Afterwards, based on the matching degree between the second recognition result and the wake-up word, determine whether to wake up the system. Therefore, by further verifying whether to wake up the system based on the recognition result of the audio data by the speech recognition service and the matching between the wake-up word recognized by the wake-up model, the accuracy and reliability of the voice system wake-up can be improved, the false wake-up rate can be reduced, and the user experience can be optimized.

[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The above and / or additional aspects and advantages of the present disclosure will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, which are used to better understand the present solution and do not constitute a limitation of the present disclosure, wherein:

[0023] Figure 1is a flowchart of a wake-up method proposed according to an embodiment of the present disclosure;

[0024] Figure 2 is a flowchart of a wake-up method proposed according to another embodiment of the present disclosure;

[0025] Figure 3 is a flowchart of a wake-up method proposed according to another embodiment of the present disclosure;

[0026] Figure 4 is an operation diagram of a wake-up method provided by an embodiment of the present disclosure;

[0027] Figure 5 is a schematic diagram of the structure of a wake-up device proposed according to an embodiment of the present disclosure;

[0028] Figure 6 A block diagram of an exemplary electronic device suitable for implementing embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0029] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0030] The embodiments of the present disclosure relate to artificial intelligence technology fields such as natural language processing and voice interaction.

[0031] Artificial Intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence.

[0032] Deep learning is the process of learning the inherent laws and representation levels of sample data. The information obtained in the learning process is of great help in interpreting data such as text, images, and sounds. The ultimate goal of deep learning is to enable machines to have analytical learning capabilities like humans and to be able to recognize data such as text, images, and sounds.

[0033] Voice interaction is a way of interaction between humans and devices that transmit information through natural speech. Users can get feedback from the device by speaking.

[0034] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0035] The wake-up method, device, electronic device, and storage medium of the embodiments of the present disclosure are described below with reference to the accompanying drawings.

[0036] It should be noted that the executor of the wake-up method of this embodiment is a wake-up device, which can be implemented by software and / or hardware. The device can be configured in an electronic device, and the electronic device can include but is not limited to a terminal, a server, etc.

[0037] Currently, in vehicle-mounted scenarios, wake-up is usually based on a voice wake-up model. However, the voice wake-up model has limited training data and the vehicle environment is noisy, resulting in a high false wake-up rate.

[0038] Therefore, the wake-up method proposed in the present invention determines whether to wake up the system based on the degree of match between the corresponding text of the audio and the recognized wake-up word after the wake-up model recognizes the wake-up word in the speech, thereby improving the reliability of system wake-up, and optimizing and updating the wake-up model when the degree of match is low, thereby realizing online optimization of the wake-up model and reducing the false wake-up rate.

[0039] It should be noted that the wake-up method proposed in the present disclosure can be applied to voice system wake-up in other scenarios in addition to the in-vehicle voice interaction scenario. The embodiments of the present disclosure only take the in-vehicle voice interaction scenario as an example for explanation.

[0040] Figure 1 It is a flowchart of a wake-up method proposed according to an embodiment of the present disclosure.

[0041] like Figure 1 As shown, the wake-up method includes the following steps:

[0042] S101: storing the acquired audio data into a preset buffer area, and inputting the audio data into a preset wake-up model to obtain a first recognition result output by the wake-up model.

[0043] The preset buffer area is a data structure of fixed length, which is used to store real-time audio data, and can be called a buffer area, a wake-up buffer area, etc., which is not limited in the present disclosure.

[0044] Among them, the preset wake-up model can be any existing wake-up model, which can identify the wake-up word in the audio data through algorithm matching, notify the voice system to be awakened, and then conduct human-computer interaction.

[0045] In the disclosed embodiment, audio data such as human voice, noise, and media sound can be collected in real time through the vehicle-mounted microphone device, and the audio data can be stored in a preset cache area. In addition, the audio data can be input into a preset wake-up model at the same time, and the wake-up model can identify whether the audio data contains any wake-up word preset in the wake-up model through algorithm matching, thereby obtaining a first recognition result output by the wake-up model.

[0046] S102: When the first recognition result includes the wake-up word, a speech recognition service is called to recognize the audio data to obtain a second recognition result.

[0047] In the embodiment of the present disclosure, when the first recognition result includes the wake-up word, the accuracy of the first recognition result can be further verified. Then, the speech recognition service can be called first to recognize the currently collected audio data, that is, the audio data corresponding to the first recognition result including the wake-up word, and recognize the text corresponding to the audio data, so as to obtain the second recognition result.

[0048] It should be noted that the speech recognition service may be automatic speech recognition (ASR), which can analyze acoustic speech and obtain corresponding text or pinyin information.

[0049] S103: Determine whether to wake up the system based on the matching degree between the second recognition result and the wake-up word.

[0050] It should be noted that since the speech recognition service has a higher recognition accuracy for audio data than the wake-up model, it is possible to verify whether the system is woken up by calculating the matching degree between the second recognition result and the wake-up word.

[0051] In the disclosed embodiment, the degree of match between the second recognition result and the wake-up word can be determined by calculating the semantic similarity between the text of the second recognition result and the wake-up word, or determining whether the text of the second recognition result contains the wake-up word or a synonym of the wake-up word. Then, it is determined whether the degree of match meets the wake-up condition to determine whether to wake up the system. For example, it can be determined to wake up the system when the degree of match is greater than a preset degree of match threshold.

[0052] It should be noted that no matter whether the recognition result corresponding to the audio data can wake up the system or not, the audio data can be deleted from the cache, thereby freeing up the storage space of the cache and improving the performance of audio data wake-up recognition.

[0053] In this embodiment, the acquired audio data is first stored in a preset buffer area and input into a preset wake-up model to obtain a first recognition result output by the wake-up model. Then, when the wake-up word is included in the first recognition result, the speech recognition service is called to recognize the audio data to obtain a second recognition result. Afterwards, based on the degree of match between the second recognition result and the wake-up word, it is determined whether to wake up the system. Therefore, by further verifying whether to wake up the system based on the recognition result of the audio data by the speech recognition service and the match between the wake-up word recognized by the wake-up model, the accuracy and reliability of the voice system wake-up can be improved, the false wake-up rate can be reduced, and the user experience can be optimized.

[0054] For example, the present disclosure provides a method for waking up the vehicle voice system. When the user is chatting or making a phone call in the car, the microphone device will store the collected audio in the buffer area and input the audio into the wake-up model at the same time. When the collected audio includes the sentence "Hey, AA,...", and the wake-up model is preset with wake-up words such as "Hello, AA", "Hey, AA" and so on (AA can be the vehicle brand or the name of the vehicle voice system, or it can be empty), the wake-up model can recognize the wake-up word "Hey, BB" in the audio, and the first recognition result output is "Contains the wake-up word 'Hey, AA', which can wake up the voice system". Then, the voice recognition service can be called to recognize the audio data, and the text corresponding to the audio data (i.e., the second recognition result) is "Hey, AA,...". Then "Hey, AA" is matched with "Hey, AA,...". Since the semantics of the text is irrelevant to the wake-up word, the matching degree between the second recognition result and the wake-up word is very low, and the first recognition result is confirmed to be a false wake-up event, and the vehicle voice system is not woken up.

[0055] Figure 2 It is a flowchart of a wake-up method proposed in another embodiment of the present disclosure.

[0056] like Figure 2 As shown, the wake-up method includes the following steps:

[0057] S201: storing the acquired audio data into a preset buffer area, and inputting the audio data into a preset wake-up model to obtain a first recognition result output by the wake-up model.

[0058] The description of S201 can be specifically referred to the above embodiment, which will not be repeated here.

[0059] It should be noted that the first recognition result may also include the first start and end positions of the audio segment containing the wake-up word in the audio data for marking, so as to reduce the amount of data converted from audio to text by the speech recognition service, thereby improving the efficiency of wake-up judgment.

[0060] S202: Based on the first start and end positions, obtain the audio segment to be recognized from the audio data in the buffer area.

[0061] In the disclosed embodiment, the starting point and the ending point of the audio segment to be identified can be determined in the audio data in the buffer area according to the first starting and ending positions, and then the audio between the starting point and the ending point can be extracted to obtain the audio segment to be identified.

[0062] It should be noted that in some possible embodiments, the audio segment corresponding to the first start and end positions determined in the first recognition result may not fully describe the intention of the user's voice, resulting in low recognition accuracy, thereby affecting the accuracy of wake-up. Therefore, in the present disclosure, the position offset can be used to calibrate the first start and end positions.

[0063] Optionally, the position offset may be determined based on the distance between the first start and end positions.

[0064] In the disclosed embodiment, if the distance between the starting point and the ending point determined by the first starting and ending positions is relatively large, it can be explained that the wake-up word itself is relatively long and contains certain semantics. In this case, the position offset can be appropriately smaller.

[0065] Then, the first start and end positions can be updated based on the position offset to obtain the second start and end positions, and then the audio segments corresponding to the second start and end positions can be obtained from the audio data.

[0066] In the disclosed embodiment, the position offset can be used to adjust the start position and end position of the first start and end position in the audio to obtain the second start and end position. For example, the start position corresponding to the first start and end position can be moved forward by the position offset to obtain the start position corresponding to the second start and end position, and the end position corresponding to the first start and end position can be moved backward by the position offset to obtain the end position corresponding to the second start and end position. After that, the audio segment between the start position and the end position corresponding to the second start and end position can be obtained from the audio data. Thereby, the semantic integrity of the extracted audio segment can be guaranteed, thereby providing conditions for improving the reliability of audio recognition and wake-up judgment.

[0067] Optionally, when the distance between the first start and end positions is greater than a distance threshold, an audio segment corresponding to the first start and end positions may be acquired from the audio data.

[0068] The distance threshold may be a value set based on experience.

[0069] In this embodiment, when the distance between the first start and end positions is greater than the distance threshold, it can be considered that the audio segment extracted based on the first start and end positions contains the complete user intention. Therefore, the audio segment corresponding to the first start and end positions can be directly obtained from the audio data, which can ensure the semantic integrity of the extracted audio segment and provide conditions for improving the reliability of audio recognition and wake-up judgment.

[0070] S203: Calling a speech recognition service to recognize the audio segment to obtain a second recognition result.

[0071] It should be noted that the specific implementation form of calling the voice recognition service to recognize the audio clip is the same as the method of calling the voice recognition service to recognize the audio data in the above embodiment. Therefore, you can refer to the above embodiment and will not repeat it here.

[0072] In this embodiment, by extracting audio segments for wake-up verification from audio data based on the start and end positions in the wake-up model recognition results, the amount of data converted from audio to text by the speech recognition service can be reduced, thereby improving the efficiency of wake-up judgment.

[0073] S204: Acquire a text segment to be matched from the second recognition result according to the first start and end positions.

[0074] In the disclosed embodiment, since the audio and the text recognized by the speech recognition service are in one-to-one correspondence, the corresponding text segment to be matched can be extracted from the text recognized by the speech recognition service according to the first start and end positions corresponding to the audio segment. Alternatively, the text segment obtained by the speech recognition service directly recognizing the audio segment can be used, that is, the second recognition result is the text segment to be matched.

[0075] S205: Determine whether to wake up the system based on the matching degree between the text segment and the wake-up word.

[0076] In the disclosed embodiment, the matching degree between the text segment and the wake-up word can be determined by calculating the semantic similarity between the text segment and the wake-up word, or determining whether the text segment contains the wake-up word or a synonym of the wake-up word. Then, it is determined whether the matching degree meets the wake-up condition to determine whether to wake up the system. For example, the condition to be met for wake-up may be that the matching degree is greater than a preset matching degree threshold.

[0077] In this embodiment, the corresponding audio text segment is obtained by obtaining the start and end positions in the recognition results based on the wake-up model, and then the matching degree between the text segment and the wake-up word is calculated to determine whether to wake up the system, thereby reducing the amount of data for the wake-up word matching, improving the wake-up efficiency, and optimizing the user experience.

[0078] Figure 3It is a flowchart of a wake-up method proposed in another embodiment of the present disclosure.

[0079] like Figure 3 As shown, the wake-up method includes:

[0080] S301: storing the acquired audio data into a preset buffer area, and inputting the audio data into a preset wake-up model to obtain a first recognition result output by the wake-up model.

[0081] S302: When the first recognition result includes the wake-up word, call the speech recognition service to recognize the audio data to obtain a second recognition result.

[0082] The description of S301 and S302 can be specifically referred to the above embodiment, which will not be repeated here.

[0083] S303: based on the matching degree between the second recognition result and the wake-up word, when the matching degree is greater than a matching degree threshold, waking up the system.

[0084] The matching degree threshold may be determined based on experience.

[0085] In the disclosed embodiment, when the matching degree between the second recognition result and the wake-up word is greater than the matching degree threshold, it can be determined that the accuracy of the wake-up word recognition is high, so the reliability of the system wake-up event determined by the wake-up model based on the wake-up word is high, and the voice system can be awakened to start human-computer interaction. Therefore, by waking up the system again when the matching degree between the second recognition result and the wake-up word is greater than the wake-up condition, the verification of the voice system wake-up is achieved, the reliability and accuracy of the wake-up is improved, the probability of false wake-up caused by noise is reduced, the utilization rate of system resources and system performance are improved, and the user experience is optimized.

[0086] It should be noted that, in the present disclosure, the wake-up model can also be optimized based on the verification of the recognition result output by the wake-up model to improve the reliability and recognition stability of the wake-up model, thereby improving the efficiency and accuracy of the wake-up.

[0087] Optionally, when the degree of match is less than or equal to a threshold value of the degree of match, the audio data, the first recognition result, and the second recognition result may be sent to the server to trigger the server to update the wake-up model.

[0088] In the embodiment of the present disclosure, when the matching degree is less than or equal to the matching degree threshold, it can be determined that the recognition accuracy of the wake-up model for whether the audio contains the wake-up word is low, and there may be a false wake-up problem, and the wake-up model needs to be optimized. Therefore, the audio data, the first recognition result, and the second recognition result can be sent to the server to trigger the server to update the wake-up model. This can provide conditions for improving the reliability and recognition stability of the wake-up model, which is conducive to improving the efficiency and accuracy of the wake-up.

[0089] Optionally, when a wake-up model update instruction sent by the server is received, the preset wake-up model may be updated based on update parameters in the update instruction.

[0090] In the disclosed embodiment, the server can calculate the error of the wake-up model based on the received audio data, the first recognition result and the second recognition result, determine the parameters for updating the wake-up model and send an update instruction. Thus, when the wake-up model update instruction sent by the server is received, the preset wake-up model can be updated based on the update parameters in the update instruction. Thus, by using the update parameters sent by the server to update the wake-up model, the efficiency of the wake-up model optimization can be improved.

[0091] It should be noted that in the embodiment of the present disclosure, after obtaining the first recognition result output by the wake-up model, there may still be a situation where the first recognition result does not include the wake-up word. At this time, the wake-up model will not output a wake-up event and will not trigger the wake-up verification process. The audio data in the preset cache area can be directly cleared to free up the storage space of the cache area and improve the performance of audio data wake-up recognition.

[0092] The following can be combined Figure 4 The wake-up method described in the embodiment of the present disclosure is further illustrated by examples. Figure 4 The figure is a schematic diagram of an operation of a wake-up method.

[0093] like Figure 4 As shown, the microphone MIC can first collect human voice, noise, media sound, etc. to obtain audio data, and then store the audio data in the wake-up buffer and input it into the wake-up model for algorithm matching. If the wake-up word is not matched, the system will not be woken up and the process will be ended. Alternatively, if the wake-up word is matched, speech recognition is performed on the audio data in the wake-up buffer, and the result after speech recognition is further matched with the wake-up word. If the recognition result matches the wake-up word, the system is woken up and the process is ended. Alternatively, if the recognition result does not match the wake-up word, the system will not be woken up, and the audio data, speech recognition results, wake-up words and other data will be uploaded to the server for analysis and optimization of the wake-up model to complete the wake-up test process.

[0094] Figure 5 It is a structural diagram of a wake-up device proposed in an embodiment of the present disclosure.

[0095] like Figure 5 As shown, the wake-up device 500 includes:

[0096] The first recognition module 501 is used to store the acquired audio data into a preset buffer area and input it into a preset wake-up model to obtain a first recognition result output by the wake-up model;

[0097] The second recognition module 502 is used to call the speech recognition service to recognize the audio data to obtain a second recognition result when the first recognition result includes the wake-up word;

[0098] The determination module 503 is used to determine whether to wake up the system according to the matching degree between the second recognition result and the wake-up word.

[0099] In some embodiments, the second identification module 502 may be specifically used for:

[0100] Based on the first start and end positions, obtaining the audio segment to be recognized from the audio data in the buffer area;

[0101] A speech recognition service is called to recognize the audio segment to obtain a second recognition result.

[0102] In some embodiments, the second identification module 502 may be specifically used for:

[0103] Determining a position offset based on a distance between the first start and end positions;

[0104] Based on the position offset, the first start and end positions are updated respectively to obtain the second start and end positions;

[0105] From the audio data, obtain the audio segment corresponding to the second start and end positions.

[0106] In some embodiments, the second identification module 502 may be specifically used for:

[0107] According to the first start and end positions, obtaining a text segment to be matched from the second recognition result;

[0108] Determine whether to wake up the system based on the degree of match between the text fragment and the wake-up word.

[0109] In some embodiments, the second identification module 502 may be specifically used for:

[0110] When the distance between the first start and end positions is greater than the distance threshold, an audio segment corresponding to the first start and end positions is obtained from the audio data.

[0111] In some embodiments, the determination module 503 may be specifically configured to:

[0112] When the matching degree is greater than the matching degree threshold, the system is woken up.

[0113] In some embodiments, the determination module 503 may be specifically configured to:

[0114] When the matching degree is less than or equal to the matching degree threshold, the audio data, the first recognition result and the second recognition result are sent to the server to trigger the server to update the wake-up model.

[0115] In some embodiments, the wake-up device 500 may further include:

[0116] The updating module is used to update the preset wake-up model based on the update parameters in the update instruction when receiving the wake-up model update instruction sent by the server.

[0117] In some embodiments, the first identification module 501 may also be used to:

[0118] When the wake-up word is not included in the first recognition result, the audio data in the preset buffer area is cleared.

[0119] It should be noted that the above explanation of the awakening method is also applicable to the awakening device of this embodiment and will not be repeated here.

[0120] In this embodiment, by further verifying whether to wake up the system based on the recognition results of the audio data based on the speech recognition service and the match between the wake-up word recognized by the wake-up model, the accuracy and reliability of the voice system wake-up can be improved, the false wake-up rate can be reduced, and the user experience can be optimized.

[0121] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0122] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0123] like Figure 6As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0124] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0125] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as the wake-up method. For example, in some embodiments, the wake-up method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the wake-up method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the wake-up method in any other appropriate manner (e.g., by means of firmware).

[0126] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0127] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0128] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0129] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0130] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), the Internet, and a blockchain network.

[0131] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server of a distributed system, or a server combined with a blockchain.

[0132] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0133] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. In the description of the present disclosure, the words "if" and "if" used can be interpreted as "at the time of" or "when" or "in response to determination" or "under the circumstances of".

[0134] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A wake-up method, comprising: The acquired audio data is stored in a preset buffer area, and input into a preset wake-up model to obtain a first recognition result output by the wake-up model; In a case where the first recognition result includes a wake-up word, calling a speech recognition service to recognize the audio data to obtain a second recognition result; Determine whether to wake up the system based on the degree of match between the second recognition result and the wake-up word.

2. The method of claim 1, wherein: The first recognition result also includes the first start and end positions, and the calling of the speech recognition service to recognize the audio data to obtain the second recognition result includes: Based on the first start and end positions, obtaining an audio segment to be identified from the audio data in the buffer area; A speech recognition service is called to recognize the audio segment to obtain the second recognition result.

3. The method of claim 2, wherein: The acquiring the audio segment to be recognized from the audio data in the buffer area based on the first start and end positions includes: Determining a position offset based on a distance between the first start and end positions; Based on the position offset, the first start and end positions are updated respectively to obtain second start and end positions; An audio segment corresponding to the second start and end positions is obtained from the audio data.

4. The method of claim 3, wherein: The determining whether to wake up the system according to the matching degree between the second recognition result and the wake-up word includes: According to the first start and end positions, obtaining a text segment to be matched from the second recognition result; Determine whether to wake up the system based on the matching degree between the text segment and the wake-up word.

5. The method of claim 2, wherein: The acquiring the audio segment to be recognized from the audio data in the buffer area based on the first start and end positions includes: When the distance between the first start and end positions is greater than a distance threshold, an audio segment corresponding to the first start and end positions is acquired from the audio data.

6. The method according to any one of claims 1 to 5, wherein: The determining whether to wake up the system according to the matching degree between the second recognition result and the wake-up word includes: When the matching degree is greater than a matching degree threshold, the system is awakened.

7. The method according to any one of claims 1 to 5, wherein: The determining whether to wake up the system according to the matching degree between the second recognition result and the wake-up word includes: When the degree of match is less than or equal to a degree of match threshold, the audio data, the first recognition result, and the second recognition result are sent to a server to trigger the server to update the wake-up model.

8. The method of claim 7, wherein: The method further comprises: When the wake-up model update instruction sent by the server is received, the preset wake-up model is updated based on the update parameters in the update instruction.

9. The method according to any one of claims 1 to 5, wherein: After obtaining the first recognition result output by the wake-up model, the method further includes: When the first recognition result does not include the wake-up word, the audio data in the preset buffer area is cleared.

10. A wake-up device, comprising: A first recognition module, used to store the acquired audio data into a preset buffer area, and input it into a preset wake-up model to obtain a first recognition result output by the wake-up model; A second recognition module, configured to, when the first recognition result includes a wake-up word, call a speech recognition service to recognize the audio data to obtain a second recognition result; A determination module is used to determine whether to wake up the system according to the matching degree between the second recognition result and the wake-up word.

11. The device of claim 10, wherein: The second identification module is specifically used for: Based on the first start and end positions, obtaining an audio segment to be identified from the audio data in the buffer area; A speech recognition service is called to recognize the audio segment to obtain the second recognition result.

12. The device of claim 11, wherein: The second identification module is specifically used for: Determining a position offset based on a distance between the first start and end positions; Based on the position offset, the first start and end positions are updated respectively to obtain second start and end positions; An audio segment corresponding to the second start and end positions is obtained from the audio data.

13. The device of claim 12, wherein: The second identification module is specifically used for: According to the first start and end positions, obtaining a text segment to be matched from the second recognition result; Determine whether to wake up the system based on the matching degree between the text segment and the wake-up word.

14. The device of claim 11, wherein: The second identification module is specifically used for: When the distance between the first start and end positions is greater than a distance threshold, an audio segment corresponding to the first start and end positions is acquired from the audio data.

15. The device according to any one of claims 10 to 14, wherein: The determination module is specifically used for: When the matching degree is greater than a matching degree threshold, the system is awakened.

16. The device according to any one of claims 10 to 14, wherein: The determination module is specifically used for: When the degree of match is less than or equal to a degree of match threshold, the audio data, the first recognition result, and the second recognition result are sent to a server to trigger the server to update the wake-up model.

17. The device of claim 16, wherein: The device also includes: The updating module is used to update the preset wake-up model based on the update parameters in the update instruction when receiving the wake-up model update instruction sent by the server.

18. The device according to any one of claims 10 to 14, wherein: The first identification module is further used for: When the first recognition result does not include the wake-up word, the audio data in the preset buffer area is cleared.

19. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the wake-up method according to any one of claims 1 to 9.

20. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: in, The computer instructions are used to enable the computer to execute the wake-up method according to any one of claims 1 to 9.

21. A computer program product, characterized in that The invention comprises a computer program, which implements the steps of the wake-up method according to any one of claims 1 to 9 when being executed by a processor.

Images