Intelligent voice prompt method, device, equipment and readable storage medium

Abstract

The present application relates to the field of artificial intelligence technology, and discloses an intelligent voice prompt method, including: obtaining a target prompt text, and determining a target smart home device for voice prompting; obtaining an input sample according to the target prompt text and the target smart home device, and sending the input sample to a pre-trained intonation prediction model to output corresponding intonation information; generating corresponding voice information according to the intonation information and the target prompt text; sending the voice information to the target smart home device to control the target smart home device to play the voice information. The present application also provides a device, a computer device, and a storage medium. Effectively improves the intelligence of voice prompts and improves user experience.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an intelligent voice prompt method, an intelligent voice prompt device, a computer device, and a computer-readable storage medium. Background Art

[0002] With the development of artificial intelligence and Internet of Things technology, the popularity of smart home is gradually increasing, becoming one of the hot topics discussed by people in modern life. The development of smart home can bring more convenience to people's lives. The current smart home has entered a new stage of rapid market development, and its potential market demand is also increasing day by day. For example, the concept of smart home is used in many aspects such as home security, intelligent control of household appliances, lighting adjustment, multimedia resource sharing, and intelligent management of kitchens and bathrooms.

[0003] However, nowadays, smart home is still in the process of development, and there are still many problems to be solved, such as its adaptability, security, privacy, interoperability and manageability. There are many problems that need to be solved by new-age IT talents. This means that the development of smart home is still facing many challenges. Only by solving these challenges can we fundamentally meet the needs of different users for smart home, thereby further improving the living standards of users and changing the lifestyles of residents.

[0004] The existing voice prompting method is relatively fixed, and voice prompting is simply performed according to the actual situation, so that the user may not be able to give feedback to the voice prompt in time, resulting in poor effect of the voice prompting. Summary of the invention

[0005] The present application provides an intelligent voice prompt method, device, computer equipment and storage medium to improve the intelligence of voice prompts and enhance user experience.

[0006] In a first aspect, the present application provides an intelligent voice prompt method, the method comprising:

[0007] Obtain the target prompt text and determine the target smart home device for voice prompting;

[0008] Obtaining an input sample according to the target prompt text and the target smart home device, and sending the input sample to a pre-trained intonation prediction model to output corresponding intonation information;

[0009] Generate corresponding voice information according to the intonation information and the target prompt text;

[0010] The voice information is sent to the target smart home device to control the target smart home device to play the voice information.

[0011] In a second aspect, the present application also provides an intelligent voice prompt device, the device comprising:

[0012] An information determination module is used to obtain a target prompt text and determine a target smart home device for voice prompting;

[0013] A model prediction module, used to obtain an input sample according to the target prompt text and the target smart home device, and send the input sample to a pre-trained intonation prediction model to output corresponding intonation information;

[0014] A speech generation module, used to generate corresponding speech information according to the intonation information and the target prompt text;

[0015] The voice feedback module is used to send the voice information to the target smart home device to control the target smart home device to play the voice information.

[0016] In a third aspect, the present application also provides a computer device, comprising a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and implement the intelligent voice prompt method as described above when executing the computer program.

[0017] In a fourth aspect, the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the processor implements the intelligent voice prompt method as described above.

[0018] The present application discloses an intelligent voice prompt method, device, computer equipment and storage medium. In a smart home system, when the system determines that a voice prompt is needed, it first obtains a target prompt text and determines a target smart home device for voice prompting, then obtains a corresponding input sample according to the target prompt text and the target smart home device, and inputs the obtained input sample into a pre-trained intonation prediction model to obtain corresponding intonation information, then generates corresponding voice information according to the target prompt text and the obtained intonation information, and finally sends the obtained voice information to the target smart home device so that the target smart home device plays the obtained voice information. It is achieved that when a voice prompt is performed, the played voice information is obtained according to the user's actual scene information and actual needs, making the voice prompt more intelligent and having a better user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 A flowchart of an intelligent voice prompt method provided by an embodiment of the present application;

[0021] Figure 2 A flowchart of steps for determining to start an intelligent voice prompt provided in an embodiment of the present application;

[0022] Figure 3 A flowchart of the steps of obtaining intonation information provided in an embodiment of the present application;

[0023] Figure 4 A flowchart of the steps of obtaining a trained intonation prediction model provided in an embodiment of the present application;

[0024] Figure 5 A flowchart of the steps of updating a model provided in one embodiment of the present application;

[0025] Figure 6 This is a schematic block diagram of an intelligent voice prompt device in one embodiment of the present application;

[0026] Figure 7 This is a schematic block diagram of the structure of a computer device in one embodiment of the present application. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0029] It should be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0030] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0031] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0032] See also Figure 1 , Figure 1 A flowchart of an intelligent voice prompt method provided for an embodiment of the present application.

[0033] like Figure 1 As shown, the intelligent voice prompt method includes steps S101 to S105.

[0034] Step S101, obtaining a target prompt text, and determining a target smart home device for voice prompting;

[0035] The user sets the corresponding reminder rules in advance, and the system generates the corresponding reminder text at the appropriate time according to the set reminder rules, and then the system sends the corresponding voice prompt information according to the obtained prompt text to send the corresponding prompt information to the user. For example, the corresponding reminder rules are set according to time. When the user sets the working time, the corresponding reminder information is sent to the user at the appropriate time according to the current time information, such as: There are 30 minutes left before the working time, please go out in time.

[0036] Therefore, when providing voice prompt feedback, the target prompt text is first obtained, wherein the target prompt text can be pre-set or intelligently generated. In addition, the prompt text has different importance levels, for example, some prompt texts are only used for suggestions, while some prompt texts are used to force people to execute, so when providing feedback, the feedback method will also be different.

[0037] Furthermore, when determining the target prompt text, corresponding text can be added according to the actual situation. For example, when prompting the user to go out, if it is detected that it is raining outside, the user can be reminded to bring an umbrella according to the actual weather forecast data. Therefore, when obtaining the target prompt text, the target prompt text can be improved accordingly.

[0038] Furthermore, when the target prompt text is obtained, it is used to send corresponding voice prompt information to the user according to the target prompt text. Therefore, before obtaining the target prompt text, it is necessary to determine whether the user is in an indoor environment, that is, whether the voice prompt function needs to be triggered.

[0039] After the target prompt text is determined, a corresponding prompt voice will be generated according to the target prompt text for playback. When playing the voice information, it is necessary to reasonably select the smart home device for voice playback. Therefore, while obtaining the target prompt text, the target smart home device for voice prompting will also be determined, so as to control the determined target smart home device to play the voice information after the generation of the prompt voice is completed.

[0040] Therefore, refer to Figure 2 , Figure 2 A flowchart of the steps for determining to start an intelligent voice prompt is provided in an embodiment of the present application. Specifically, before obtaining the target prompt text and determining the target smart home device for voice prompting, it also includes:

[0041] Step S201, acquiring an image captured by an image acquisition device;

[0042] Step S202: Analyze and process the image to determine whether to start the intelligent voice prompt, and obtain the target prompt text when the intelligent voice prompt is started.

[0043] When it is necessary to play a voice information prompt, it is necessary to ensure that the user is in an indoor environment, that is, it is necessary to determine whether to start the system's intelligent voice prompt function at this time. Therefore, before obtaining the target prompt text and determining the target smart home device, it is also necessary to determine whether to start the system's intelligent voice prompt. Only when it is determined to start the system's intelligent voice prompt, will the target prompt text and the target smart home device be obtained, otherwise the relevant information will not be obtained.

[0044] During the intelligent voice prompting process, the system will always obtain the user's location information so that when prompting the user, it can issue prompt information more quickly, conveniently and effectively. When determining the user's location, the system obtains the image captured by the set image acquisition device, wherein the system establishes a communication connection between the image acquisition devices, and then processes the captured image accordingly to obtain the user's location in the indoor environment.

[0045] Therefore, when determining the user's location, the image captured by the image acquisition device installed indoors is obtained, and then the image is analyzed and processed to determine the user's location information, wherein the image acquisition device can be a monitoring device installed indoors. After the monitoring video is acquired by the monitoring device, the image frame of the monitoring video is acquired to obtain the relevant image. When determining the location, the actual indoor situation is collected to determine the user's actual location, and then the actual location can be used to determine the scene where the user is actually located, such as bedroom, living room, kitchen, etc., and then the appropriate voice information for feedback is determined according to different scenes.

[0046] Furthermore, when determining the location of the user, the infrared sensing device installed on each smart home device can also be used to collect infrared signals to determine whether there is a user in the indoor environment and the location of the user. Specifically, the process includes: receiving infrared information collected by each smart home device based on the infrared sensing device installed, and determining whether to start the intelligent voice prompt according to the received infrared signal. By analyzing the infrared signal, the actual state of the indoor environment is determined, and then it is determined whether the intelligent voice prompt is needed.

[0047] Furthermore, when determining whether to start the intelligent voice prompt, it is also possible to monitor the status of the smart home devices, such as whether one or some smart home devices are in operation, such as whether the air conditioner is in operation, etc. It is also possible to determine whether to start the intelligent voice prompt by monitoring the status of the smart door, for example, using the smart door to determine whether the user is at home. Specifically, the image acquisition device provided on the smart door can be used to determine whether the user has left home or arrived home, so as to determine whether there is a user in the room.

[0048] It should be noted that there are many ways to determine whether to start the intelligent voice prompt, which are not limited here. You can choose the appropriate method according to the actual situation to determine when to start the intelligent voice prompt.

[0049] Furthermore, when acquiring images, since the user is not always in an indoor environment, and is in an outdoor environment more often, and when in an outdoor environment, there is obviously no need to provide corresponding prompts to the user. Therefore, when acquiring images, the image may not be acquired in real time, but may be determined based on the actual indoor environment.

[0050] For example, by monitoring the smart door, when it is determined that the user has left home and has not returned home, there is no need to acquire and collect images, that is, there is no need to analyze and process the images. Only when the user is at home (i.e., in an indoor environment) is it necessary to acquire relevant images and analyze and process the images.

[0051] In addition, when determining the target smart home device for voice prompting, it includes: determining that the image contains a user tag and a location tag corresponding to the user tag; and querying in the device list according to the location tag to determine the target smart home device for voice prompting.

[0052] When determining the smart home device for voice prompts, it is necessary to reasonably select a more suitable smart home device so that when the smart home device is used for voice prompts, the user can hear and give feedback in a timely manner.

[0053] For example, if the user is in the bedroom and controls the smart refrigerator to give voice prompts, it is obviously not smart enough. It is very likely that the user cannot hear the voice information played, or cannot give timely feedback on the voice information. Therefore, it is necessary to reasonably choose smart home devices that give voice prompts.

[0054] When determining the target smart home device for voice prompting, it is determined by the collected image, by obtaining the user tag of the user in the image and the location tag corresponding to the user tag, and then querying in the device list according to the location tag to determine the target smart home device for voice prompting.

[0055] In actual applications, after acquiring the image, in addition to determining the user's location, since the entire system includes a large number of smart home devices, it is obviously impractical for all smart home devices to play voice information when giving voice prompts. Therefore, it is necessary to reasonably select appropriate smart home devices to realize the playback of voice information.

[0056] Therefore, after the user's position is determined based on the image captured by the image capture device, the target smart home device is determined based on the determined position information, wherein the target smart home device is the smart home device that plays the voice information.

[0057] In fact, when determining the target smart home device, you can pre-construct a location display map of the smart home device, such as using a coordinate system for display. Then, after determining the user's location information, merge the position corresponding to the location information into the pre-constructed coordinate system. Then, based on the coordinate position of the smart home device in the coordinate system and the coordinate position of the user, determine the target smart home device. For example, select the smart home device that is closest to the user and in the user's current scene as the target smart home device, such as a smart refrigerator in the kitchen, a smart desk lamp in the bedroom, or a smart speaker in the living room.

[0058] Step S102: obtaining an input sample according to the target prompt text and the target smart home device, and sending the input sample to a pre-trained intonation prediction model to output corresponding intonation information.

[0059] After obtaining the target prompt text and the target smart home device for voice prompting, the current input sample will be obtained based on the target prompt text and the target smart home device, and then the obtained input sample will be input into a pre-trained intonation prediction model to output the current corresponding intonation information.

[0060] In actual applications, different prompt texts and different combinations of environmental information correspond to different tone information. For example, when urging users to go out to work, the tone information of the prompt voice when the user is in the bedroom and the dining room is different. For example, when in the bedroom, because the user may not get up on time, the tone will be very urgent, while in the dining room, because the user may be eating breakfast, it will be judged that the user is about to go out, so the tone may be relatively gentle. For another example, when suggesting users to drink water, a gentle tone will be used for feedback, and when the user is about to be late for work, an urging tone will be used for feedback.

[0061] Therefore, when the target prompt text and the target smart home device are obtained, the corresponding input samples will be obtained according to the target prompt text and the target smart home device, and then the current corresponding intonation information will be determined according to the obtained input samples.

[0062] Specifically, refer to Figure 3 , Figure 3 A flowchart of the steps of obtaining intonation information provided in an embodiment of the present application. Step S102 includes:

[0063] Step S301: Analyze and process the target prompt text to determine the prompt level corresponding to the target prompt text;

[0064] Step S302: determining the current corresponding environment information according to the location tag, and summarizing the prompt level and the environment information to obtain a corresponding input sample;

[0065] Step S303: input the input sample into the trained intonation prediction model to output corresponding intonation information.

[0066] After obtaining the target prompt text and the target smart home device, first determine the current corresponding intonation information, and then obtain corresponding voice information according to the obtained intonation information and the target prompt text to perform voice prompt.

[0067] When determining the intonation information, the target prompt text is first analyzed and processed to determine the prompt level corresponding to the target prompt text, wherein the prompt level is obtained based on the target prompt text and the current actual state. For example, after setting the working time, the closer to the working time, the higher the corresponding prompt level will be. For example, when the user has not drunk water for a long time, the longer the interval between drinking water, the higher the corresponding prompt level will be.

[0068] In actual applications, different target prompt texts correspond to different prompt levels. For example, the prompt text that reminds users to drink water and the prompt text that reminds users that they are late for work have different corresponding prompt levels. Different prompt levels will also have different dominant roles when making voice prompts. For example, when the prompt level is high, an urging tone may be needed to remind the user, while when the prompt level is low, a gentle tone may be needed to remind the user.

[0069] When the prompt level corresponding to the target prompt text is determined, the user's current environmental information will also be determined. Specifically, the current corresponding environmental information is determined according to the location tag, wherein the environmental information is the user's current indoor environment state, such as being in a bedroom, dining room or kitchen, etc. Then, after determining the current corresponding environmental information, the obtained prompt level and environmental information are summarized to obtain an input sample for prediction, and finally the obtained input sample is input into a pre-trained intonation prediction model to output the current corresponding intonation information.

[0070] In practical applications, the steps of determining the prompt level and the environmental information of the target prompt text can be performed simultaneously, and there may be no order of precedence. After obtaining the prompt level and the environmental information, the pre-trained intonation prediction model will be used to predict the intonation information.

[0071] Furthermore, the pre-trained intonation prediction model is obtained by pre-training when it is used for the first time, and in the subsequent continuous use process, the trained intonation prediction model will be fine-tuned and updated according to actual feedback, so that the intonation prediction model used can better meet the needs of users. Figure 4 , Figure 4 A flowchart of the steps of obtaining a trained intonation prediction model provided in one embodiment of the present application.

[0072] Among them, the trained intonation training model includes:

[0073] Step S401: loading a to-be-trained intonation prediction model, wherein the intonation prediction model is obtained based on a neural network model;

[0074] Step S402, receiving input speech information to be trained, and inputting the speech information to be trained into the intonation prediction model to be trained, so as to detect whether the trained intonation prediction model converges;

[0075] Step S403: When it is determined that the trained intonation prediction model has converged, the trained intonation prediction model is saved.

[0076] When a usable intonation prediction model is obtained through training, the intonation prediction model to be trained is first loaded, wherein the intonation prediction model to be trained is obtained based on a neural network model, and then input voice information to be trained is received, and then the received voice information to be trained is input into the intonation prediction model to be trained, so as to realize the training of the intonation prediction model to be trained, and at the same time determine whether the trained intonation prediction model converges, so as to save the trained intonation prediction model when it is determined that the trained intonation prediction model converges.

[0077] Among them, when determining whether the trained intonation prediction model converges, it can be determined whether a parameter in the intonation prediction model meets the set conditions, or it can be determined whether the number of training times meets the set conditions. There is no limitation here and it can be determined according to actual conditions.

[0078] Furthermore, when training the intonation prediction model to be trained based on the voice information to be trained, it includes: receiving the input voice information to be trained, and the prompt level corresponding to the voice information; identifying the environmental information corresponding to the voice information to be trained, and associating the prompt level corresponding to the voice information to be trained with the environmental information, so as to input it into the intonation prediction model to be trained.

[0079] When receiving voice information to be trained, where the voice information to be trained is a training sample for model training, relevant information corresponding to the voice information to be trained is identified, including identifying the prompt level corresponding to the voice information to be trained and the environmental information corresponding to the voice information to be trained, and then the obtained prompt level and environmental information are associated to obtain the corresponding training sample for model training, and then the obtained training sample is input into the intonation prediction model to be trained to achieve corresponding training of the intonation prediction model to be trained.

[0080] During training, due to the huge number of speech information samples, there are many combinations of prompt levels and environmental information corresponding to the same text information in the obtained training samples. Therefore, after obtaining the trained intonation prediction model, different text information will have different intonation information under different environmental information.

[0081] Step S103: Generate corresponding voice information according to the intonation information and the target prompt text.

[0082] After the intonation information is determined, the intonation information is merged with the prompt text to generate corresponding voice information, so that the target smart home device plays the obtained voice information.

[0083] Step S104: Send the voice information to the target smart home device to control the target smart home device to play the voice information.

[0084] After the voice information for playback and feedback is obtained, the obtained voice information is sent to a predetermined target smart home device, so that the target smart home device plays the voice information after receiving the voice information.

[0085] Further, refer to Figure 5 , Figure 5 A flowchart of the steps of model updating provided in one embodiment of the present application.

[0086] Specifically, updating the model includes:

[0087] Step S501, receiving first information in response to the voice information input, and determining a first acceptance level corresponding to the first information;

[0088] After the target smart home device plays the obtained voice information, it can receive the first information sent by the user based on the voice information, wherein the first information includes the user's acceptance level of the voice information, so the first information may include different level information of acceptance level such as low, medium and high. Therefore, after receiving the first information, the corresponding first acceptance level is determined at this time, so as to update the intonation prediction model according to the obtained acceptance level.

[0089] Step S502: if it is determined that the first acceptance level is higher than the preset first acceptance level, then it is determined that the intonation prediction is accurate;

[0090] After the first acceptance level is obtained, since the acceptance levels are divided into high and low levels, and different acceptance levels represent the user's acceptance level, when the acceptance level is high, it means that the user is more accepting of the reminder method obtained at this time, and when the acceptance level is low, it means that the user is not very accepting of the reminder method obtained at this time. Therefore, when it is determined that the obtained first acceptance level is higher than the preset acceptance level, it is determined that the intonation prediction is accurate.

[0091] Step S503: if it is determined that the first acceptance level is lower than a preset second acceptance level, recording the intonation information and the target prompt text;

[0092] Similarly, when it is determined that the first acceptance level is lower than the preset second acceptance level, it means that the user does not like the reminder method of the voice information obtained at this time, that is, it means that the prediction result at this time does not meet the user's needs. Therefore, the obtained intonation information and prompt text will be recorded at this time, and then the recorded information will be used as a sample in the model update sample set, so that when it is determined that the model is regularly updated and maintained, the updated sample set can be used to perform incremental training on the model.

[0093] Step S504: when a model update instruction is received, the recorded update sample set is obtained, and the intonation prediction model is trained based on the update sample set to obtain an updated intonation prediction model when the model converges.

[0094] Since the prediction results of the model may not necessarily perfectly meet and conform to the actual needs of users, the model needs to be updated regularly so that the model can make predictions more accurately based on user needs. Therefore, when determining to update, obtain the pre-recorded update sample set and use the updated sample set to perform incremental training on the intonation prediction model to obtain an updated intonation prediction model.

[0095] As for the update sample set, it is updated in real time. Before completing a model update, the update sample set can be cleared, or it can be not cleared to delete the oldest samples when the number of samples is too large.

[0096] In the intelligent voice prompt method described above, in the smart home system, when the system determines that a voice prompt is needed, it first obtains the target prompt text and determines the target smart home device for voice prompting, then obtains the corresponding input sample according to the target prompt text and the target smart home device, and inputs the obtained input sample into the pre-trained intonation prediction model to obtain the corresponding intonation information, then generates the corresponding voice information according to the target prompt text and the obtained intonation information, and finally sends the obtained voice information to the target smart home device so that the target smart home device plays the obtained voice information. This achieves that when a voice prompt is given, the voice information played is obtained according to the user's actual scene information and actual needs, making the voice prompt more intelligent and providing a better user experience.

[0097] See also Figure 6 , Figure 6 This is a schematic block diagram of an intelligent voice prompt device in one embodiment of the present application, which is used to execute the aforementioned intelligent voice prompt method.

[0098] like Figure 6 As shown, the intelligent voice prompt device 600 includes:

[0099] The information determination module 601 is used to obtain the target prompt text and determine the target smart home device for voice prompting;

[0100] A model prediction module 602 is used to obtain an input sample according to the target prompt text and the target smart home device, and send the input sample to a pre-trained intonation prediction model to output corresponding intonation information;

[0101] A speech generation module 603 is used to generate corresponding speech information according to the intonation information and the target prompt text;

[0102] The voice feedback module 604 is used to send the voice information to the target smart home device to control the target smart home device to play the voice information.

[0103] Furthermore, in one embodiment, the intelligent voice prompting device 600 is further used for:

[0104] Loading a to-be-trained intonation prediction model, wherein the to-be-trained intonation prediction model is obtained based on a neural network model; receiving input speech information to be trained, and inputting the to-be-trained speech information into the to-be-trained intonation prediction model to detect whether the to-be-trained intonation prediction model converges;

[0105] When it is determined that the trained intonation prediction model converges, the trained intonation prediction model is saved.

[0106] Furthermore, in one embodiment, the intelligent voice prompting device 600 is further used for:

[0107] Receive input voice information to be trained and the prompt level corresponding to the voice information; identify the environmental information corresponding to the voice information to be trained, and associate the prompt level corresponding to the voice information to be trained with the environmental information to input into the intonation prediction model to be trained.

[0108] Furthermore, in one embodiment, the intelligent voice prompting device 600 is further used for:

[0109] Acquire an image captured by an image acquisition device; analyze and process the image to determine whether to start the intelligent voice prompt, and acquire the target prompt text when the intelligent voice prompt is started.

[0110] Furthermore, in one embodiment, the information determination module 601 is further configured to:

[0111] Determine that the image includes a user tag and a location tag corresponding to the user tag; and query in a device list according to the location tag to determine a target smart home device for voice prompting.

[0112] Furthermore, in one embodiment, the model prediction module 602 is further configured to:

[0113] The target prompt text is analyzed and processed to determine the prompt level corresponding to the target prompt text; the current corresponding environmental information is determined according to the position tag, and the prompt level and the environmental information are summarized to obtain a corresponding input sample; the input sample is input into the trained intonation prediction model to output the corresponding intonation information.

[0114] Furthermore, in one embodiment, the intelligent voice prompting device 600 is further used for:

[0115] Receive first information in response to the voice information input, and determine a first acceptance level corresponding to the first information; if it is determined that the first acceptance level is higher than a preset first acceptance level, determine that the intonation prediction is accurate; if it is determined that the first acceptance level is lower than a preset second acceptance level, record the intonation information and the target prompt text; when a model update instruction is received, obtain the recorded update sample set, and train the intonation prediction model based on the update sample set to obtain an updated intonation prediction model when the model converges.

[0116] It should be noted that those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described device and each module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0117] The above-mentioned device can be implemented in the form of a computer program. Figure 7 Runs on the computer device shown.

[0118] See also Figure 7 , Figure 7 This is a schematic block diagram of the structure of a computer device in one embodiment of the present application. The computer device may be a server.

[0119] See also Figure 7 The computer device includes a processor, a memory and a network interface connected through a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.

[0120] The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute any intelligent voice prompt method.

[0121] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

[0122] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor, the processor can execute any intelligent voice prompt method.

[0123] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0124] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0125] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps:

[0126] Acquire a target prompt text and determine a target smart home device for voice prompting; obtain an input sample based on the target prompt text and the target smart home device, and send the input sample to a pre-trained intonation prediction model to output corresponding intonation information; generate corresponding voice information based on the intonation information and the target prompt text; send the voice information to the target smart home device to control the target smart home device to play the voice information.

[0127] In one embodiment, when implementing the computer program, the processor is further configured to implement:

[0128] Loading a to-be-trained intonation prediction model, wherein the to-be-trained intonation prediction model is obtained based on a neural network model; receiving input speech information to be trained, and inputting the to-be-trained speech information into the to-be-trained intonation prediction model to be trained, so as to detect whether the to-be-trained intonation prediction model converges; when it is determined that the trained intonation prediction model converges, saving the trained intonation prediction model.

[0129] In one embodiment, when the processor implements the receiving of the input speech information to be trained and inputting the speech information to be trained into the intonation prediction model to be trained, the processor is further configured to implement:

[0130] Receive input voice information to be trained and the prompt level corresponding to the voice information; identify the environmental information corresponding to the voice information to be trained, and associate the prompt level corresponding to the voice information to be trained with the environmental information to input into the intonation prediction model to be trained.

[0131] In one embodiment, before obtaining the target prompt text and determining the target smart home device for voice prompting, the processor is further configured to implement:

[0132] Acquire an image captured by an image acquisition device; analyze and process the image to determine whether to start the intelligent voice prompt, and acquire the target prompt text when the intelligent voice prompt is started.

[0133] In one embodiment, when the processor implements the step of determining the target smart home device for voice prompting, the processor is further configured to implement:

[0134] Determine that the image includes a user tag and a location tag corresponding to the user tag; and query in a device list according to the location tag to determine a target smart home device for voice prompting.

[0135] In one embodiment, when the processor obtains an input sample according to the target prompt text and the target smart home device, and sends the input sample to a pre-trained intonation prediction model to output corresponding intonation information, it is also used to implement:

[0136] The target prompt text is analyzed and processed to determine the prompt level corresponding to the target prompt text; the current corresponding environmental information is determined according to the position tag, and the prompt level and the environmental information are summarized to obtain a corresponding input sample; the input sample is input into the trained intonation prediction model to output the corresponding intonation information.

[0137] In one embodiment, after implementing the sending of the voice information to the target smart home device to control the target smart home device to play the voice information, the processor is further configured to implement:

[0138] Receive first information in response to the voice information input, and determine a first acceptance level corresponding to the first information; if it is determined that the first acceptance level is higher than a preset first acceptance level, determine that the intonation prediction is accurate; if it is determined that the first acceptance level is lower than a preset second acceptance level, record the intonation information and the target prompt text; when a model update instruction is received, obtain the recorded update sample set, and train the intonation prediction model based on the update sample set to obtain an updated intonation prediction model when the model converges.

[0139] A computer-readable storage medium is also provided in an embodiment of the present application, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and the processor executes the program instructions to implement any one of the intelligent voice prompt methods provided in the embodiments of the present application.

[0140] The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc., equipped on the computer device.

[0141] Furthermore, the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the blockchain node, etc.

[0142] In addition, the blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0143] The above is only a specific implementation of the present application, but the protection scope of the present application 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 in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. An intelligent voice prompt method, characterized in that: The method comprises: Obtain the target prompt text and determine the target smart home device for voice prompting; Obtaining an input sample according to the target prompt text and the target smart home device, and sending the input sample to a pre-trained intonation prediction model to output corresponding intonation information; Generate corresponding voice information according to the intonation information and the target prompt text; The voice information is sent to the target smart home device to control the target smart home device to play the voice information.

2. The method according to claim 1, characterized in that The method further comprises: Loading a to-be-trained intonation prediction model, wherein the intonation prediction model is obtained based on a neural network model; Receiving input speech information to be trained, and inputting the speech information to be trained into the intonation prediction model to be trained to detect whether the intonation prediction model to be trained converges; When it is determined that the trained intonation prediction model converges, the trained intonation prediction model is saved.

3. The method according to claim 2, characterized in that The step of receiving the input speech information to be trained and inputting the speech information to be trained into the intonation prediction model to be trained comprises: Receiving input voice information to be trained and a prompt level corresponding to the voice information; The environmental information corresponding to the voice information to be trained is identified, and the prompt level corresponding to the voice information to be trained is associated with the environmental information to be input into the intonation prediction model to be trained.

4. The method according to claim 1, characterized in that Before obtaining the target prompt text and determining the target smart home device for voice prompting, the following steps are included: Acquiring an image captured by an image acquisition device; The image is analyzed and processed to determine whether to start the intelligent voice prompt, and when it is determined that the intelligent voice prompt is started, the target prompt text is obtained.

5. The method according to claim 4, characterized in that The step of determining a target smart home device for voice prompting includes: Determining that the image includes a user tag and a location tag corresponding to the user tag; A query is performed in the device list according to the location tag to determine the target smart home device for voice prompting.

6. The method according to claim 5, characterized in that The step of obtaining an input sample according to the target prompt text and the target smart home device, and sending the input sample to a pre-trained intonation prediction model to output corresponding intonation information includes: Analyzing and processing the target prompt text to determine the prompt level corresponding to the target prompt text; Determine the current corresponding environment information according to the location tag, and summarize the prompt level and the environment information to obtain a corresponding input sample; The input sample is input into the trained intonation prediction model to output corresponding intonation information.

7. The method according to any one of claims 1 to 6, characterized in that After sending the voice information to the target smart home device to control the target smart home device to play the voice information, the method further includes: receiving first information input in response to the voice information, and determining a first acceptance level corresponding to the first information; If it is determined that the first acceptance level is higher than the preset first acceptance level, then it is determined that the intonation prediction is accurate; If it is determined that the first acceptance level is lower than a preset second acceptance level, recording the intonation information and the target prompt text; When a model update instruction is received, the recorded update sample set is obtained, and the intonation prediction model is trained based on the update sample set to obtain an updated intonation prediction model when the model converges.

8. An intelligent voice prompting device, characterized in that: The device comprises: An information determination module is used to obtain a target prompt text and determine a target smart home device for voice prompting; A model prediction module, used to obtain an input sample according to the target prompt text and the target smart home device, and send the input sample to a pre-trained intonation prediction model to output corresponding intonation information; A speech generation module, used to generate corresponding speech information according to the intonation information and the target prompt text; The voice feedback module is used to send the voice information to the target smart home device to control the target smart home device to play the voice information.

9. A computer device, characterized in that: Including memory and processor: The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the intelligent voice prompt method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, one or more processors execute the steps of the intelligent voice prompt method as described in any one of claims 1 to 7.

Images