Volume adjustment method, electronic equipment and storage medium
By establishing a volume prediction model based on user personal information, using deep learning technology to analyze environmental sound data, and adaptively adjust the volume of terminal equipment, the problem of inability to personalize the volume in the existing technology is solved, and the user experience is improved.
Patent Information
- Application Number
- CN202411029097.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-15
AI Technical Summary
The existing automatic volume adjustment system cannot accurately adjust according to the user's personalized needs, resulting in users needing to frequently manually adjust the volume, affecting the user experience.
By obtaining the user's personal information, establishing a volume prediction model, using deep learning technology to analyze the environmental sound data, adaptively adjust the volume of the terminal device, and adjust the volume according to the user's personalized preferences.
It realizes adaptive adjustment of volume according to the ambient sound data, and personalized automatic adjustment of volume for different users, improving the problem that terminal devices cannot personalize the automatic adjustment of volume for different users, and improving the user experience.
Smart Images

Figure CN118972738A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a volume adjustment method, an electronic device and a storage medium. Background Art
[0002] As users' demands for personalized and intelligent audio systems continue to increase, existing automatic volume adjustment systems are mainly based on a single parameter, such as vehicle speed or external environmental noise level, and these systems have some limitations. For example, personalized needs are not taken into account. Different people may have different hearing preferences. The current system cannot make precise adjustments based on the personalized needs of users, resulting in users frequently manually adjusting the volume when the external environmental noise changes, which affects the user experience. Summary of the invention
[0003] In view of this, an object of the embodiments of the present application is to provide a volume adjustment method, an electronic device, and a storage medium, which can improve the problem that the terminal device cannot automatically adjust the volume in a personalized manner for different users.
[0004] In order to achieve the above technical objectives, the technical solutions adopted in this application are as follows:
[0005] In a first aspect, an embodiment of the present application provides a volume adjustment method, the method comprising:
[0006] Obtain environmental sound data around the terminal device;
[0007] Analyzing and processing the environmental sound data based on a volume prediction model corresponding to the personal information of the current user to obtain a predicted volume;
[0008] Based on the predicted volume, the volume of the sound output by the terminal device is adjusted.
[0009] In combination with the first aspect, in some optional implementations, before analyzing and processing the ambient sound data based on the volume prediction model corresponding to the personal information of the current user, the method further includes:
[0010] Get the current user's personal information;
[0011] A data model corresponding to the personal information is selected from a pre-established model library as the volume prediction model, wherein the model library stores at least two data models for volume prediction based on deep learning.
[0012] In combination with the first aspect, in some optional implementations, before obtaining the personal information of the current user, the method further includes:
[0013] Acquire a data set of multiple users, the data set including personal information, environmental sound data, and adjusted volume corresponding to each user;
[0014] Cleaning the data set to obtain a cleaned data set, wherein the cleaning includes at least one of processing missing values, outliers, duplicate records, and inconsistent formats;
[0015] Performing feature extraction on the cleaned data set to obtain classification features, wherein the classification features include at least one of the user's personal information, the ambient sound type, and the volume setting;
[0016] Based on a clustering algorithm, the multiple users are clustered and grouped according to the personal information and the environmental sound type in the classification features to obtain at least two user groups and volume preference data corresponding to each user group;
[0017] Creating a data model corresponding to each user group, and setting corresponding weights for features in the data model based on the volume preference data;
[0018] Based on the cleaned data set, model training and verification are performed on the data model of the corresponding user group to obtain a verified data model to form a model library.
[0019] In combination with the first aspect, in some optional implementations, selecting a data model corresponding to the personal information from a pre-established model library as the volume prediction model includes:
[0020] Converting the personal information into a feature vector, wherein the personal information includes at least one of the age, gender, and region of the current user;
[0021] Based on the feature vector, determining the similarity between the current user and the user groups in the model library, wherein the model library records the correspondence between different user groups and corresponding data models;
[0022] The data model corresponding to the user group with the greatest similarity is selected from the model library to serve as the volume prediction model.
[0023] In combination with the first aspect, in some optional implementations, adjusting the volume of the sound output by the terminal device based on the predicted volume includes:
[0024] Determining whether the terminal device is in a preset scenario where volume adjustment is not allowed;
[0025] If the terminal device is not in the preset scene, adjusting the volume of the sound output by the terminal device to the predicted volume;
[0026] If the terminal device is in the preset scene, the volume of the sound output by the terminal device is not adjusted.
[0027] In combination with the first aspect, in some optional implementations, the method further includes:
[0028] Acquire first historical behavior data of the current user, where the first historical behavior data includes adjustment volume and environmental sound data corresponding to historical adjustment behaviors;
[0029] Based on the adjusted volume and the environmental sound data in the first historical behavior data, the volume prediction model is optimized to obtain an optimized volume prediction model.
[0030] In combination with the first aspect, in some optional implementations, the method further includes:
[0031] Acquire second historical behavior data corresponding to each type of user group, wherein the second historical behavior data includes volume adjustment and environmental sound data when corresponding users in the user group perform interactive behaviors;
[0032] Based on the adjusted volume and environmental sound data in the second historical behavior data, the data model corresponding to the user group in the model library is optimized to obtain an optimized data model.
[0033] In a second aspect, an embodiment of the present application further provides a volume adjustment method, which is applied to a server, and the method includes:
[0034] Receiving a download request sent by a terminal device;
[0035] Based on the download request, a volume prediction model corresponding to the current user's personal information is sent to the terminal device, and the volume prediction model is used to analyze and process the ambient sound data around the terminal device to obtain a predicted volume, and the predicted volume is used to adjust the volume of the sound output by the terminal device.
[0036] In conjunction with the second aspect, in some optional implementations, based on a download request sent by a terminal device, sending a volume prediction model corresponding to the personal information of the current user to the terminal device includes:
[0037] Selecting a data model corresponding to the personal information carried in the download request from a pre-established model library as the volume prediction model;
[0038] The volume prediction model is sent to the terminal device.
[0039] In conjunction with the second aspect, in some optional implementations, selecting a data model corresponding to the personal information carried in the download request from a pre-established model library as the volume prediction model includes:
[0040] Converting the personal information into a feature vector, wherein the personal information includes at least one of the age, gender, and region of the current user;
[0041] Based on the feature vector, determining the similarity between the current user and the user groups in the model library, wherein the model library records the correspondence between different user groups and corresponding data models;
[0042] The data model corresponding to the user group with the greatest similarity is selected from the model library to serve as the volume prediction model.
[0043] In conjunction with the second aspect, in some optional implementations, before sending the volume prediction model corresponding to the personal information of the current user to the terminal device based on the download request sent by the terminal device, the method further includes:
[0044] Acquire a data set of multiple users, the data set including personal information, environmental sound data, and adjusted volume corresponding to each user;
[0045] Cleaning the data set to obtain a cleaned data set, wherein the cleaning includes at least one of processing missing values, outliers, duplicate records, and inconsistent formats;
[0046] Performing feature extraction on the cleaned data set to obtain classification features, wherein the classification features include at least one of the user's personal information, the ambient sound type, and the volume setting;
[0047] Based on a clustering algorithm, the multiple users are clustered and grouped according to the personal information and the environmental sound type in the classification features to obtain at least two user groups and volume preference data corresponding to each user group;
[0048] Creating a data model corresponding to each user group, and setting corresponding weights for features in the data model based on the volume preference data;
[0049] Based on the cleaned data set, model training and verification are performed on the data model of the corresponding user group to obtain a verified data model to form a model library.
[0050] In conjunction with the second aspect, in some optional implementations, the method further includes:
[0051] Acquire second historical behavior data corresponding to each type of user group, wherein the second historical behavior data includes volume adjustment and environmental sound data when corresponding users in the user group perform interactive behaviors;
[0052] Based on the adjusted volume and environmental sound data in the second historical behavior data, the data model corresponding to the user group in the model library is optimized to obtain an optimized data model.
[0053] In a third aspect, an embodiment of the present application further provides an electronic device, comprising a processor and a memory coupled to each other, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the electronic device executes the above method.
[0054] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium. When the computer program is executed on a computer, the computer executes the above method.
[0055] The invention adopting the above technical solution has the following advantages:
[0056] In the technical solution provided by the present application, the ambient sound data around the terminal device is analyzed and processed using the volume prediction model corresponding to the personal information of the current user to obtain the predicted volume, so that the volume can be adaptively adjusted according to the ambient sound data. In addition, the volume prediction model is a model for the current user, so it is conducive to adjusting the volume in a personalized way for different users, and can improve the problem that the terminal device cannot automatically adjust the volume in a personalized way for different users. BRIEF DESCRIPTION OF THE DRAWINGS
[0057] The present application may be further described by the non-limiting embodiments given in the accompanying drawings. It should be understood that the following drawings only illustrate certain embodiments of the present application and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings may be obtained based on these drawings without creative effort.
[0058] Figure 1 This is one of the flow charts of the volume adjustment method provided in the embodiment of the present application.
[0059] Figure 2 The second flowchart of the volume adjustment method provided in the embodiment of the present application.
[0060] Figure 3 The third flowchart of the volume adjustment method provided in the embodiment of the present application.
[0061] Figure 4The fourth flowchart of the volume adjustment method provided in the embodiment of the present application. DETAILED DESCRIPTION
[0062] The present application will be described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that in the drawings or descriptions, similar or identical parts use the same figure numbers, and the implementation methods not shown or described in the drawings are forms known to ordinary technicians in the relevant technical field. In the description of this application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.
[0063] First embodiment
[0064] Please refer to Figure 1 The present application provides a volume adjustment method. Each step of the volume adjustment method can be executed by a device at one end, or interactively executed by devices at both ends. There is no specific restriction on the execution subject of each step of the method. The volume adjustment method can include the following steps:
[0065] Step 110, obtaining ambient sound data around the terminal device;
[0066] Step 120, analyzing and processing the environmental sound data based on a volume prediction model corresponding to the personal information of the current user to obtain a predicted volume;
[0067] Step 130: Adjust the volume of the sound output by the terminal device based on the predicted volume.
[0068] In this embodiment, the terminal device may be, but is not limited to, a smart phone, a multimedia device on a vehicle, etc. In this embodiment, the terminal device may have data processing capability, communication capability, and audio playback function.
[0069] For example, the terminal device is a multimedia device in a vehicle, and steps 110 to 130 may all be executed by an intelligent system in the vehicle.
[0070] For another example, the terminal device is a multimedia device on a vehicle, and step 110 is performed by the intelligent system on the vehicle side. Then, the ambient sound data around the terminal device obtained in step 110 is sent to the server on the cloud. After the server receives the ambient sound data, step 120 is performed, and the predicted volume obtained in step 120 is sent to the vehicle side. After the vehicle side receives the predicted volume, the intelligent system performs step 130 to realize automatic adjustment of the volume of the multimedia device on the vehicle.
[0071] This application takes the terminal device as a multimedia device on a vehicle as an example, and describes in detail the steps of the volume adjustment method by example, as follows:
[0072] Before step 110, the method may further include the step of selecting a volume prediction model that matches the current user from the model library. For example, before step 110, the method may further include:
[0073] Step 101, obtaining personal information of the current user;
[0074] Step 102: Select a data model corresponding to the personal information from a pre-established model library as the volume prediction model, wherein the model library stores at least two data models for volume prediction based on deep learning.
[0075] In this embodiment, when the user registers to use the APP (Application) corresponding to the vehicle, he or she can fill in the user's personal information through the touch screen on the smartphone or vehicle. Personal information may include but is not limited to age, gender and permanent residence. Personal information can be encrypted using an industry-standard encryption algorithm (such as the AES-256 algorithm) and stored in the service database after encryption to ensure the security of personal information during transmission and storage. In addition, the system of the service database will add an additional layer of verification through a security token or OTP (One Time Password) mechanism to prevent unauthorized access to the above personal information.
[0076] As an example, when a user uses the volume prediction function for the first time, the vehicle sends a download request to the server. The server legally obtains the current user's personal information from the service database based on the user information (such as the user account) carried in the download request. Then, the server selects a data model corresponding to the personal information from the model library as the volume prediction model. The data model can be flexibly selected according to actual conditions, and can be, but not limited to, a recurrent neural network model or a convolutional neural network model.
[0077] In this embodiment, step 102 may include:
[0078] Converting the personal information into a feature vector, wherein the personal information includes at least one of the age, gender, and region of the current user;
[0079] Based on the feature vector, determining the similarity between the current user and the user groups in the model library, wherein the model library records the correspondence between different user groups and corresponding data models;
[0080] The data model corresponding to the user group with the greatest similarity is selected from the model library to serve as the volume prediction model.
[0081] In this embodiment, the server stores feature data of each user group, and the feature data includes common features of the user group, such as data of the same region, the same age group, etc. By converting the user's personal information into a feature vector, and then calculating the similarity between the feature vector and the feature data of the user group, as the similarity between the current user and the user group, the user group with the greatest similarity to the current user can be determined, and the data model corresponding to the user group can be used as the volume prediction model of the current user. The similarity can be calculated by algorithms such as Euclidean distance and cosine similarity.
[0082] Before step 101, the method may further include the step of creating a model library. For example, before step 101, the method may further include:
[0083] Step 010, obtaining a data set of multiple users, wherein the data set includes personal information, environmental sound data, and adjusted volume corresponding to each user;
[0084] Step 020, cleaning the data set to obtain a cleaned data set, wherein the cleaning includes at least one of processing missing values, outliers, duplicate records and inconsistent formats;
[0085] Step 030, extracting features from the cleaned data set to obtain classification features, wherein the classification features include at least one of the user's personal information, the type of ambient sound, and the volume setting;
[0086] Step 040, based on a clustering algorithm, clustering and grouping the multiple users according to the personal information and the environmental sound type in the classification features, to obtain at least two user groups and volume preference data corresponding to each user group;
[0087] Step 050, creating a data model corresponding to each user group, and setting corresponding weights for features in the data model based on the volume preference data;
[0088] Step 060, based on the cleaned data set, the data model of the corresponding user group is trained and verified to obtain a verified data model to form a model library.
[0089] In this embodiment, in order to reduce the computational overhead on the vehicle side, steps 010 to 060 can be executed by the server. In step 010, the server can extract the user's personal information, environmental sound data, and volume adjustment data from the service database, and then clean the data. Among them, data cleaning may include but is not limited to: replacing missing values with mean values; deleting outliers, deleting duplicate records; deleting inconsistent data, or converting inconsistent data to form data in a unified format. Among them, the detection method of outliers is conventional, which will not be repeated here.
[0090] In step 020, by cleaning the data set, data annotation can be achieved, data consistency and reliability can be improved, and interference data can be reduced.
[0091] In step 030, the server can extract useful features from the cleaned data set, such as user age, region, ambient sound type, and volume setting.
[0092] The type of environmental sound can be identified by using a deep learning model, for example, the deep learning model groups the environmental sound data to distinguish different noise feature groups. As an example, the grouping may include, but is not limited to, different types such as traffic noise, human voice, and music.
[0093] In step 030, the extracted features can be converted into feature vectors or encoded. For example, for numerical features (such as age and volume), normalization can be used to form numerical feature vectors. Non-numerical features such as gender and region can be encoded using one-hot encoding to obtain encoded features.
[0094] In step 040, the clustering algorithm can be flexibly selected according to the actual situation, and is not specifically limited here. For example, the optimal number of clusters is determined by using the elbow rule, silhouette coefficient, etc. Then, based on the optimal number of clusters, based on the clustering algorithm, all users are grouped by personal information and environmental sound type, and the characteristic distribution of each cluster group is analyzed to identify the volume preference data of different groups. In addition, in the user group grouping obtained by clustering, possible outliers or abnormal groups can be identified for further cleaning.
[0095] In step 050, the data model is the above-mentioned recurrent neural network model, convolutional neural network model, etc., and the prediction model based on deep learning can predict the volume suitable for the output sound of the terminal device based on the environmental sound data. When configuring the feature parameters of the data model, different weights can be set for the features in the data model of different user groups according to the importance of the features, so as to realize personalized volume prediction.
[0096] In step 060, the data models of different user groups are trained, tested and verified using the environmental sound data, volume setting data and other data in the cleaned data set, so that a data model with volume prediction function can be obtained. The data models of all user groups form a model library.
[0097] In this embodiment, the environmental sound data may include, but is not limited to, the type, volume, frequency, amplitude, and other data of the environmental sound. The volume setting may include the volume set by the user. The trained data model may be cross-validated to evaluate the performance of the model to prevent overfitting of the model.
[0098] In step 110, the sensors on the vehicle may collect the sound data of the environment in which the vehicle is located at intervals or in real time to serve as the environmental sound data around the terminal device.
[0099] In addition, the collection of ambient sound data can automatically adjust the collection frequency according to the vehicle's usage mode and user behavior. For example, the collection frequency is higher in driving mode than in parking mode. The collected data can be stored in an encrypted local database and cleaned up regularly to optimize storage space.
[0100] In step 120, the intelligent system on the vehicle side can use the volume prediction model to analyze and process the environmental sound, so as to obtain a predicted volume suitable for the sound output by the terminal device in the current environment.
[0101] In this embodiment, step 130 may include:
[0102] Determining whether the terminal device is in a preset scenario where volume adjustment is not allowed;
[0103] If the terminal device is not in the preset scene, adjusting the volume of the sound output by the terminal device to the predicted volume;
[0104] If the terminal device is in the preset scene, the volume of the sound output by the terminal device is not adjusted.
[0105] Understandably, when the vehicle-side intelligent system is in a scenario where the navigation system suppresses multimedia, the voice system suppresses multimedia, etc., and the system setting does not allow the volume to be adjusted, the intelligent system will not automatically adjust the volume. If the intelligent system is not in a preset scenario, the terminal device is allowed to automatically adjust the volume based on the preset volume.
[0106] It can be understood that the intelligent system on the vehicle side can collect ambient sound data in real time and update the predicted volume in real time based on the ambient sound data. In this way, the volume of the sound output by the terminal device can be adaptively adjusted as the ambient sound data changes.
[0107] In this embodiment, the user can adjust the volume through the vehicle interface, square control, voice, etc., and form user interaction behavior data. The system records the vehicle volume data and environmental volume data in the relevant time period. The user interaction behavior data will be analyzed using cluster analysis or association rule mining to identify user behavior patterns and used to optimize the data model.
[0108] For example, see Figure 2 As an optional implementation, after step 130, the method may further include:
[0109] Step 140, obtaining first historical behavior data of the current user, wherein the first historical behavior data includes adjustment volume and environmental sound data corresponding to historical adjustment behaviors;
[0110] Step 150: Optimize the volume prediction model based on the adjustment volume and the environmental sound data in the first historical behavior data to obtain an optimized volume prediction model.
[0111] Understandably, the local volume prediction model on the vehicle side can be optimized and updated based on the historical behavior data of the user adjusting the volume. The update cycle can be fixed or triggered by the user. The update frequency and update mechanism can be dynamically adjusted according to user activities and system performance, and are not specifically limited here.
[0112] In this embodiment, the volume prediction model selected by the user in the cloud is downloaded to the local vehicle end, and the volume prediction model is used as the initial personalized volume adjustment model. The first historical behavior data is pre-processed data. The pre-processing can be the above-mentioned data cleaning process.
[0113] When optimizing the model, you can set a weight adjustment strategy to assign higher weights to local data (including volume adjustment and ambient sound data) to emphasize the personalized needs of local users. In addition, use incremental learning or fine-tuning strategies to use local data to adjust the local volume prediction model to adapt to the characteristics of local data. Use regularization algorithms (such as L1, L2, or elastic net regularization) to train and test models using local data to avoid overfitting. In addition, use a cross-validation algorithm to evaluate the performance of the model on local data to ensure the stability and generalization ability of the model.
[0114] If a user has multiple volume prediction models, model fusion can be used to combine the predictions of different models to improve the accuracy of the predicted volume.
[0115] On the vehicle side or smartphone side, a user interface can be provided to allow the user to access and control specified parameters of the volume prediction model, such as adjusting the learning rate, synchronization frequency, etc.
[0116] The cloud server can implement an automated version checking and updating mechanism to ensure that the data model downloaded locally by the vehicle is always up to date. The implementation method of the cloud data model update can be found in steps 160 to 170 below.
[0117] Please refer to Figure 3 As an optional implementation, after step 130, the method may further include:
[0118] Step 160, obtaining second historical behavior data corresponding to each type of user group, wherein the second historical behavior data includes volume adjustment and environmental sound data when corresponding users in the user group perform interactive behaviors;
[0119] Step 170: Based on the volume adjustment and environmental sound data in the second historical behavior data, the data model corresponding to the user group in the model library is optimized to obtain an optimized data model.
[0120] In this embodiment, the server may periodically or upon receiving a trigger instruction from a manager, execute step 160. The collected second historical behavior data needs to ensure compliance with privacy protection regulations and encrypt sensitive information.
[0121] For example, the historical behavior data of all users in a user group within a certain period of time is used as the second historical behavior data, and the second historical behavior data is the data after data cleaning. Then, a higher weight is assigned to the second historical behavior data to emphasize the personalized needs of the user group, and then the second historical behavior data is used to train, test and verify the data model of the user group, so as to obtain an optimized data model. It can be understood that when the second historical behavior data is used to optimize the cloud data model, the optimization method is similar to that of optimizing the local volume prediction model, which will not be repeated here.
[0122] As the number of users and user behavior data increases, based on the latest user data set, when the set cycle is met or triggered by the administrator, steps 010 to 060 are executed again. In this way, user groups can be added or deleted, and the data model of the corresponding user groups can be optimized.
[0123] Generally speaking, after the vehicle side downloads the volume prediction model, it will no longer download the volume prediction model from the cloud when predicting the volume in the future. In addition, the vehicle side can update the volume prediction model based on the user's operation behavior data (such as the manually adjusted volume and the ambient sound data during the adjustment).
[0124] In other implementations, the user may also use the latest data model in the cloud to replace the volume prediction model on the vehicle side. The latest data model refers to the latest data model of the user group to which the user belongs in the model library.
[0125] In this embodiment, the vehicle side, smart phone, etc., the relevant interactive interface can provide a function switch for the user to turn off or on the automatic volume adjustment function switch, adjust the volume sensitivity, and immediately update the local data model, replace the local data model, etc. If the automatic volume adjustment function switch is turned on, the corresponding steps of the method will be executed after the vehicle is started; if the function switch is turned off, the steps of the method will not be executed.
[0126] Based on the above design, this method can utilize the in-car ambient sound data to realize intelligent adjustment of the in-car volume, provide a new interactive form of controlling the volume through personalized settings in the car, and organically integrate various audio resources in the smart cockpit to enhance the overall user experience. While increasing the entertainment attributes of the cockpit, it can also meet users' needs for personalization, intelligence and safety.
[0127] Second embodiment
[0128] Please refer to Figure 4 The present application also provides a volume adjustment method for a server. The method can be applied to a server, and each step of the method can be executed or implemented by a server in the cloud. The volume adjustment method can include the following steps:
[0129] Step 210, receiving a download request sent by a terminal device;
[0130] Step 220, based on the download request, sending a volume prediction model corresponding to the current user's personal information to the terminal device, the volume prediction model is used to analyze and process the ambient sound data around the terminal device to obtain a predicted volume, and the predicted volume is used to adjust the volume of the sound output by the terminal device.
[0131] Wherein, step 220 may include:
[0132] Step 221, selecting a data model corresponding to the personal information carried in the download request from a pre-established model library as the volume prediction model;
[0133] Step 222: Send the volume prediction model to the terminal device.
[0134] Wherein, step 221 may include:
[0135] Converting the personal information into a feature vector, wherein the personal information includes at least one of the age, gender, and region of the current user;
[0136] Based on the feature vector, determining the similarity between the current user and the user groups in the model library, wherein the model library records the correspondence between different user groups and corresponding data models;
[0137] The data model corresponding to the user group with the greatest similarity is selected from the model library to serve as the volume prediction model.
[0138] In the second embodiment, before step 220, the method may further include:
[0139] Acquire a data set of multiple users, the data set including personal information, environmental sound data, and adjusted volume corresponding to each user;
[0140] Cleaning the data set to obtain a cleaned data set, wherein the cleaning includes at least one of processing missing values, outliers, duplicate records, and inconsistent formats;
[0141] Performing feature extraction on the cleaned data set to obtain classification features, wherein the classification features include at least one of the user's personal information, the ambient sound type, and the volume setting;
[0142] Based on a clustering algorithm, the multiple users are clustered and grouped according to the personal information and the environmental sound type in the classification features to obtain at least two user groups and volume preference data corresponding to each user group;
[0143] Creating a data model corresponding to each user group, and setting corresponding weights for features in the data model based on the volume preference data;
[0144] Based on the cleaned data set, model training and verification are performed on the data model of the corresponding user group to obtain a verified data model to form a model library.
[0145] After step 220, the method may further include:
[0146] Acquire second historical behavior data corresponding to each type of user group, wherein the second historical behavior data includes volume adjustment and environmental sound data when corresponding users in the user group perform interactive behaviors;
[0147] Based on the adjusted volume and environmental sound data in the second historical behavior data, the data model corresponding to the user group in the model library is optimized to obtain an optimized data model.
[0148] It can be understood that in the second embodiment, the implementation process of the server executing each step of the volume adjustment method is the same as or similar to that of the first embodiment, and will not be repeated here.
[0149] Third embodiment
[0150] The embodiment of the present application provides an electronic device, which may include a processor and a memory. The memory stores a computer program, and when the computer program is executed by the processor, the electronic device can perform corresponding steps in the following volume adjustment method.
[0151] In this embodiment, the processor may be an integrated circuit chip having the ability to process signals. For example, the processor may be a central processing unit (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and may implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of the present application.
[0152] The memory may be, but is not limited to, a random access memory, a read-only memory, a programmable read-only memory, an erasable programmable read-only memory, an electrically erasable programmable read-only memory, etc. In this embodiment, the storage module may be used to store a volume prediction model, environmental sound data, user operation behavior data, etc. Of course, the storage module may also be used to store a program, and the processing module executes the program after receiving the execution instruction.
[0153] 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 electronic device described above can refer to the corresponding process of each step in the aforementioned method, and will not be elaborated here.
[0154] Fourth embodiment
[0155] The embodiment of the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed on a computer, the computer executes the volume adjustment method described in the above embodiment.
[0156] Through the description of the above implementation methods, technical personnel in this field can clearly understand that the present application can be implemented by hardware, and can also be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for enabling a computer device (which can be a personal computer, an electronic device, or a network device, etc.) to execute the methods described in each implementation scenario of the present application.
[0157] In the embodiments provided in the present application, it should be understood that the disclosed equipment and methods can also be implemented in other ways. The equipment and method embodiments described above are merely schematic, for example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer program products according to the multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a part of a module, a program segment or a code, and a part of the module, program segment or code includes one or more executable instructions for implementing the specified logical functions. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of the boxes in the block diagram and / or the flowchart can be implemented by a dedicated hardware-based system that performs the specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions. In addition, each functional module in each embodiment of the present application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part.
[0158] The above description is only an embodiment of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims
1. A volume adjustment method, characterized in that: The method comprises: Obtain environmental sound data around the terminal device; Analyzing and processing the environmental sound data based on a volume prediction model corresponding to the personal information of the current user to obtain a predicted volume; Based on the predicted volume, the volume of the sound output by the terminal device is adjusted.
2. The method according to claim 1, characterized in that Before analyzing and processing the environmental sound data based on the volume prediction model corresponding to the personal information of the current user, the method further includes: Get the current user's personal information; A data model corresponding to the personal information is selected from a pre-established model library as the volume prediction model, wherein the model library stores at least two data models for volume prediction based on deep learning.
3. The method according to claim 2, characterized in that Before obtaining the personal information of the current user, the method further includes: Acquire a data set of multiple users, the data set including personal information, environmental sound data, and adjusted volume corresponding to each user; Cleaning the data set to obtain a cleaned data set, wherein the cleaning includes at least one of processing missing values, outliers, duplicate records, and inconsistent formats; Performing feature extraction on the cleaned data set to obtain classification features, wherein the classification features include at least one of the user's personal information, the ambient sound type, and the volume setting; Based on a clustering algorithm, the multiple users are clustered and grouped according to the personal information and the environmental sound type in the classification features to obtain at least two user groups and volume preference data corresponding to each user group; Creating a data model corresponding to each user group, and setting corresponding weights for features in the data model based on the volume preference data; Based on the cleaned data set, model training and verification are performed on the data model of the corresponding user group to obtain a verified data model to form a model library.
4. The method according to claim 2, characterized in that: Selecting a data model corresponding to the personal information from a pre-established model library as the volume prediction model includes: Converting the personal information into a feature vector, wherein the personal information includes at least one of the age, gender, and region of the current user; Based on the feature vector, determining the similarity between the current user and the user groups in the model library, wherein the model library records the correspondence between different user groups and corresponding data models; The data model corresponding to the user group with the greatest similarity is selected from the model library to serve as the volume prediction model.
5. The method according to claim 1, characterized in that Adjusting the volume of the sound output by the terminal device based on the predicted volume includes: Determining whether the terminal device is in a preset scenario where volume adjustment is not allowed; If the terminal device is not in the preset scene, adjusting the volume of the sound output by the terminal device to the predicted volume; If the terminal device is in the preset scene, the volume of the sound output by the terminal device is not adjusted.
6. The method according to claim 1, characterized in that The method further comprises: Acquire first historical behavior data of the current user, where the first historical behavior data includes adjustment volume and environmental sound data corresponding to historical adjustment behaviors; Based on the adjusted volume and the environmental sound data in the first historical behavior data, the volume prediction model is optimized to obtain an optimized volume prediction model.
7. The method according to claim 1, characterized in that The method further comprises: Acquire second historical behavior data corresponding to each type of user group, wherein the second historical behavior data includes volume adjustment and environmental sound data when corresponding users in the user group perform interactive behaviors; Based on the adjusted volume and environmental sound data in the second historical behavior data, the data model corresponding to the user group in the model library is optimized to obtain an optimized data model.
8. A volume adjustment method, characterized in that: Applied to a server, the method comprises: Receiving a download request sent by a terminal device; Based on the download request, a volume prediction model corresponding to the current user's personal information is sent to the terminal device, and the volume prediction model is used to analyze and process the ambient sound data around the terminal device to obtain a predicted volume, and the predicted volume is used to adjust the volume of the sound output by the terminal device.
9. The method according to claim 8, characterized in that Based on a download request sent by a terminal device, a volume prediction model corresponding to personal information of a current user is sent to the terminal device, including: Selecting a data model corresponding to the personal information carried in the download request from a pre-established model library as the volume prediction model; The volume prediction model is sent to the terminal device.
10. The method according to claim 9, characterized in that Selecting a data model corresponding to the personal information carried in the download request from a pre-established model library as the volume prediction model includes: Converting the personal information into a feature vector, wherein the personal information includes at least one of the age, gender, and region of the current user; Based on the feature vector, determining the similarity between the current user and the user groups in the model library, wherein the model library records the correspondence between different user groups and corresponding data models; The data model corresponding to the user group with the greatest similarity is selected from the model library to serve as the volume prediction model.
11. The method according to claim 8, characterized in that Before sending the volume prediction model corresponding to the personal information of the current user to the terminal device based on the download request sent by the terminal device, the method further includes: Acquire a data set of multiple users, the data set including personal information, environmental sound data, and adjusted volume corresponding to each user; Cleaning the data set to obtain a cleaned data set, wherein the cleaning includes at least one of processing missing values, outliers, duplicate records, and inconsistent formats; Performing feature extraction on the cleaned data set to obtain classification features, wherein the classification features include at least one of the user's personal information, the ambient sound type, and the volume setting; Based on a clustering algorithm, the multiple users are clustered and grouped according to the personal information and the environmental sound type in the classification features to obtain at least two user groups and volume preference data corresponding to each user group; Creating a data model corresponding to each user group, and setting corresponding weights for features in the data model based on the volume preference data; Based on the cleaned data set, model training and verification are performed on the data model of the corresponding user group to obtain a verified data model to form a model library.
12. The method according to claim 8, characterized in that The method further comprises: Acquire second historical behavior data corresponding to each type of user group, wherein the second historical behavior data includes volume adjustment and environmental sound data when corresponding users in the user group perform interactive behaviors; Based on the adjusted volume and environmental sound data in the second historical behavior data, the data model corresponding to the user group in the model library is optimized to obtain an optimized data model.
13. An electronic device, characterized in that: The electronic device includes a processor and a memory coupled to each other, wherein the memory stores a computer program. When the computer program is executed by the processor, the electronic device executes the method as described in any one of claims 1 to 7, or executes the method as described in any one of claims 8 to 12.
14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed on a computer, the computer is enabled to execute the method according to any one of claims 1 to 7, or execute the method according to any one of claims 8 to 12.
