User personalized content recommendation processing method and device based on listening to people

Abstract

The invention discloses a user personalized content recommendation processing method and device based on sound listening and person recognition. The method comprises the steps that collected sound information of a user in front of a smart television is recognized; when the user identity corresponding to the current sound information is identified as a new user, creating a new virtual user account; when the user identity corresponding to the current sound information is identified as the user with the existing user portrait, calling out personalized content or service corresponding to the user with the existing user portrait for first recommendation display; in the first recommendation display process, sound information of the current user is continuously collected, and the emotion of the current user is recognized; and further screening and filtering the content of the first recommendation display to obtain personalized content or service second recommendation display matched with the emotion of the current user. According to the invention, a new function is added to the smart television, and accurate recognition of different family members is realized through voice voiceprint recognition and mood recognition, so that contents and services more conforming to preferences are provided for the family members.

Description

Technical Field

[0001] The present invention relates to the field of smart television technology, and in particular to a method and device for user personalized content recommendation based on audio recognition, a smart television and a storage medium. Background Art

[0002] With the development of electronic technology and the continuous improvement of people's living standards, the use of various smart home appliances such as smart TVs is becoming more and more popular, and smart TVs have become an indispensable home appliance in people's lives.

[0003] As a family device shared by multiple people, TV is usually used by more than one user, including the elderly, young people, and young children. Different groups have very different needs for content. Internet TVs in the prior art have some personalized content recommendations, but they are usually unable to provide very accurate services to each user group, and cannot provide accurate personalized recommendations based on the user's mood, which is sometimes inconvenient for users.

[0004] Therefore, the existing technology still needs to be improved and developed. Summary of the invention

[0005] The technical problem to be solved by the present invention is that, in view of the problems and defects of the above-mentioned prior art, a method, device, smart TV and storage medium for user personalized content recommendation based on audio recognition are provided, aiming to solve the problem that Internet TV in the prior art cannot provide very accurate service to each user group and is sometimes inconvenient for users to use.

[0006] The technical solution adopted by the present invention to solve the problem is as follows:

[0007] A method for user personalized content recommendation based on identifying people by listening, comprising:

[0008] When the smart TV is detected to be started, the microphone set on the smart TV is used to collect the voice information of the user in front of the smart TV;

[0009] Through the voiceprint recognition algorithm, the collected voice information of the user in front of the smart TV is recognized, and the user identity corresponding to the current voice information is identified;

[0010] When the user identity corresponding to the current sound information is identified as a new user, a new virtual user account is created; and the operation habit data and playback data of the new user during use are recorded, and associated with the virtual user account, and a user profile is gradually constructed;

[0011] When it is identified that the user identity corresponding to the current sound information is a user with an existing user profile, personalized content or services corresponding to the user with the existing user profile are called out through a preset multi-user model, and a first recommendation display is performed;

[0012] During the first recommendation display process, the voice information of the current user is continuously collected, and the voice information of the current user is analyzed to identify the emotion of the current user;

[0013] According to the recognized emotion of the current user, the content of the first recommended display is further screened and filtered to obtain personalized content or services for adjusting the emotion of the current user, and a second recommended display is performed.

[0014] In the method for processing personalized content recommendations for users based on audio recognition, the preset multi-user model is constructed by the following steps:

[0015] Pre-collect and obtain the voice information of the user in front of the TV during the smart TV playback process, and the user's operating habits and playback data on the smart TV;

[0016] Based on the collected user's voice information, and the user's operating habits and playback data on smart TVs, a one-to-one correspondence is established with the user to build a user portrait;

[0017] Based on the constructed user portraits of multiple users, a recommendation algorithm is used to generate personalized recommendation content corresponding to the users of each user portrait;

[0018] The generated personalized recommendation content corresponding to the user of each user portrait is used to train the constructed multi-user model to obtain a preset multi-user model.

[0019] The method for processing personalized content recommendations for users based on audio recognition, wherein before the step of detecting that the smart TV is started and controlling the microphone provided on the smart TV to collect the voice information of the user in front of the smart TV, further comprises:

[0020] The voice information of the user in front of the smart TV is collected in advance, and the collected voice information is associated with the corresponding user identity and stored.

[0021] The method for processing personalized content recommendations for users based on audio recognition, wherein when the user identity corresponding to the current audio information is identified as a new user, the step of creating a new virtual user account comprises:

[0022] The new users include child users, elderly users, young users and multiple users;

[0023] Personalized recommendations for children are set in advance with animation and educational content as the priority;

[0024] Personalized recommendations for elderly users are set in advance with health and wellness and opera content as the priority;

[0025] The personalized recommendation content set in advance for multiple users is to recommend content that multiple users like to watch in proportion, or to recommend family-friendly content;

[0026] Pre-set for children users, the control provides differentiated services such as automatic switching of eye protection mode and timely reminder of rest;

[0027] The control provides differentiated services for elderly users by presetting the size of the screen text to a preset value and the volume to a preset threshold.

[0028] The method for processing personalized content recommendations for users based on voice recognition, wherein the steps of identifying the user's face information based on an image recognition algorithm and determining whether a new user is identified include:

[0029] Get the voice information of the user in front of the smart TV;

[0030] Identify whether the user is a new user based on the collected voice information of the user;

[0031] When a new user is identified, the sound information is further analyzed in combination with a data analysis algorithm to determine the age of the user in front of the smart TV, whether the user is a child user, an elderly user, or a young user.

[0032] The method for processing personalized content recommendation based on audio recognition, wherein the step of calling out personalized content or services corresponding to the user with the existing user portrait through a preset multi-user model and performing the first recommendation display comprises:

[0033] Control the recommendation priority of animation and educational content for personalized recommended content for child users;

[0034] Control the recommendation priority of personalized recommendation content for elderly users, such as health and wellness and opera content;

[0035] The personalized recommended content for multiple users is controlled to recommend content that multiple users like to watch in proportion, or to recommend family-friendly content.

[0036] The method for processing personalized content recommendation based on audio recognition, wherein the step of calling out personalized content or services corresponding to the user with the existing user portrait through a preset multi-user model and performing the first recommendation display comprises:

[0037] For children, the control provides differentiated services such as automatic switching to eye protection mode and timely reminders for rest;

[0038] For elderly users, the control provides differentiated services such as enlarging the screen text to a preset value and increasing the volume to a preset threshold.

[0039] A user personalized content recommendation processing device based on audio recognition, wherein the device comprises:

[0040] A data pre-collection module is used to pre-collect the voice information of the user in front of the smart TV, establish an association relationship between the collected voice information and the corresponding user identity, and store it;

[0041] The user model building module is used to collect and obtain the voice information of the user in front of the TV during the playback of the smart TV, and the user's operating habit data and playback data on the smart TV in advance; based on the collected user's voice information, and the user's operating habit data and playback data on the smart TV, establish a one-to-one correspondence with the user to build a user portrait; based on the constructed user portraits of multiple users, use a recommendation algorithm to generate personalized recommendation content corresponding to the user of each user portrait; and use the generated personalized recommendation content corresponding to the user of each user portrait to train the constructed multi-user model to obtain a preset multi-user model;

[0042] A sound collection module is used to detect that the smart TV is started and control the microphone set on the smart TV to collect the sound information of the user in front of the smart TV;

[0043] The recognition module is used to recognize the collected voice information of the user in front of the smart TV and identify the user identity corresponding to the current voice information through a voiceprint recognition algorithm;

[0044] A user portrait building module is used to create a new virtual user account when the user identity corresponding to the current sound information is identified as a new user; and record the operation habit data and playback data of the new user during use, establish an association with the virtual user account, and gradually build a user portrait;

[0045] The first recommendation display module is used to, when the user identity corresponding to the current sound information is identified as a user with an existing user profile, call out personalized content or services corresponding to the user with the existing user profile through a preset multi-user model, and perform the first recommendation display;

[0046] The emotion recognition module is used to continuously collect the voice information of the current user during the first recommendation display process, and analyze the continuously collected voice information of the current user to identify the emotion of the current user;

[0047] The second recommendation display module is used to further filter the content of the first recommendation display according to the recognized emotion of the current user, obtain personalized content or services for adjusting the emotion of the current user, and perform a second recommendation display.

[0048] A smart TV comprises a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, including the method for executing any one of the methods described above.

[0049] A non-transitory computer-readable storage medium, wherein when instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute any one of the methods described above.

[0050] Beneficial effects of the present invention: The present invention provides a method, device, smart TV and storage medium for user personalized content recommendation based on audio recognition. The present invention adds new functions to the smart TV: through voice print recognition and mood recognition, accurate identification of different family members can be achieved, thereby providing them with content and services that are more in line with their preferences; and providing convenience for users. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 It is a flowchart of a method for processing user personalized content recommendation based on audio recognition provided by an embodiment of the present invention.

[0053] Figure 2 It is a flowchart of a method for processing user personalized content recommendation based on audio recognition provided by an embodiment of the present invention.

[0054] Figure 3 It is a principle block diagram of an embodiment of a user personalized content recommendation processing device based on audio recognition provided by the present invention.

[0055] Figure 4 It is a block diagram of the internal structure principle of the smart TV provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0057] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0058] The present invention provides a method for user personalized content recommendation based on identifying people by listening to their voices; through voiceprint recognition of speech, the user's gender, age, emotions and other information can be quickly and accurately identified, and accurate identification of different family members can be achieved, thereby providing them with content and services that are more in line with their preferences. The accuracy of user identification has been greatly improved, providing convenience for users.

[0059] like Figure 1 As shown, a method for user personalized content recommendation based on audio recognition in Embodiment 1 of the present invention includes the following steps:

[0060] Step S100: Detecting that the smart TV is started, controlling the microphone set on the smart TV to collect the voice information of the user in front of the smart TV;

[0061] In the embodiment of the present invention, a smart TV is used as an example for explanation. In specific implementation, when the smart TV is turned on, a built-in microphone is automatically activated to capture the user's voice. This process relies on voice recognition technology to recognize the user's instructions or requests.

[0062] Specifically, the smart TV monitors its power-on status through built-in sensors or software. Once the startup signal is detected, the smart TV of the present invention is ready to receive the user's voice input. After the TV is started, the microphone is automatically activated. This means that the user does not need to operate manually, and the TV is ready to receive voice commands at any time. The microphone captures the user's voice information in real time. These voice data will be transmitted to the voice recognition system for processing. By analyzing the user's voice information, it is convenient to identify the user's identity in the subsequent steps.

[0063] Step S200: using a voiceprint recognition algorithm, the collected voice information of the user in front of the smart TV is recognized, and the user identity corresponding to the current voice information is recognized;

[0064] In this step, voiceprint recognition technology is used to confirm the user's identity. Voiceprint recognition is a biometric technology that analyzes the user's voice characteristics to verify the identity.

[0065] Specifically, the voiceprint recognition algorithm is used to create a unique voiceprint model by analyzing the frequency, pitch, timbre and other characteristics of the voice. Each person's voice has its own unique physiological characteristics, which can be extracted by the algorithm and used for identity recognition.

[0066] Then, when the user interacts with the smart TV, the present invention will collect the user's voice information in real time. These voice data will be transmitted to the voiceprint recognition system for processing. By comparing with the pre-stored voiceprint database, the present invention can identify the user identity corresponding to the current voice information. This process is usually completed within a few seconds, ensuring a smooth user experience.

[0067] For example, when multiple members of a family use the same smart TV, whenever a member walks in front of the TV and says "open my personal homepage", the present invention can confirm the identity of the member through the voiceprint recognition algorithm.

[0068] In another embodiment of the present invention, parents can set voiceprint recognition to restrict children from accessing certain applications or content. When children try to access these restricted contents, the system will recognize their voiceprints and deny access to ensure the security of the content.

[0069] The benefit of this step is that voiceprint recognition provides a secure authentication method, and only authenticated users can access specific content or functions. For example, parents can set up voiceprint recognition to ensure that only they can access adult content.

[0070] By identifying the user, smart TVs can provide personalized recommendations based on the viewing habits and preferences of different users. For example, each member of a family can get different program recommendations based on their preferences.

[0071] Users do not need to enter passwords or perform tedious identity verification. They can complete identity confirmation through voice, which improves the convenience of use.

[0072] It can be seen that identifying user voice information through voiceprint recognition algorithm not only improves the security and personalized services of smart TVs, but also provides users with a more convenient operating experience.

[0073] Step S300: When the user identity corresponding to the current sound information is identified as a new user, a new virtual user account is created; and the operation habit data and playback data of the new user during use are recorded, and associated with the virtual user account, and a user portrait is gradually constructed;

[0074] This step involves how the smart TV handles the identification of new users and the establishment of personalized services.

[0075] Specifically, when the smart TV confirms that the current user is a new user through the voiceprint recognition algorithm, the embodiment of the present invention will start the process of creating a virtual user account. This means that the user has not yet registered in the system or has no history. The system of the present invention will generate a virtual account for the new user as its identity on the smart TV. This account will be used to store the user's preference settings, viewing history and other related data.

[0076] And when a new user uses the smart TV, the present invention will automatically record his / her operating habits (such as commonly used applications, viewing time, etc.) and playback data (such as watched programs, ratings, etc.), which will be associated with the newly created virtual user account.

[0077] By continuously collecting and analyzing the operation habits and playback data of new users, the present invention can gradually build a profile of the user, including information such as the user's interests, preferences, and viewing habits, thereby laying the foundation for subsequent personalized recommendations.

[0078] For example, when a new user uses a smart TV for the first time at home, the present invention confirms that the user is a new user through voiceprint recognition and automatically creates a virtual account. The user starts watching a science fiction movie, and the system records this choice and the length of time it has been watched. In the following days, the user watches multiple programs of different types, and the embodiment of the present invention continues to record this data. After a period of time, the smart TV can recognize the user's preference for science fiction and action movies, and actively recommend related movies and TV series the next time the user turns on the TV.

[0079] Another example is that if a new user frequently uses certain specific applications (such as music streaming or games) during viewing, the present invention will record these habits and establish corresponding preference settings in the user's virtual account to provide more convenient access in future use.

[0080] The benefit of this step is that by creating a virtual user account and recording user data, the smart TV can provide personalized content recommendations and services to new users in the following steps. For example, the system can recommend related movies or programs based on the user's viewing history.

[0081] When a new user uses a smart TV for the first time, there is no need to go through a tedious registration process, and the system will automatically create an account for the new user, thereby improving the smoothness of the user experience. In addition, by analyzing the operating habits and playback data of new users, the smart TV can continuously optimize the recommendation algorithm to improve the relevance of content and user satisfaction.

[0082] It can be seen that in this step, by creating a virtual account for a new user and recording the data during its use, the smart TV can provide personalized services and experiences, gradually build a user portrait, and thus improve user satisfaction and loyalty.

[0083] Step S400: When it is identified that the user identity corresponding to the current sound information is a user with an existing user profile, personalized content or services corresponding to the user with the existing user profile are called out through a preset multi-user model, and a first recommendation display is performed;

[0084] This step embodiment involves how the smart TV provides personalized recommendations for registered users to enhance the viewing experience of the users.

[0085] Specifically, when the smart TV confirms the identity of the current user through voiceprint recognition technology, if the user's identity is identified to match the user profile that already exists in the system, the control will enter the personalized recommendation process.

[0086] In the embodiment of the present invention, a multi-user model is preset inside the smart TV, which can manage and distinguish the preferences and behavior data of different users, and helps the system to perform personalized analysis and recommendations for each user portrait.

[0087] The present invention retrieves content or services that match the user's preferences and habits based on the existing user profile, which may include the types of programs the user has watched in the past, commonly used applications, and the user's specific areas of interest.

[0088] In an embodiment of the present invention, when it is recognized that the user identity corresponding to the current sound information is a user with an existing user portrait, the smart TV will be controlled to immediately display personalized recommended content related to the user's personal portrait, so that the user can find his or her favorite programs or applications more conveniently.

[0089] For example, when a user has frequently watched family comedies and documentaries in the past few months, when he uses the smart TV again, the present invention confirms his identity through voiceprint recognition and retrieves his user portrait. Based on the historical viewing data of this user, the multi-user model of the smart TV of the present invention can recognize that he has a strong interest in humorous and educational programs. Therefore, the present invention immediately recommends a new family comedy series and a latest nature documentary as the first recommended content.

[0090] If the user subsequently chooses to watch the recommended family comedy, the smart TV will further record the feedback to optimize future recommendation results and make the recommendations more accurate and personalized. For example, if the user gives a positive review of a comedy, the present invention may add more programs of similar style or theme in future recommendations.

[0091] The benefit of the present invention is that through accurate personalized recommendations, users can find content of interest faster. This personalized experience not only improves satisfaction, but also enhances the stickiness between users and smart TVs. And users do not need to spend a lot of time searching or browsing irrelevant content. The present invention will quickly filter out recommended programs that suit their tastes, achieving a more efficient viewing experience. And personalized recommendations can effectively stimulate users' content consumption. The recommended content is highly relevant to the user's interests, increasing the possibility of users watching new programs or using new applications.

[0092] It can be seen that the present invention significantly improves the satisfaction and pleasure felt by users in the viewing experience through such personalized recommendations, while also promoting content consumption within the platform, forming a virtuous usage cycle.

[0093] In a specific implementation, the preset multi-user model of the present invention is constructed by the following steps:

[0094] S401, pre-collecting and obtaining the voice information of the user in front of the TV during the smart TV playback process, and the user's operating habit data and playback data on the smart TV;

[0095] S402: Based on the collected voice information of the user, and the user's operation habit data and playback data on the smart TV, a one-to-one correspondence is established with the user to construct a user portrait;

[0096] S403, based on the constructed user portraits of the multiple users, using a recommendation algorithm to generate personalized recommendation content corresponding to the users of the user portraits;

[0097] S404: Use the generated personalized recommendation content corresponding to the user of each user portrait to train the constructed multi-user model to obtain a preset multi-user model.

[0098] That is, in the embodiment of the present invention, when the user is watching a program, the smart TV will continuously collect sound information, the user's operating habits (for example, how the user uses the remote control, selects channels or applications), and playback data (which programs the user has watched, how long they have watched, etc.). By collecting this data, the system can obtain the user's preferences and behavior patterns.

[0099] After collecting enough data, the present invention matches existing users based on voice features and integrates each user's voice data, operating habits and viewing history to form a unique user profile. The user profile is a comprehensive description of the user's interests, preferences and behaviors. For example, a user may prefer comedy and music programs and tend to watch them at night.

[0100] In addition, the present invention utilizes the constructed user portraits, and the system uses recommendation algorithms (such as collaborative filtering, content recommendation, etc.) to generate personalized recommended content corresponding to each user portrait. This recommended content will be dynamically adjusted based on factors such as the user's historical viewing records, the behavior of similar users, and current popular trends.

[0101] The generated personalized recommendation content is not only fed back to the user, but also used for system learning and adaptation, and training multi-user models. By continuously iterating and optimizing the model, the present invention can provide more accurate recommendations in the future, improve user experience and overall system efficiency.

[0102] In this step, by gaining an in-depth understanding of the user's needs through the user's voice and operating habits, the smart TV can provide more accurate content recommendations that are in line with the user's interests, allowing the user to enjoy a more personalized viewing experience.

[0103] Step S500: During the first recommendation display process, continuously collect the voice information of the current user, analyze the continuously collected voice information of the current user, and identify the emotion of the current user;

[0104] In the embodiment of the present invention, it is related to how the smart TV further identifies the user's emotional state through continuous sound data, thereby optimizing the recommended content. This process not only improves the user experience, but also provides a deeper basis for personalized recommendations.

[0105] Specifically, when the user is watching the recommended content, after the first recommendation display (i.e., the initial recommendation display), the smart TV of the present invention will continue to collect the user's voice information through the built-in microphone. This includes the user's voice, tone, volume change, etc. This real-time collection can capture the user's immediate reaction during the viewing process, such as laughter, sighs, surprise and other emotional expressions.

[0106] The collected sound information will be transmitted to the analysis system, and the sound features will be processed using machine learning and sentiment analysis algorithms. The embodiment of the present invention will analyze the frequency, pitch, rhythm, etc. of the sound, and combine natural language processing technology to identify the user's emotional state, such as happiness, sadness, anger or boredom.

[0107] Based on this emotion recognition, the smart TV can give priority to pushing more soothing programs in the next recommendation, as described in step S600. At the same time, if Xiao Zhang shows disappointment at a certain moment (for example, sighing or lowering of tone), the system may quickly adjust the recommendation and provide some more attractive content, such as a newly released popular movie or an attractive documentary.

[0108] In this step, through emotion recognition, the smart TV can more accurately understand the user's current state, thereby providing content recommendations that better meet the user's emotional needs and enhance the personalized experience. When the recommended content can match the user's emotions, the user's satisfaction will be significantly improved, thereby enhancing the user's dependence on and loyalty to the smart TV. In addition, emotion recognition enables the recommendation system to adjust strategies in real time to avoid negative experiences caused by inappropriate recommended content and enhance the user's viewing pleasure.

[0109] In this way, smart TVs can not only respond to users' emotional changes in real time, but also make flexible adjustments in recommended content to ensure that users always have the best viewing experience. This emotion-driven recommendation mechanism significantly improves the quality of interaction and satisfaction between users and smart TVs.

[0110] Step S600: further filter the content recommended and displayed for the first time according to the identified emotion of the current user, obtain personalized content or service for adjusting the emotion of the current user, and perform a second recommendation and display.

[0111] This step is an important part of realizing emotional intelligence in the user's viewing experience of the smart TV. Through this process, the present invention can provide more targeted content recommendations according to the user's emotional state, thereby improving the user's satisfaction and viewing experience.

[0112] Specifically, when a user is watching content, the smart TV uses sound analysis technology to identify the user's emotional state. This emotional state may be happiness, sadness, anxiety, boredom, etc. The identified emotions provide an important basis for subsequent content screening.

[0113] In an embodiment of the present invention, the first recommended content is further screened according to the recognized emotion to select programs that are more suitable for the user's current emotional adjustment. For example, if the user's emotion is recognized as "anxiety", the present invention will filter out irritating or complicated content from the first recommended content. By analyzing the user's emotions, the present invention can select content that can help the user adjust his emotions, such as relaxing music, humorous variety shows, or relaxing nature documentaries.

[0114] The filtered content of the present invention will form a personalized recommendation list specifically targeting the current emotional needs of the user. This recommendation list will be used for the second recommendation display to ensure that the user can receive the content or service that best suits their emotional state.

[0115] For example, in the embodiment of the present invention, when the previous steps identify that user Xiao Zhang likes to watch horror movies, police and gangster movies, and comedy movies, the present invention recommends some horror movies, police and gangster movies, and comedy movies for the first time in the previous steps. If the user is watching, the smart TV captures Xiao Zhang's voice through a microphone, such as tired, stressed, and busy at work. The present invention identifies Xiao Zhang's emotion as "anxiety" through sound analysis. In this case, the smart TV will screen the horror movies, police and gangster movies, and comedy movies that are recommended for the first time, and exclude those tense plots that may increase Xiao Li's anxiety.

[0116] After screening, horror movies and police movies are filtered out, leaving comedy movies related movies to be recommended and displayed again by re-screening some soothing music. These contents can help Xiao Li relax and reduce anxiety. In this way, the present invention can recommend a light-hearted comedy movie, or play a soothing music to help Xiao Zhang gradually relax during the viewing process. In this way, the smart TV not only meets Xiao Zhang's viewing needs, but also provides support on the emotional level, improving the overall user experience.

[0117] It can be seen that the present invention, by providing content that matches the user's emotions, the smart TV can effectively help users regulate their emotions, relieve stress or improve their mood. And personalized recommendations make users feel that the smart TV pays attention to their emotional state, enhancing the interaction and participation between users and the device. And when users receive content that meets their emotional needs, they are more likely to continue watching, thereby increasing the content consumption rate and user retention rate. This personalized recommendation mechanism for emotional regulation enables smart TVs to better understand and respond to users' emotional needs and create a more humane viewing environment.

[0118] In summary, in a specific embodiment of the present invention, once the user's emotions are identified, the content of the first recommended display can be further filtered, screened and adjusted based on the emotional information to be a recommended program that matches the user's emotions. For example, if the user shows a happy emotion when watching the program that is recommended and displayed for the first time, the present invention will filter out more similar relaxing entertainment programs from the content that is recommended and displayed for the first time. On the contrary, if the user shows boredom or disappointment, the present invention will filter out different types of recommended content from the content that is recommended and displayed for the first time to attract the user's attention again.

[0119] For example, in the embodiment of the present invention, when the previous steps identify that the user Xiao Zhang likes to watch horror movies, police movies and comedy movies, the present invention will recommend and display some horror movies, police movies and comedy movies for the first time in the previous steps. If the user is watching, the smart TV captures Xiao Zhang's voice through the microphone, such as "tired", "stressed", "busy at work" and other voice word information. After analyzing these sound data, the present invention identifies that Xiao Zhang's emotion is "tired", and the present invention will filter out more similar relaxing entertainment programs from the content of the first recommendation display and recommend them to the user.

[0120] In another embodiment, before the specific implementation, the present invention will: collect the voice information of the user in front of the smart TV in advance, and establish an association relationship between the collected voice information and the corresponding user identity, and store it. Specifically, during the user's use of the smart TV, the user's voice information will be collected through the built-in microphone. These sound information include the user's voice, tone, emotional expression, etc. The collected sound information can be the user's comments, emotional reactions, or sounds of communication with other people when watching a program. The present invention will associate the collected sound information with the user's identity. This is usually achieved through the user's login or registration process on the smart TV. For example, when user Xiao Zhang logs in to the smart TV, Xiao Zhang's voice information will be bound to his account to form a complete user profile.

[0121] Then, the association between the collected sound information and the user's identity will be stored in the database of the smart TV. This data can be used for subsequent sentiment analysis and personalized recommendations. The stored data needs to follow the principle of privacy protection to ensure that the user's sound information will not be abused. By collecting sound information and associating it with the user's identity, the smart TV can better understand the user's preferences and emotional state, and provide data support for subsequent content recommendations. The pre-collected sound information can help the system build a user's emotional model, thereby improving the accuracy of emotion recognition. This enables the smart TV to respond to the user's emotional changes more accurately.

[0122] In another embodiment of the present invention, when the user identity corresponding to the current sound information is identified as a new user, the step of creating a new virtual user account further includes:

[0123] The new users include child users, elderly users, young users and multiple users;

[0124] That is, in the embodiment of the present invention, when the smart TV recognizes that the user is a new user through sound information, a new virtual user account will be automatically created. This process ensures that each user can enjoy personalized services. Then the user classification is performed, and the new user is further classified into child users, elderly users, young users and multiple users. Each category of users has different needs and preferences, so the system will provide corresponding personalized recommended content according to the user type.

[0125] In the embodiment of the present invention, personalized recommended content is set in advance for child users as the recommendation priority of animation and educational content; that is, in the embodiment of the present invention, for child users: animation and educational content are recommended first to meet the learning and entertainment needs of children. This recommendation method can help children acquire knowledge while watching and keep their interest.

[0126] The personalized recommended content set in advance for elderly users is the recommendation priority of health and wellness and opera content; in the embodiment of the present invention, the recommended content for elderly users will give priority to health and wellness and opera content. These contents can not only satisfy the interests of the elderly, but also help them obtain health-related information.

[0127] In the embodiment of the present invention, the recommended content for young users generally includes popular movies, music and social content to meet their lifestyle and entertainment needs.

[0128] The personalized recommended content set in advance for multiple users is to recommend content that multiple users like to watch in proportion, or recommend family-friendly content; when multiple users use the smart TV together, the present invention will recommend content that each user likes in proportion, or recommend family-friendly content, to ensure that the needs of each user are met.

[0129] The control provides differentiated services for child users in advance, such as automatic switching to eye protection mode and timely reminder of rest; in the embodiment of the present invention, differentiated services for child users: the system automatically switches to eye protection mode to protect children's eyesight. At the same time, a reminder of rest appears in time to encourage children to take a proper rest after watching for a period of time, thus promoting healthy viewing habits.

[0130] The control provides differentiated services for elderly users by presetting the size of the screen text to a predetermined value and the volume to a predetermined threshold. In the embodiment of the present invention, for differentiated services for elderly users, the system will enlarge the screen text to a predetermined value to ensure that they can read clearly. At the same time, the volume will also be increased to a predetermined threshold to help elderly users with poor hearing to better listen to the content.

[0131] In another embodiment of the present invention, it also includes the following steps:

[0132] S101, obtaining voice information of a user in front of a smart TV;

[0133] In this step, the smart TV collects the voice information of family members through the built-in microphone. The device actively monitors the user's speech and captures information such as speaking tone, timbre characteristics, and sound intensity. This information is collected to provide basic data for subsequent analysis to ensure that the real user can be accurately identified.

[0134] S102, identifying whether the user is a new user based on the collected voice information of the user;

[0135] In the embodiment of the present invention, after the sound information is obtained, the smart TV will use a machine learning algorithm to analyze the data to determine whether the user is a user already recorded in the system. If the system fails to find a record matching the current sound features in the stored database, the user is identified as a new user. This process relies on the algorithm's modal analysis of the sound data to ensure accurate user identification.

[0136] S103: When a new user is identified, the sound information is further analyzed in combination with a data analysis algorithm to determine the age of the user in front of the smart TV, whether the user is a child user, an elderly user, or a young user.

[0137] That is, in the embodiment of the present invention, once the user is confirmed to be a new user, that is, an external guest user other than a family user is identified, a further analysis algorithm will be used to extract information from various features of the sound to determine the user's age group. By analyzing the tone, audio frequency and pronunciation of the sound, the present invention can infer which age group the user belongs to (children, elderly, young people). This analysis not only relies on the statistical pattern of the sound features, but also may be combined with a small number of machine learning models, which can increase the accuracy of the judgment.

[0138] It can be seen that the accurate user age judgment in the embodiment of the present invention makes the recommended content more relevant, and users of different age groups can easily find programs they are interested in, which reduces the time spent by users in searching for programs.

[0139] In another embodiment of the present invention, it also includes the following steps:

[0140] S111, controlling the recommendation priority of personalized recommended content for child users to be animation and educational content;

[0141] In the embodiment of this step, if the smart TV recognizes that the user is a child, it will increase the recommendation priority of animation and educational content. This means that when a child user uses the TV, the present invention will prioritize the display of cartoons, educational programs and interactive learning content suitable for their age after identification to promote the intellectual and emotional development of the child. Further optimization will be performed based on the user's viewing preferences and viewing history to ensure that the recommended content is both interesting and beneficial.

[0142] S112, controlling the recommendation priority of personalized recommended content for elderly users to be health and wellness and opera content;

[0143] In this step, when the user is identified as an elderly person, the smart TV will adjust its recommendation algorithm and give priority to pushing content related to health and wellness and traditional opera. Elderly people usually have a higher interest and demand for health knowledge, health care methods, and cultural and artistic programs. The present invention analyzes the preferences of elderly users and recommends health lectures, health care guides, and various opera performances to help them obtain relevant information and enrich their cultural life.

[0144] S113, controlling the personalized recommended content for multiple users to recommend content that the multiple users like to watch in proportion, or recommending family-friendly content;

[0145] In the embodiment of the present invention, in a family, there are usually multiple users using the same smart TV. This step ensures that the smart TV can provide personalized recommendations based on the viewing preferences of each family member. For example, if there are children, young people and the elderly in the family, the system will recommend content in proportion to the interests of each member, or push programs suitable for the whole family to watch together, such as family movies and parent-child activity-related programs, to enhance the fun of watching together as a family.

[0146] S114. For a child user, the control provides differentiated services such as automatic switching to eye protection mode and timely reminder to take a rest;

[0147] In the embodiment of the present invention, in order to protect the eyesight of child users, the smart TV will automatically switch to the eye protection mode after identifying the child, and adjust the brightness, contrast and other parameters to reduce eye stimulation. At the same time, it will also regularly issue a rest reminder to encourage children to rest their eyes after watching for a certain period of time, develop healthy viewing habits, and reduce the risk of vision damage.

[0148] S115. For the user who is an elderly user, the control provides differentiated services such as enlarging the screen text to a predetermined value and increasing the volume to a predetermined threshold.

[0149] In the embodiment of the present invention, for the elderly group, the screen settings will be automatically adjusted after they are identified, and the text size will be enlarged to a level that is easy to recognize, ensuring the readability of the information. At the same time, the volume will be raised to a preset comfortable level to meet the auditory needs of elderly users. This measure significantly improves the user experience of the elderly, allowing them to enjoy watching TV more easily and comfortably.

[0150] It can be seen that the present invention provides personalized recommendations for users of different age groups, which improves the relevance and fun of viewing, allowing each user to have a caring viewing experience. It also provides eye protection services for children, reminding them to rest in time, which can effectively reduce the risk of visual impairment; and provides appropriate visual and auditory support for elderly users, which improves their viewing comfort. In addition, through family-friendly content and proportional recommendations, it strengthens the interaction and communication between family members, enhances the joyful atmosphere of family viewing, and increases the opportunity to enjoy entertainment together.

[0151] The present invention is further described in detail below through specific application examples.

[0152] Embodiment 2, as Figure 2 As shown, the second embodiment provides a method for user personalized content recommendation based on audio recognition, including the following steps:

[0153] Step S11: The user initiates a voice command and proceeds to step S12;

[0154] After the smart TV in the embodiment of the present invention is turned on, the user can directly issue a voice command, such as "Xiaoyue Xiaoyue, turn on XX video".

[0155] Step S12: determine whether the user is a new user based on the SR, if yes, proceed to step S13, if no, proceed to step S14;

[0156] The "SR" in "Determine whether the user is a new user based on voice SR" refers to "Speech Recognition". This step involves using speech recognition technology to determine whether the user is a new user;

[0157] In this specific embodiment, when a user interacts with the smart TV system through voice, the user's voice will be recorded and analyzed, and the characteristics of the user's voice, such as tone, speaking speed, accent, etc., will be extracted. These characteristics can help the system identify the user's identity. The extracted voice characteristics will be compared with the stored user voice characteristics. If there is no record matching the user's voice characteristics in the system, the system can determine that the user is a new user. Once the user is confirmed to be a new user, the system can assign a specific mark or label to it for subsequent personalized services or recommendations.

[0158] Step S13, create a new virtual user, and proceed to step S14;

[0159] In the embodiment of the present invention, when a new user is identified, a new virtual user profile is created for the user. The profile stores the basic information and preference settings of the user for subsequent use.

[0160] Step S14, setting the identified user as the current user, and proceeding to step S15;

[0161] In the embodiment of the present invention, whether it is a new user or an identified user, it will be set as the current operating user so that its behavior can be recorded and analyzed.

[0162] Step S15: other usage behaviors are associated with the current user, and the process proceeds to step S16;

[0163] In the embodiment of the present invention, various usage behaviors of the current user, such as programs watched, contents searched, etc., are monitored and these behaviors are connected with the current user profile, which helps to provide personalized services to the user later.

[0164] Step S16: forming a user portrait and proceeding to step S17;

[0165] In the embodiment of the present invention, a user portrait is generated based on the user's behavior data. This portrait includes information such as the user's interests and viewing habits, and is used to better understand the user's needs.

[0166] Step S17: providing personalized content or services based on existing user portraits;

[0167] In the embodiment of the present invention, the generated user portrait is used to analyze the user's preferences, so as to recommend personalized content or services to the user, so as to improve the user satisfaction and interactive experience.

[0168] Then, end the process.

[0169] It can be seen that in the embodiment of the present invention, the user identity is verified through voice recognition to ensure that only authorized users can access specific functions. And according to the voice input of new users, the response mode of the system is adjusted to improve their user experience.

[0170] Exemplary Devices

[0171] like Figure 3 As shown, an embodiment of the present invention provides a user personalized content recommendation processing device based on audio recognition, the device comprising:

[0172] The data pre-collection module 310 is used to collect the voice information of the user in front of the smart TV in advance, establish an association relationship between the collected voice information and the corresponding user identity, and store it;

[0173] The user model building module 320 is used to collect and acquire the voice information of the user in front of the TV during the playback of the smart TV, and the user's operating habit data and playback data on the smart TV in advance; based on the collected user's voice information, and the user's operating habit data and playback data on the smart TV, establish a one-to-one correspondence with the user to build a user portrait; based on the constructed user portraits of multiple users, use a recommendation algorithm to generate personalized recommendation content corresponding to the user of each user portrait; and use the generated personalized recommendation content corresponding to the user of each user portrait to train the constructed multi-user model to obtain a preset multi-user model;

[0174] The sound collection module 330 is used to detect that the smart TV is started and control the microphone set on the smart TV to collect the sound information of the user in front of the smart TV;

[0175] The recognition module 340 is used to recognize the collected voice information of the user in front of the smart TV and identify the user identity corresponding to the current voice information through a voiceprint recognition algorithm;

[0176] The user portrait building module 350 is used to create a new virtual user account when the user identity corresponding to the current sound information is identified as a new user; and record the operation habit data and playback data of the new user during use, establish an association with the virtual user account, and gradually build a user portrait;

[0177] The first recommendation display module 360 ​​is used to, when the user identity corresponding to the current sound information is identified as a user with an existing user profile, call out personalized content or services corresponding to the user with the existing user profile through a preset multi-user model, and perform a first recommendation display;

[0178] The emotion recognition module 370 is used to continuously collect the voice information of the current user during the first recommendation display process, and analyze the continuously collected voice information of the current user to identify the emotion of the current user;

[0179] The second recommendation display module 380 is used to further filter the content of the first recommendation display according to the identified emotion of the current user, obtain personalized content or services for adjusting the emotion of the current user, and perform a second recommendation display, as described above.

[0180] Based on the above embodiments, the present invention further provides a smart TV, whose principle block diagram can be shown as follows: Figure 4As shown. The smart TV includes a processor, a memory, a network interface, a display screen, and a database connected through a system bus. Among them, the processor of the smart TV is used to provide computing and control capabilities. The memory of the smart TV includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the smart TV is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a user personalized content recommendation processing method based on audio recognition is implemented. The database of the smart TV is used to store a user personalized content recommendation processing program based on audio recognition.

[0181] Those skilled in the art will understand that Figure 4 The principle block diagram shown in the figure is only a block diagram of a partial structure related to the solution of the present invention, and does not constitute a limitation on the smart TV to which the solution of the present invention is applied. A specific smart TV may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0182] In one embodiment, a smart TV is provided, comprising a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, and the one or more programs include instructions for performing the following operations:

[0183] When the smart TV is detected to be started, the microphone set on the smart TV is used to collect the voice information of the user in front of the smart TV;

[0184] Through the voiceprint recognition algorithm, the collected voice information of the user in front of the smart TV is recognized, and the user identity corresponding to the current voice information is identified;

[0185] When the user identity corresponding to the current sound information is identified as a new user, a new virtual user account is created; and the operation habit data and playback data of the new user during use are recorded, and associated with the virtual user account, and a user profile is gradually constructed;

[0186] When it is identified that the user identity corresponding to the current sound information is a user with an existing user profile, personalized content or services corresponding to the user with the existing user profile are called out through a preset multi-user model, and a first recommendation display is performed;

[0187] During the first recommendation display process, the voice information of the current user is continuously collected, and the voice information of the current user is analyzed to identify the emotion of the current user;

[0188] According to the recognized emotion of the current user, the content of the first recommended display is further screened and filtered to obtain personalized content or services for adjusting the emotion of the current user, and a second recommended display is performed, as described above.

[0189] The preset multi-user model is constructed by the following steps:

[0190] Pre-collect and obtain the voice information of the user in front of the TV during the smart TV playback process, and the user's operating habits and playback data on the smart TV;

[0191] Based on the collected user's voice information, and the user's operating habits and playback data on smart TVs, a one-to-one correspondence is established with the user to build a user portrait;

[0192] Based on the constructed user portraits of multiple users, a recommendation algorithm is used to generate personalized recommendation content corresponding to the users of each user portrait;

[0193] The generated personalized recommendation content corresponding to the user of each user portrait is used to train the constructed multi-user model to obtain a preset multi-user model.

[0194] The method for processing personalized content recommendations for users based on audio recognition, wherein before the step of detecting that the smart TV is started and controlling the microphone provided on the smart TV to collect the voice information of the user in front of the smart TV, further comprises:

[0195] The voice information of the user in front of the smart TV is collected in advance, and the collected voice information is associated with the corresponding user identity and stored.

[0196] Wherein, when the user identity corresponding to the current sound information is identified as a new user, the step of creating a new virtual user account includes:

[0197] The new users include child users, elderly users, young users and multiple users;

[0198] Personalized recommendations for children are set in advance with animation and educational content as the priority;

[0199] Personalized recommendations for elderly users are set in advance with health and wellness and opera content as the priority;

[0200] The personalized recommendation content set in advance for multiple users is to recommend content that multiple users like to watch in proportion, or to recommend family-friendly content;

[0201] Pre-set for children users, the control provides differentiated services such as automatic switching of eye protection mode and timely reminder of rest;

[0202] The control provides differentiated services for elderly users by presetting the size of the screen text to a preset value and the volume to a preset threshold.

[0203] The step of identifying the user's face information based on an image recognition algorithm and determining whether a new user is identified includes:

[0204] Get the voice information of the user in front of the smart TV;

[0205] Identify whether the user is a new user based on the collected voice information of the user;

[0206] When a new user is identified, the sound information is further analyzed in combination with a data analysis algorithm to determine the age of the user in front of the smart TV, whether the user is a child user, an elderly user, or a young user.

[0207] The step of calling out personalized content or services corresponding to the user with the existing user portrait through a preset multi-user model and performing a first recommendation display includes:

[0208] Control the recommendation priority of animation and educational content for personalized recommended content for child users;

[0209] Control the recommendation priority of personalized recommendation content for elderly users, such as health and wellness and opera content;

[0210] The personalized recommended content for multiple users is controlled to recommend content that multiple users like to watch in proportion, or to recommend family-friendly content.

[0211] The step of calling out personalized content or services corresponding to the user with the existing user portrait through a preset multi-user model and performing a first recommendation display includes:

[0212] For children, the control provides differentiated services such as automatic switching to eye protection mode and timely reminders for rest;

[0213] For the user who is an elderly user, the control provides differentiated services such as enlarging the screen text to a predetermined value and increasing the volume to a predetermined threshold, as described above.

[0214] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0215] In summary, the present invention provides a method, device, smart TV and storage medium for user personalized content recommendation based on audio recognition. The present invention adds new functions to the smart TV: through voice print recognition and mood recognition, accurate identification of different family members can be achieved, thereby providing them with content and services that are more in line with their preferences; and it provides convenience for users.

Claims

1. A method for user personalized content recommendation based on voice recognition, characterized in that: include: When the smart TV is detected to be started, the microphone set on the smart TV is used to collect the voice information of the user in front of the smart TV; Through the voiceprint recognition algorithm, the collected voice information of the user in front of the smart TV is recognized, and the user identity corresponding to the current voice information is identified; When the user identity corresponding to the current sound information is identified as a new user, a new virtual user account is created; And record the operation habit data and playback data of the new user during use, establish association with the virtual user account, and gradually build a user portrait; When it is identified that the user identity corresponding to the current sound information is a user with an existing user profile, personalized content or services corresponding to the user with the existing user profile are called out through a preset multi-user model, and a first recommendation display is performed; During the first recommendation display process, the voice information of the current user is continuously collected, and the voice information of the current user is analyzed to identify the emotion of the current user; According to the identified emotion of the current user, the content of the first recommended display is further screened and filtered to obtain personalized content or services matching the emotion of the current user, and a second recommended display is performed.

2. The method for user personalized content recommendation based on audio recognition according to claim 1, characterized in that: The preset multi-user model is constructed by the following steps: Pre-collect and obtain the voice information of the user in front of the TV during the smart TV playback process, and the user's operating habits and playback data on the smart TV; Based on the collected user's voice information, and the user's operating habits and playback data on smart TVs, a one-to-one correspondence is established with the user to build a user portrait; Based on the constructed user portraits of multiple users, a recommendation algorithm is used to generate personalized recommendation content corresponding to the users of each user portrait; The generated personalized recommendation content corresponding to the user of each user portrait is used to train the constructed multi-user model to obtain a preset multi-user model.

3. The method for user personalized content recommendation based on audio recognition according to claim 1, characterized in that: Before the step of detecting that the smart TV is started and controlling the microphone provided on the smart TV to collect the voice information of the user in front of the smart TV, the step further includes: The voice information of the user in front of the smart TV is collected in advance, and the collected voice information is associated with the corresponding user identity and stored.

4. The method for user personalized content recommendation based on audio recognition according to claim 1, characterized in that: When the user identity corresponding to the current sound information is identified as a new user, the step of creating a new virtual user account includes: The new users include child users, elderly users, young users and multiple users; Personalized recommendations for children are set in advance with animation and educational content as the priority; Personalized recommendations for elderly users are set in advance with health and wellness and opera content as the priority; The personalized recommendation content set in advance for multiple users is to recommend content that multiple users like to watch in proportion, or to recommend family-friendly content; Pre-set for children users, the control provides differentiated services such as automatic switching of eye protection mode and timely reminder of rest; The control provides differentiated services for elderly users by presetting the size of the screen text to a preset value and the volume to a preset threshold.

5. The method for user personalized content recommendation based on audio recognition according to claim 4 is characterized in that: The step of identifying the user's face information based on the image recognition algorithm and determining whether a new user is identified includes: Get the voice information of the user in front of the smart TV; Identify whether the user is a new user based on the collected voice information of the user; When a new user is identified, the sound information is further analyzed in combination with a data analysis algorithm to determine the age of the user in front of the smart TV, whether the user is a child user, an elderly user, or a young user.

6. The method for user personalized content recommendation based on audio recognition according to claim 1, characterized in that: The step of calling out personalized content or services corresponding to the user with the existing user portrait through the preset multi-user model and performing the first recommendation display comprises: Control the recommendation priority of animation and educational content for personalized recommended content for child users; Control the recommendation priority of personalized recommendation content for elderly users, such as health and wellness and opera content; The personalized recommended content for multiple users is controlled to recommend content that multiple users like to watch in proportion, or to recommend family-friendly content.

7. The method for user personalized content recommendation based on audio recognition according to claim 1, characterized in that: The step of calling out personalized content or services corresponding to the user with the existing user portrait through the preset multi-user model and performing the first recommendation display comprises: For children, the control provides differentiated services such as automatic switching to eye protection mode and timely reminders for rest; For elderly users, the control provides differentiated services such as enlarging the screen text to a preset value and increasing the volume to a preset threshold.

8. A user personalized content recommendation processing device based on audio recognition, characterized in that: The device comprises: A data pre-collection module is used to pre-collect the voice information of the user in front of the smart TV, establish an association relationship between the collected voice information and the corresponding user identity, and store it; The user model building module is used to collect and obtain the voice information of the user in front of the TV during the playback of the smart TV, and the user's operating habit data and playback data on the smart TV in advance; based on the collected user's voice information, and the user's operating habit data and playback data on the smart TV, establish a one-to-one correspondence with the user to build a user portrait; based on the constructed user portraits of multiple users, use a recommendation algorithm to generate personalized recommendation content corresponding to the user of each user portrait; and use the generated personalized recommendation content corresponding to the user of each user portrait to train the constructed multi-user model to obtain a preset multi-user model; A sound collection module is used to detect that the smart TV is started and control the microphone set on the smart TV to collect the sound information of the user in front of the smart TV; The recognition module is used to recognize the collected voice information of the user in front of the smart TV and identify the user identity corresponding to the current voice information through a voiceprint recognition algorithm; A user portrait building module is used to create a new virtual user account when the user identity corresponding to the current sound information is identified as a new user; and record the operation habit data and playback data of the new user during use, establish an association with the virtual user account, and gradually build a user portrait; The first recommendation display module is used to, when the user identity corresponding to the current sound information is identified as a user with an existing user profile, call out personalized content or services corresponding to the user with the existing user profile through a preset multi-user model, and perform the first recommendation display; The emotion recognition module is used to continuously collect the voice information of the current user during the first recommendation display process, and analyze the continuously collected voice information of the current user to identify the emotion of the current user; The second recommendation display module is used to further filter the content of the first recommendation display according to the recognized emotion of the current user, obtain personalized content or services for adjusting the emotion of the current user, and perform a second recommendation display.

9. A smart TV, characterized in that: The device comprises a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, and the one or more programs include being used to execute the method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method as described in any one of claims 1 to 7.

Images