Television program recommendation method, storage medium and equipment

Abstract

The invention discloses a television program recommendation method, a storage medium and equipment, and relates to the technical field of smart home / smart home, and the television program recommendation method comprises the steps: obtaining physiological feature data of audiences in front of a television, the physiological feature data comprises data detected by a millimeter wave radar and / or face images of audiences shot by a camera; and determining a mentality change index of the audience based on the physiological feature data of the audience, and determining a recommended program based on the currently played program, the mentality change index of the audience and the program recommendation index of each candidate program in the plurality of candidate programs. According to the method, the mind state change of the user is detected in real time through the millimeter wave radar and the image acquisition equipment, accurate detection of the mind state change of audiences is realized, the accuracy of a program recommendation result is improved, and the use experience of the user is improved.

Description

Technical Field

[0001] This application relates to the field of smart home / smart family. Specifically, it relates to a method, a storage medium, and a device for recommending TV programs. Background Art

[0002] A TV is a common household entertainment device and an indispensable part of people's daily lives. It plays a dual role as a medium for information dissemination and a carrier for leisure and entertainment. By receiving signals from TV stations or streaming services, it can play a rich variety of content including news, movies, TV dramas, variety shows, sports events, and educational programs.

[0003] With the development of technology, smart TVs are equipped with a TV program recommendation function. This TV program recommendation function usually makes recommendations based on the user's viewing history to help users quickly find content that matches their interests from a vast amount of TV programs. However, the TV program recommendations based on the user's viewing history often have the problem of inaccurate recommendation results, thus resulting in a poor user experience. Summary of the Invention

[0004] This application provides a method, a storage medium, and a device for recommending TV programs to solve the technical problem that the accuracy of the recommendation results of TV program recommendations based on the user's viewing history is low, thus resulting in a poor user experience.

[0005] In a first aspect, this application provides a method for recommending TV programs, including:

[0006] Obtain physiological characteristic data of the audience in front of the TV, where the physiological characteristic data includes data detected by a millimeter-wave radar and / or a facial image of the audience captured by a camera;

[0007] Determine the mental state change index of the audience based on the physiological characteristic data;

[0008] Based on the currently playing program, the mental state change index, and the program recommendation index of each candidate program among a plurality of candidate programs, determine the recommended program, where the program recommendation index is determined based on the mental state change index of the key segments watched by historical audiences for the candidate programs, and the key segments are determined based on at least one or more of the mental state change index, the change information of the audience volume, and the program switching moment during the process of historical audiences watching the candidate programs.

[0009] Optionally, the physiological characteristic data includes heart rate data and breathing data detected by the millimeter-wave radar, and determining the mental state change index of the audience based on the physiological characteristic data includes:

[0010] Divide the difference between the heart rate data and the average heart rate data of the audience within a preset time period by the average heart rate data to obtain a heart rate change index;

[0011] Divide the difference between the respiration data and the average respiration data of the audience within a preset time period by the average respiration data to obtain a respiration change index;

[0012] Determine the mental state change index based on the heart rate change index and the respiration change index.

[0013] Optionally, the physiological characteristic data includes the facial image of the audience captured by a camera. Determining the mental state change index of the audience based on the physiological characteristic data includes:

[0014] Determine the emotion classification and emotion intensity of the audience according to the facial image of the audience;

[0015] Determine the mental state change index of the audience according to the emotion classification and emotion intensity corresponding to each of the multiple facial images of the audience.

[0016] Optionally, when there are multiple audiences, determining the mental state change index of the audience based on the physiological characteristic data includes:

[0017] Determine the mental state change index of each audience based on the physiological characteristic data of each audience;

[0018] Determine the age and / or gender of each audience based on the facial image of each audience, and determine the weight coefficient of each audience according to the age and / or gender of each audience;

[0019] Perform weighted averaging on the mental state change index of each audience according to the weight coefficient of each audience to obtain the mental state change index of the multiple audiences.

[0020] Optionally, determining the recommended program based on the currently played program, the mental state change index, the facial image of the audience, and the program recommendation index of each candidate program in multiple candidate programs includes:

[0021] Input the type of the currently played program, the mental state change index, and the facial image of the audience into a pre-trained recommendation model to obtain the recommended program type output by the recommendation model. The recommendation model is trained based on audience sample data, and the audience sample data includes program type sample data, mental state change sample data, facial image sample data, and recommended program type labels. The mental state change sample data is determined based on physiological characteristic sample data, and the physiological characteristic sample data is obtained by a millimeter-wave radar on a TV sample;

[0022] Determine the recommended program according to the recommended program type and the program recommendation index of each candidate program.

[0023] Optionally, the TV program recommendation method further includes:

[0024] For any one of the candidate programs, obtain the mental state change index of the historical audience in each time period during the process of watching the candidate program;

[0025] If the mental state change index of the historical audience in the first time period exceeds the preset mental state change index range, determine the segment corresponding to the first time period as the key segment.

[0026] Optionally, the TV program recommendation method further includes:

[0027] For any one of the candidate programs, obtain the change information of the audience volume in each time period during the process of watching the candidate program by the historical audience;

[0028] If the change information of the audience volume in the second time period exceeds the preset audience volume change range, determine the segment corresponding to the second time period as the key segment.

[0029] Optionally, the TV program recommendation method further includes:

[0030] For any one of the candidate programs, obtain the program switching moments during the process of watching the candidate program by the historical audience, and determine the switching volume of each time period according to the program switching moments;

[0031] If the switching volume in the third time period exceeds the preset switching volume threshold, determine the segment corresponding to the third time period as the key segment.

[0032] In a second aspect, the present application provides a TV program recommendation device, including:

[0033] An acquisition module, configured to acquire physiological characteristic data of the audience in front of the TV, where the physiological characteristic data includes data detected by a millimeter-wave radar and / or a facial image of the audience captured by a camera.

[0034] A determination module, configured to determine the mental state change index of the audience based on the physiological characteristic data.

[0035] The determining module is further configured to determine a recommended program based on the currently played program, the state of mind change index, and the program recommendation index of each candidate program among a plurality of candidate programs, where the program recommendation index is determined based on the state of mind change index of the key segments watched by historical viewers for the candidate programs, and the key segments are determined based on at least one or more of the state of mind change index, the change information of the audience volume, and the program switching time during the process of historical viewers watching the candidate programs.

[0036] Optionally, the TV program recommendation device further includes: a processing module.

[0037] The processing module is configured to divide the difference between the heart rate data and the average heart rate data of the viewer within a preset time period by the average heart rate data to obtain a heart rate change index.

[0038] The processing module is further configured to divide the difference between the respiration data and the average respiration data of the viewer within a preset time period by the average respiration data to obtain a respiration change index.

[0039] The determining module is further configured to determine the state of mind change index based on the heart rate change index and the respiration change index.

[0040] Optionally, the determining module is further configured to determine the emotion classification and emotion intensity of the viewer according to the facial image of the viewer.

[0041] The determining module is further configured to determine the state of mind change index of the viewer according to the emotion classification and emotion intensity of the viewer corresponding to each of the multiple facial images of the viewer.

[0042] Optionally, the determining module is further configured to determine the state of mind change index of each viewer based on the physiological characteristic data of each viewer.

[0043] The determining module is further configured to determine the age and / or gender of each viewer based on the facial image of each viewer, and determine the weight coefficient of each viewer according to the age and / or gender of each viewer.

[0044] The processing module is further configured to perform weighted averaging on the state of mind change index of each viewer according to the weight coefficient of each viewer to obtain the state of mind change index of the multiple viewers.

[0045] Optionally, the processing module is further configured to input the type of the currently played program, the mental state change index, and the facial image of the viewer into a pre-trained recommendation model to obtain a recommended program type output by the recommendation model, where the recommendation model is trained based on viewer sample data, and the viewer sample data includes program type sample data, mental state change sample data, facial image sample data, and recommended program type labels, the mental state change sample data is determined based on physiological feature sample data, and the physiological feature sample data is obtained by a millimeter-wave radar on a television sample.

[0046] The determining module is further configured to determine the recommended program according to the recommended program type and the program recommendation index of each candidate program.

[0047] Optionally, the obtaining module is further configured to, for any one of the candidate programs, obtain the mental state change index of the historical viewer during each time period when watching the candidate program.

[0048] If the mental state change index of the historical viewer in the first time period exceeds a preset mental state change index range, the determining module is further configured to determine the segment corresponding to the first time period as the key segment.

[0049] Optionally, the obtaining module is further configured to, for any one of the candidate programs, obtain the viewer volume change information of the historical viewer during each time period when watching the candidate program.

[0050] If the viewer volume change information in the second time period exceeds a preset viewer volume change range, the determining module is further configured to determine the segment corresponding to the second time period as the key segment.

[0051] Optionally, the obtaining module is further configured to, for any one of the candidate programs, obtain the program switching moment during the process of the historical viewer watching the candidate program.

[0052] The determining module is further configured to determine the switching amount of each time period according to the program switching moment.

[0053] If the switching amount in the third time period exceeds a preset switching amount threshold, the determining module is further configured to determine the segment corresponding to the third time period as the key segment.

[0054] In a third aspect, the present application provides a computer-readable storage medium, on which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, they are used to implement the television program recommendation method as described in the first aspect and various possible implementation manners of the first aspect.

[0055] Fourth aspect, the present application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0056] The memory stores computer-executable instructions;

[0057] The processor executes the computer-executable instructions stored in the memory to implement the TV program recommendation method as described in the first aspect and various possible implementation manners of the first aspect.

[0058] Fifth aspect, the present application provides a program product, including a computer program, and when the computer program is executed by a processor, it implements the TV program recommendation method as described above.

[0059] The TV program recommendation method, storage medium and device provided by the present application obtain physiological characteristic data of the viewers in front of the TV, where the physiological characteristic data includes data detected by a millimeter-wave radar and / or facial images of the viewers captured by a camera; determine the mental state change index of each viewer based on the physiological characteristic data of each viewer in front of the TV, and determine the age and / or gender of each viewer based on the facial image of each viewer, and determine the weight coefficient of each viewer according to the age and / or gender of each viewer, and perform weighted averaging on the mental state change index of each viewer according to the weight coefficient of each viewer to obtain the mental state change index of multiple viewers; input the type of the currently played program, the mental state change index of the viewers and the facial images of the viewers into a pre-trained recommendation model to obtain the recommended program type output by the recommendation model, and then determine the recommended program for the current time according to the recommended program type and the program recommendation index of each candidate program. This method realizes the accurate detection of the mental state changes of the viewers by real-time detecting the mental state changes of the users through a millimeter-wave radar and a camera device, provides personalized program recommendations for the viewers based on the preferences and real-time mental state changes of the viewers, improves the accuracy of the program recommendation results, and improves the user experience. Description of the Drawings

[0060] The drawings here are incorporated into the specification and constitute a part of this specification, showing the embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0061] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0062] Figure 1 It is a schematic diagram of the hardware environment of the TV program recommendation method provided by the present application;

[0063] Figure 2 Flow schematic of the TV program recommendation method provided by this application Figure 1 ;

[0064] Figure 3 Flow schematic of the TV program recommendation method provided by this application Figure 2 ;

[0065] Figure 4 Structural schematic diagram of the TV program recommendation device provided by this application;

[0066] Figure 5 Structural schematic diagram of the TV program recommendation equipment provided by this application. Specific implementation manners

[0067] In order to enable those skilled in the art to better understand the solutions of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0068] It should be noted that the terms "first", "second", etc. in the description and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of this application described here can be implemented in an order different from those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0069] According to one aspect of the embodiments of this application, a TV program recommendation method is provided. This TV program recommendation method is widely applied to whole-house intelligent digital control application scenarios such as Smart Home, smart home, smart home appliance ecosystem, and Intelligence House ecosystem. Optionally, in this embodiment, the above TV program recommendation method can be applied to, for example Figure 1 the hardware environment composed of the terminal device 102 and the server 104 as shown Figure 1As shown, the server 104 is connected to the terminal device 102 via a network and can be used to provide services (such as application services, etc.) for the terminal or the client installed on the terminal. A database can be set up on the server or independently of the server to provide data storage services for the server 104. Cloud computing and / or edge computing services can be configured on the server or independently of the server to provide data computing services for the server 104.

[0070] The above network can include but is not limited to at least one of the following: wired network, wireless network. The above wired network can include but is not limited to at least one of the following: wide area network, metropolitan area network, local area network. The above wireless network can include but is not limited to at least one of the following: WIFI (Wireless Fidelity), Bluetooth. The terminal device 102 is not limited to being a PC, mobile phone, tablet computer, smart air conditioner, smart range hood, smart refrigerator, smart oven, smart stove, smart washing machine, smart water heater, smart washing equipment, smart dishwasher, smart projection device, smart TV, smart clothes dryer, smart curtain, smart audio and video, smart socket, smart speaker, smart sound box, smart fresh air device, smart kitchen and bathroom equipment, smart bathroom equipment, smart floor cleaning robot, smart window cleaning robot, smart floor mopping robot, smart air purification device, smart steam box, smart microwave oven, smart kitchen water heater, smart purifier, smart water dispenser, smart door lock, etc.

[0071] Television is a common household entertainment device and an indispensable part of people's daily lives. It plays a dual role as a medium for information dissemination and a carrier for leisure and entertainment. By receiving signals from television stations or streaming services, it can play a rich variety of content including news, movies, TV dramas, variety shows, sports events, and educational programs.

[0072] With the continuous progress of technology, the functions of smart TVs are also constantly improving; among them, the TV program recommendation function has become a major highlight of smart TV functions. This TV program recommendation function is usually based on the user's viewing history and aims to help users quickly find content that matches their interests from a vast amount of TV programs, enhancing the user's viewing experience.

[0073] However, existing TV program recommendation systems mainly rely on the user's viewing history, and this method still has certain limitations in terms of personalization and accuracy; in addition, the accuracy rate of the recommendation results is relatively low and cannot fully meet the user's interests and needs, thus affecting the user's usage experience.

[0074] The method for recommending TV programs provided by this application aims to solve the above technical problems in the prior art. This method installs multiple millimeter-wave radars and camera devices on the TV, and uses the multiple millimeter-wave radars to detect the psychological changes and facial expressions of the audience in real time. By combining the psychological changes of the audience and the facial expression states, the current psychology of the audience is determined, and new programs that better match their current psychology are recommended to the audience. It also uses millimeter-wave radars to collect the feedback information of the audience during the program playback, and combines it with the TV program content for aggregation calculation to realize the timely update of the recommendation index of the TV program. This method realizes the accurate detection of the psychological changes of the audience through the millimeter-wave radar and camera devices. Based on the preferences and real-time psychological changes of the audience, personalized program recommendations are provided to the audience, improving the accuracy of the program recommendation results and the user experience.

[0075] The following uses specific embodiments to elaborate in detail on the technical solutions of this application and how the technical solutions of this application solve the above technical problems. These several specific embodiments below can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of this application will be described below in conjunction with the drawings.

[0076] Figure 2 Flow diagram of the TV program recommendation method provided by the embodiment of this application Figure 1 For example, the execution subject of this embodiment can be an intelligent cloud platform equipped with a program recommendation module. As Figure 2 shown, the TV program recommendation method provided by this embodiment includes:

[0077] S201: Obtain the physiological characteristic data of the audience in front of the TV.

[0078] Among them, the physiological characteristic data includes the data detected by the millimeter-wave radar and / or the facial images of the audience captured by the camera.

[0079] It can be understood that the physiological characteristic data detected by the millimeter-wave radar is used to reflect the physiological state of the audience, and the facial image is used to indicate the image file containing the visual information of the audience's face.

[0080] It should be noted that the execution subject of the embodiments of this application is an intelligent cloud platform, which can transmit data with an intelligent TV, and the intelligent cloud platform and the intelligent TV are in the same smart home Internet of Things; multiple millimeter-wave radars and image acquisition devices are arranged on the intelligent TV. Through the multiple millimeter-wave radars, real-time detection of the physiological state of the audience in front of the intelligent TV can be realized, and through the image acquisition device (which can be a camera, for example), real-time detection of the facial expression changes of the audience in front of the intelligent TV can be realized; the intelligent cloud platform can receive the data collected by the multiple millimeter-wave radars arranged on the intelligent TV and the image data obtained by the image acquisition device, and perform data analysis and processing on the obtained data information, so as to assist the intelligent cloud platform in recommending TV programs. This application does not impose special restrictions on the positions and quantities of the millimeter-wave radars and image acquisition devices on the intelligent TV.

[0081] In addition, the intelligent cloud platform can be set in the intelligent TV or in other intelligent devices in the same Internet of Things as the intelligent TV. This application does not impose special restrictions on the setting position of the intelligent cloud platform.

[0082] In some embodiments, the millimeter-wave radar can be installed at specific positions on the intelligent TV. For example, it is usually located at the top or both sides of the intelligent TV frame to better detect the body signals of the audience; the millimeter-wave radar has high-precision detection capabilities and can accurately detect minute physiological changes. For example, for the detection of the heart rate, its accuracy can reach within an error range of several times per minute; at the same time, the millimeter-wave radar also has the advantage of non-contact detection and will not cause any interference or discomfort to the user.

[0083] The multiple millimeter-wave radars arranged on the intelligent TV send millimeter-wave signals to the viewing area of the current intelligent TV, receive the reflected millimeter-wave signals, analyze and process the obtained millimeter-wave signals, obtain the physiological characteristic data of the audience in front of the intelligent TV, and capture the facial images of the audience in front of the intelligent TV through the camera.

[0084] It can be understood that the detection principle of the physiological characteristic data is as follows: the millimeter-wave signals emitted by the millimeter-wave radar will be reflected when they encounter the human body, and the frequency and phase of the reflected signals will be affected by human physiological activities. For example, when the human heart beats, it will cause minute vibrations on the body surface, and this vibration will cause a change in the frequency of the reflected signal. Therefore, the millimeter-wave radar can calculate the heart rate by detecting this frequency change.

[0085] Exemplarily, physiological characteristic data of the audience in front of the current smart TV is detected by multiple millimeter-wave radars. The physiological characteristic data specifically includes: "heartbeat frequency is 72 beats per minute, breathing frequency is 16 breaths per minute, and skin conductance response is 5 microsiemens"; the facial image of the audience captured by the camera can be the facial features and instant expressions of Audience A. This application does not impose special restrictions on the types of physiological characteristic data.

[0086] S202: Determine the mental state change index of the audience based on the physiological characteristic data of the audience.

[0087] Among them, the mental state change index is used to reflect the psychological change situation of the audience.

[0088] It can be understood that the mental state change index is an index used to quantify and describe the change of the psychological state of an individual or a group; the mental state change index is usually based on the measurement and analysis of emotions, attitudes or psychological states, and can help identify and understand the change of the psychological state of the audience over time or in different situations.

[0089] In some embodiments, the physiological characteristic data includes heart rate data and breathing data detected by millimeter-wave radars. The specific determination process of the mental state change index can be: divide the difference between the heart rate data and the average heart rate data of the audience within a preset time period by the average heart rate data to obtain the heart rate change index; divide the difference between the breathing data and the average breathing data of the audience within a preset time period by the average breathing data to obtain the breathing change index; determine the mental state change index based on the heart rate change index and the breathing change index.

[0090] Among them, the preset time period can be, for example, 5 days.

[0091] It can be understood that the average heart rate data and the average breathing data of the audience within the preset time period respectively refer to the average values of the heart rate and breathing frequency of the audience when watching the TV program played on the smart TV; specifically, the heart rate data can reflect the physiological excitement degree or relaxation state of the audience when watching the program. For example, a higher heart rate may indicate that the audience is excited or nervous about the program content, while a lower heart rate may indicate relaxation or boredom; the breathing data can further reflect the emotional state of the audience. For example, faster breathing may be related to tension, excitement or anxiety, while slower breathing is usually related to relaxation or concentration.

[0092] Exemplarily, the currently acquired physiological characteristic data specifically includes: heart rate data - 72 beats per minute, respiratory data - 16 breaths per minute; if the average physiological characteristic data of the viewers who watched the smart TV within 5 days is stored in the current smart cloud platform, specifically including: average heart rate data - 70 beats per minute, average respiratory data - 15 breaths per minute; calculate the difference between the determined heart rate data this time and the average heart rate data of the viewers within the preset time period, and then divide it by the average heart rate data to obtain a heart rate change index of 2.8%; divide the difference between the respiratory data and the average respiratory data of the viewers within the preset time period by the average respiratory data to obtain a respiratory change index of 6.7%; based on the heart rate change index and the respiratory change index, calculate the mental state change index, and this mental state change index can be 4.75%.

[0093] In some other embodiments, the physiological characteristic data includes the facial images of the viewers captured by the camera. At this time, the specific determination process of the mental state change index can be: determine the emotion classification and emotion intensity of the viewers according to the facial images of the viewers; determine the mental state change index of the viewers according to the emotion classification and emotion intensity corresponding to the respective multiple facial images of the viewers.

[0094] Among them, the emotion classification is used to distinguish different emotion types of the viewers. The emotion classification includes, for example: happiness, fear, and sadness; the emotion intensity is used to describe the strength degree of the emotion type.

[0095] It can be understood that for different TV programs, there are various types of emotion states feedback by the viewers, and there is an association relationship between the TV programs and the emotion classification of the viewers. Specifically, for any TV program, there are at least two corresponding preset emotion classifications. Therefore, it can be determined whether the emotion change of the current viewer is consistent with the preset emotion change corresponding to the current TV program according to the emotion classification of the viewers obtained in real time; if they are consistent, it indicates that the emotion change of the viewer conforms to the emotion change of watching the current TV program. At this time, it is necessary to determine whether the current TV program is suitable for the viewer to continue watching according to the real-time emotion intensity of the viewer, or determine the degree of attraction of the current TV program to the viewer. The present application does not make special restrictions on the determination method of the corresponding relationship between the TV program and the preset emotion classification.

[0096] Exemplarily, for the classification of the audience's emotions and the determination of the emotion intensity, a deep learning model can be used, specifically including: collecting a large number of facial image datasets labeled with emotion classifications (such as happy, sad, angry, surprised, frightened, disgusted, neutral, etc.), and preprocessing the labeled image datasets (such as normalization, cropping, enhancement, etc.) to improve the training effect of the model; using the preprocessed dataset above to train a deep learning model, especially a facial expression recognition model, so that the trained facial expression recognition model can learn the complex mapping relationship between facial features and emotional states; for the input audience facial image, the facial expression recognition model can output the emotion classification and the corresponding emotion intensity corresponding to the facial image. For example, the facial expression recognition model can output a probability distribution indicating the possibility that the facial image belongs to each preset emotion classification. Usually, the classification with the highest probability is selected as the recognition result, that is, the emotion classification corresponding to the facial image at that time, and the emotion intensity can be quantified by the probability value. The obtained recognition results can be: Audience A - happy, 4%; Audience B - surprised, 5%; Audience C - neutral, 4.3%.

[0097] It should be noted that in the process of calculating the mental state change index based on multiple physiological feature data, the mental state change index can be the sum of the change indexes corresponding to various physiological feature data, or the weights of different physiological features can be determined according to the influence degree of different physiological features on the psychological changes of the audience, and then weighted summation is performed to obtain the corresponding mental state change index. This application does not make special restrictions on the calculation method of the mental state change index based on multiple physiological feature data.

[0098] S203: Determine the recommended program based on the currently played program, the mental state change index of the audience, and the program recommendation indexes of each candidate program among multiple candidate programs.

[0099] Among them, the candidate programs are used to indicate the TV programs that conform to the current psychological changes of the audience, the program recommendation index is used to reflect the suitability of the TV program for the current audience, and the recommended program is used to indicate the TV program determined from multiple candidate programs and recommended for the user.

[0100] It can be understood that the program recommendation index is determined based on the mental state change index of the historical audience watching the key segments of the candidate programs, where the key segments are determined based on at least one or more of the mental state change index, the change information of the audience volume, and the program switching moment during the historical audience watching the candidate programs.

[0101] During the process of the smart TV playing TV programs for the user, based on the mental state change index of the audience, determine whether the emotional change of the current audience conforms to the historical emotional change of the currently played program.

[0102] When it is determined that the emotional change of the current audience conforms to the historical emotional change of the currently played program, determine the current program playing duration, and when the program playing duration is about to end, call the program recommendation list corresponding to the currently played program, and recommend to the audience multiple candidate programs that are of the same type as the currently played program and whose program recommendation index is higher than or equal to that of the currently played program according to the program recommendation index of each candidate program among the multiple candidate programs.

[0103] When it is determined that the emotional change of the current audience does not conform to the historical emotional change of the currently played program, re-determine the emotional change of the audience according to the user's mental state change index, and re-determine multiple candidate programs corresponding to the emotional change. Moreover, the multiple candidate programs are displayed to the audience in descending order of the program recommendation index on the display screen of the smart TV.

[0104] It can be understood that the intelligent cloud platform stores a large amount of user data and TV program data. By learning and analyzing these data, the correlation between the user's emotional change and the TV program can be determined. That is to say, different TV programs correspond to different mental state change indexes. According to the real-time obtained mental state change index, the preference degree of the audience for the currently played TV program can be determined; when the real-time obtained mental state change index does not match the mental state change index corresponding to the currently played TV program, it can be determined that the current audience does not like to watch this TV program; otherwise, the current audience likes to watch this TV program.

[0105] When the intelligent cloud platform recommends TV programs for the audience, it can be when the smart TV is first started, or when the currently played program is about to end, or when the emotional change of the audience does not match the currently played TV program; in addition, different TV programs correspond to different program recommendation indexes, and the program recommendation index can be used as the basis for the intelligent cloud platform to recommend TV programs for the audience. That is to say, recommend TV programs that conform to the current mental state change and have a relatively high program recommendation index for the audience. This application does not make special restrictions on the recommendation timing of TV programs.

[0106] Exemplarily, if the program currently being played on the smart TV is a reasoning variety show, and the change in the audience's mood corresponding to this program type is 4%, and the audience is generally in a tense and focused mood; one minute before the end of the recommendation of this program, the intelligent cloud platform re-recommends for the audience based on multiple candidate programs during the recommendation process of this reasoning variety show; if the current obtained mental state change index of the audience does not match the mental state change index of this reasoning variety show, it indicates that the current audience does not like to watch this program or is not interested in this program. At this time, based on the mental state change index obtained this time, a new recommended program is output for the user, and the recommended programs are the top 5 programs sorted from largest to smallest according to the program recommendation index of each candidate program among multiple candidate programs. This application does not impose special restrictions on the number of recommended programs output this time.

[0107] In some embodiments, when the intelligent cloud platform recommends TV programs for the audience, different program recommendation index thresholds can be set according to different types of recommended programs, so as to screen TV programs that can be recommended for the audience.

[0108] In some embodiments, as the display terminal of the entire TV program recommendation result and a partial data processing center in the TV program recommendation process, the smart TV has a high-definition display screen and can present various TV programs and system feedback information. It internally includes a signal receiving and processing module for receiving broadcast TV signals or network streaming media signals and converting them into playable images and sounds; at the same time, the smart TV also has interfaces for communicating with the millimeter-wave radar and the intelligent cloud platform, can quickly respond to the data transmission requests of the millimeter-wave radar, perform efficient data interaction with the cloud platform, receive the recommended program information returned by the intelligent cloud platform, and display the returned information obtained this time on the screen.

[0109] Exemplarily, when the millimeter-wave radar and the camera detect a change in the user's mental state, the smart TV can upload data to the cloud platform in an extremely short time and quickly switch the display content after the cloud platform returns the recommended program, providing a seamless viewing experience for the user.

[0110] The TV program recommendation method provided in this embodiment obtains the physiological characteristic data of the audience in front of the TV, and the physiological characteristic data includes the data detected by the millimeter-wave radar and / or the facial images of the audience captured by the camera; determines the mental state change index of the audience based on the physiological characteristic data of the audience, and determines the recommended program based on the currently played program, the mental state change index of the audience, and the program recommendation index of each candidate program among multiple candidate programs. This method realizes the accurate detection of the change in the audience's mental state by real-time detecting the change in the user's mental state through the millimeter-wave radar and the image acquisition device, improves the accuracy of the program recommendation result, and improves the user experience.

[0111] Figure 3 Flow schematic of the TV program recommendation method provided by the embodiment of the present application Figure 2 As shown in Figure 3 the figure, on the basis of the Figure 2 embodiment, the TV program recommendation method is described in detail. The TV program recommendation method shown in this embodiment includes:

[0112] S301: Obtain the physiological characteristic data of the audience in front of the TV.

[0113] Step S301 is similar to the above step S201 and will not be elaborated here.

[0114] In some embodiments, the specific process of using the millimeter-wave radar to detect the physiological characteristic data includes: obtaining the time difference and frequency offset value of the millimeter-wave radar transmitting signals and receiving signals in each direction; determining the number of audiences in front of the TV and the physiological characteristic data of each audience according to the time difference and frequency offset value.

[0115] It can be understood that the time difference refers to the time experienced by the millimeter-wave signal after being emitted by the millimeter-wave radar and reflected back by the audience; by measuring the time difference between the reflected signal and the transmitted signal, the distance between the millimeter-wave radar and the audience can be calculated, the positions of multiple audiences can be determined, and thus the number of audiences currently watching the program can be obtained; the frequency offset is caused by the Doppler effect, that is, when the target object moves relative to the millimeter-wave radar, the frequency of the reflected signal will change. Therefore, by obtaining the frequency offset value of a specific area through the millimeter-wave radar, the physiological characteristic data of the corresponding audience can be determined.

[0116] Obtain the time difference and frequency offset value of the millimeter-wave radar transmitting signals and receiving signals in each direction, analyze and process the obtained time difference, identify different reflection sources, that is, identify the number of audiences in front of the current smart TV, and based on the corresponding millimeter-wave signals, determine the frequency offset values of different audiences, so as to extract the physiological characteristic data of multiple audiences.

[0117] Exemplarily, multiple millimeter-wave radars installed on a smart TV send millimeter-wave signals into a pre-set area corresponding to each radar to detect the number of viewers in the preset area. That is, by analyzing the intensity and angle of the reflected signals, it can be determined that the reflected signals come from different human bodies. A stronger signal intensity may indicate a user closer to the radar, and different angles indicate users in different positions. By comprehensively analyzing this information, the number of viewers can be counted. After determining the number of viewers and their corresponding positions in front of the current smart TV, the frequency offset values of the corresponding millimeter-wave signals are extracted to determine the physiological characteristic data of different viewers. Specifically, when the millimeter-wave signals emitted by the millimeter-wave radar encounter a human body, they will be reflected, and the frequency and phase of the reflected signals will be affected by human physiological activities. For the detection of user physiological data, it is mainly to obtain physiological characteristic data such as heart rate and breathing rate by analyzing the changes in the frequency and phase of the reflected signals. For example, when detecting a human body, minute body movements (such as breathing and heartbeat) will cause subtle time differences and frequency offset changes. Based on the time differences and frequency offset values, the physiological characteristic data of the corresponding viewers are extracted, and the physiological characteristic data can be the breathing rate and heart rate.

[0118] S302: Determine the mental state change index of each viewer based on the physiological characteristic data of each viewer.

[0119] S303: Determine the age and / or gender of each viewer based on the facial image of each viewer, and determine the weight coefficient of each viewer according to the age and / or gender of each viewer.

[0120] S304: Perform weighted averaging on the mental state change index of each viewer according to the weight coefficient of each viewer to obtain the mental state change index of multiple viewers.

[0121] Among them, the weight coefficient is used to reflect the preference degrees of viewers in different age and gender groups for different program types.

[0122] It can be understood that when the smart TV plays different programs, the number of viewers and the viewer groups in front of the smart TV are different. According to different numbers of viewers and viewer groups, different calculation methods for the mental state change index can be adopted. For the situation where a single viewer watches a program, the corresponding mental state change index can be calculated only based on the mental state change of the viewer himself / herself. For the situation of multiple viewers, different weight values are set for multiple viewers according to the actual viewing group situation, and then combined with the mental state change indexes of different viewers for comprehensive calculation of the mental state change index.

[0123] In the case where there are multiple viewers, based on the physiological characteristic data of each viewer obtained currently, determine the mental state change index of each viewer respectively; then, according to the facial image of each viewer obtained currently, determine the age and / or gender of the corresponding viewer, and determine the weight coefficient of each viewer according to the age and / or gender of each viewer; perform weighted average on the mental state change indexes of multiple viewers according to the weight coefficient of each viewer to obtain the mental state change index reflecting the overall situation of multiple viewers watching the program currently.

[0124] Exemplarily, if there are 3 viewers in front of the current smart TV, and the mental state change indexes corresponding to different viewers are: Viewer A - 4%, Viewer B - 5%, Viewer C - 4.3%; according to the facial images of different viewers, the age and gender of the corresponding viewers can be determined, specifically: Viewer A - female, 45 years old; Viewer B - male, 45 years old; Viewer C - female, 5 years old, and thus the weight coefficients of different viewers are obtained as: 0.4, 0.3, 0.3 respectively, and the mental state change index reflecting the overall situation of the current viewers is calculated as 4.39%.

[0125] In some embodiments, the weight coefficients of different viewers can be determined according to the situation of the viewing group in the historical viewing record of the current smart TV; for example, in the historical viewing data of the current smart TV, Viewer A has the longest viewing duration, that is, Viewer A often uses the smart TV to watch programs, so when multiple viewers watch simultaneously, the weight coefficient set by Viewer A accounts for a relatively high proportion. The present application does not make special restrictions on the determination method of the weight coefficient.

[0126] S305: Input the type of the program being played currently, the mental state change index of the viewer, and the facial image of the viewer into a pre-trained recommendation model to obtain the recommended program type output by the recommendation model.

[0127] S306: Determine the recommended program according to the recommended program type and the program recommendation index of each candidate program.

[0128] Among them, the recommended program type includes, for example: comedy, suspense, reasoning, positive energy.

[0129] It can be understood that the recommendation model is trained based on viewer sample data, and the viewer sample data includes program type sample data, mental state change sample data, facial image sample data, and recommended program type labels. Among them, the mental state change sample data is determined based on physiological characteristic sample data, and the physiological characteristic sample data is obtained through a millimeter-wave radar on the TV sample; therefore, the recommendation model can re-output the recommended program type for the viewer based on the obtained mental state change index and facial image of the viewer.

[0130] Input the type of the currently playing program, the mental state change index of the audience, and the facial image of the audience into a pre-trained recommendation model. Based on the correlation relationships between the type of the currently playing program, the mental state change index, and the facial image and program types obtained through pre-training, the recommendation model determines the recommended program types that match the current mental state change of the audience. According to the recommended program types determined this time, combined with the program recommendation indices of each candidate program, a preset number of candidate programs are selected as the recommended programs for this time.

[0131] It can be understood that the program recommendation indices of each candidate program under different program types are stored in the intelligent cloud platform, and the program recommendation indices are determined based on the historical viewing records of the corresponding TV programs. When determining the recommended programs for this time, multiple candidate programs can be screened according to a preset recommendation index threshold, and multiple candidate programs that meet the screening conditions are selected and all the currently obtained candidate programs are output as recommended programs to the audience. It can also be determined in combination with a preset recommended quantity. Specifically, it can be: sort the multiple candidate programs under the currently determined recommended program type in descending order according to the program recommendation index, and select the preset recommended quantity of candidate programs ranked at the top as the recommended programs output to the audience for this time.

[0132] In the embodiments of the present application, the program recommendation index is determined based on the mental state change index when the audience watches the key segments of different candidate programs in the historical viewing records. Specifically, it can be: calculate the mental state change indices of the key segments of the same TV program, and the obtained comprehensive result is used as the program recommendation index of the TV program. Among them, the determination of the key segments is at least based on one or more of the mental state change index, the change information of the audience volume, and the program switching moment during the historical audience's viewing of the candidate program. The present application does not make special restrictions on the calculation of the program recommendation index.

[0133] In some embodiments, the determination of the key segment can be: for any candidate program, obtain the mental state change index of each time period during the historical audience's viewing of the candidate program. If the mental state change index of the historical audience in the first time period exceeds the preset mental state change index range, then the segment corresponding to the first time period is determined as the key segment.

[0134] Among them, the preset mental state change index range is used to indicate the prediction range of the program production party for the mental state change index of the audience.

[0135] It can be understood that the first time period refers to any time period in the candidate program, and the time length of the first time period is the same as the time length of the time period corresponding to the currently obtained historical mental state change index. Among them, the time length of the time period corresponding to the currently obtained mental state change index is preset, for example, it can be 5 minutes. In addition, different candidate programs and different time periods correspond to different preset mental state change index ranges.

[0136] Exemplarily, if the preset mental state change index range corresponding to the time period "13:00 - 13:05" in the reasoning variety show is 4% - 4.5%, and if the historical mental state change index of the audience during the corresponding time period in the historical viewing data is 5%, that is, the historical mental state change index exceeds the preset mental state change index range. At this time, the time period "13:00 - 13:05" is taken as a key segment. The present application does not impose special restrictions on the time length of the key segment.

[0137] In some other embodiments, the determination of the key segment can also be: for any candidate program, obtain the change information of the number of viewers in each time period during the historical viewing of the candidate program by the audience; if the change information of the number of viewers in the second time period exceeds the preset range of the number of viewers, then the segment corresponding to the second time period is determined as the key segment.

[0138] Among them, the change information of the number of viewers is used to indicate the change situation of the number of viewers watching the TV program.

[0139] It can be understood that the millimeter-wave radar can detect the number of viewers in front of the smart TV, and by analyzing the intensity and time difference of the reflected signals at different positions, determine the approximate position and number of viewers; and the change in the number of viewers can reflect the degree of attraction of different program segments to the audience. Therefore, by determining the change information of the number of viewers in different time periods, the preference degree of the audience for different segments can be determined.

[0140] Exemplarily, if the preset range of the change in the number of viewers corresponding to the time period "13:00 - 13:10" in the reasoning variety show is a decrease of 0 - 2 viewers, and if the change information of the number of viewers in the corresponding time period in the historical viewing data is a decrease of 3 viewers, and the total number of viewers in the corresponding historical viewing data is 4 viewers, it indicates that the current time period has a poor attraction to the audience. At this time, the time period "13:00 - 13:10" is taken as the key segment.

[0141] In some other embodiments, the determination of the key segment can also be: for any candidate program, obtain the program switching moments during the historical viewing of the candidate program by the audience, and determine the switching amount of each time period according to the program switching moments; if the switching amount in the third time period exceeds the preset switching amount threshold, then the segment corresponding to the third time period is determined as the key segment.

[0142] Among them, the program switching moment is used to indicate the time node when the audience actively changes the TV program, and the switching amount threshold can be 30% for example.

[0143] Exemplarily, for the agreed candidate program, there are multiple historical viewing records. Determine the number of switches, i.e., the switching volume, in the time period "13:00 - 13:05" in the multiple historical viewing records, and compare this switching volume with a preset switching volume threshold. If it exceeds the preset switching volume threshold, then determine the time period "13:00 - 13:05" as a key segment.

[0144] It can be understood that the determination of the key segment is carried out for multiple historical viewing records of the same TV program, which can comprehensively reflect the preference degrees for different segments of the TV program under different audience groups and different numbers of audiences.

[0145] The TV program recommendation method provided in this embodiment obtains the physiological characteristic data of the audience in front of the TV. The physiological characteristic data includes the data detected by a millimeter-wave radar and / or the facial images of the audience captured by a camera; determines the mental state change index of each audience based on the physiological characteristic data of each audience in front of the TV, and determines the age and / or gender of each audience based on the facial image of each audience, and determines the weight coefficient of each audience according to the age and / or gender of each audience, and performs weighted averaging on the mental state change index of each audience according to the weight coefficient of each audience to obtain the mental state change index of multiple audiences; inputs the type of the currently playing program, the mental state change index of the audience, and the facial image of the audience into a pre-trained recommendation model to obtain the recommended program type output by the recommendation model, and then determines the recommended program for this time according to the recommended program type and the program recommendation index of each candidate program. This method realizes the accurate detection of the mental state change of the audience by real-time detecting the mental state change of the user through a millimeter-wave radar and a camera device, provides personalized program recommendations for the audience based on the preferences and real-time mental state changes of the audience, improves the accuracy of the program recommendation result, and improves the user experience.

[0146] Figure 4 It is a schematic structural diagram of the TV program recommendation device provided in this application. As Figure 4 shown, this application provides a TV program recommendation device. The TV program recommendation device 400 includes:

[0147] An acquisition module 401, configured to acquire the physiological characteristic data of the audience in front of the TV, where the physiological characteristic data includes the data detected by a millimeter-wave radar and / or the facial images of the audience captured by a camera.

[0148] A determination module 402, configured to determine the mental state change index of the audience based on the physiological characteristic data.

[0149] The determining module 402 is further configured to determine a recommended program based on the currently played program, the mental state change index, and the program recommendation index of each candidate program among a plurality of candidate programs, where the program recommendation index is determined based on the mental state change index of the key segments of the historical audience watching the candidate program, and the key segments are determined based on at least one or more of the mental state change index, the change information of the audience volume, and the program switching time during the historical audience watching the candidate program.

[0150] Optionally, the TV program recommendation device further includes: a processing module 403.

[0151] The processing module 403 is configured to divide the difference between the heart rate data and the average heart rate data of the audience within a preset time period by the average heart rate data to obtain a heart rate change index.

[0152] The processing module 403 is further configured to divide the difference between the respiration data and the average respiration data of the audience within a preset time period by the average respiration data to obtain a respiration change index.

[0153] The determining module 402 is further configured to determine the mental state change index based on the heart rate change index and the respiration change index.

[0154] Optionally, the determining module 402 is further configured to determine the emotion classification and emotion intensity of the audience according to the facial image of the audience.

[0155] The determining module 402 is further configured to determine the mental state change index of the audience according to the emotion classification and emotion intensity of the audience corresponding to each of the multiple facial images of the audience.

[0156] Optionally, the determining module 402 is further configured to determine the mental state change index of each audience based on the physiological characteristic data of each audience.

[0157] The determining module 402 is further configured to determine the age and / or gender of each audience based on the facial image of each audience, and determine the weight coefficient of each audience according to the age and / or gender of each audience.

[0158] The processing module 403 is further configured to perform weighted averaging on the mental state change index of each audience according to the weight coefficient of each audience to obtain the mental state change index of the multiple audiences.

[0159] Optionally, the processing module 403 is further configured to input the type of the currently played program, the mental state change index, and the facial image of the viewer into a pre-trained recommendation model, and obtain a recommended program type output by the recommendation model, where the recommendation model is trained based on viewer sample data, and the viewer sample data includes program type sample data, mental state change sample data, facial image sample data, and recommended program type labels, the mental state change sample data is determined based on physiological feature sample data, and the physiological feature sample data is obtained by a millimeter wave radar on a TV sample.

[0160] The determining module 402 is further configured to determine the recommended program according to the recommended program type and the program recommendation index of each candidate program.

[0161] Optionally, the obtaining module 401 is further configured to, for any one of the candidate programs, obtain the mental state change index of the historical viewers at each time period during the process of the historical viewers watching the candidate program.

[0162] If the mental state change index of the historical viewers in the first time period exceeds a preset mental state change index range, the determining module 402 is further configured to determine the segment corresponding to the first time period as the key segment.

[0163] Optionally, the obtaining module 401 is further configured to, for any one of the candidate programs, obtain the viewer volume change information of the historical viewers at each time period during the process of the historical viewers watching the candidate program.

[0164] If the viewer volume change information in the second time period exceeds a preset viewer volume change range, the determining module 402 is further configured to determine the segment corresponding to the second time period as the key segment.

[0165] Optionally, the obtaining module 401 is further configured to, for any one of the candidate programs, obtain the program switching moment during the process of the historical viewers watching the candidate program.

[0166] The determining module 402 is further configured to determine the switching amount of each time period according to the program switching moment.

[0167] If the switching amount in the third time period exceeds a preset switching amount threshold, the determining module 402 is further configured to determine the segment corresponding to the third time period as the key segment.

[0168] Figure 5 This is a schematic structural diagram of a TV program recommendation device provided by the present application. As Figure 5 shown, the present application provides a TV program recommendation device, and the TV program recommendation device 500 includes: a receiver 501, a transmitter 502, a processor 503, and a memory 504.

[0169] A receiver 501 for receiving instructions and data;

[0170] A transmitter 502 for transmitting instructions and data;

[0171] A memory 504 for storing computer-executable instructions;

[0172] A processor 503 for executing the computer-executable instructions stored in the memory 504 to implement the respective steps performed by the television program recommendation method in the above embodiments. For details, reference can be made to the relevant descriptions in the foregoing embodiments of the television program recommendation method.

[0173] Optionally, the above memory 504 can be either independent or integrated with the processor 503.

[0174] When the memory 504 is independently provided, the electronic device further includes a bus for connecting the memory 504 and the processor 503.

[0175] For the implementation principle and technical effects of the electronic device provided in the embodiments of the present application, reference can be made to the foregoing embodiments, and details are not described herein again.

[0176] The embodiments of the present application further provide a computer-readable storage medium, in which computer-executable instructions are stored, and when the processor executes the computer-executable instructions, the method described in any of the foregoing embodiments is implemented.

[0177] The embodiments of the present application further provide a computer program product, including a computer program, and when the computer program is executed by the processor, the method described in any of the foregoing embodiments is implemented.

[0178] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there can be other division methods in actual implementation. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0179] The above integrated modules implemented in the form of software function modules can be stored in a computer-readable storage medium. The above software function modules are stored in a storage medium, including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute some steps of the methods described in the embodiments of the present application.

[0180] It should be understood that the above-mentioned processor may be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the method disclosed in combination with the application can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules in the processor. The memory may include high-speed RAM memory, and may also include non-volatile storage NVM, such as at least one disk memory, and may also be a USB flash drive, a mobile hard disk, a read-only memory, a magnetic disk or an optical disc, etc.

[0181] The above storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, magnetic disk or optical disc. The storage medium may be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0182] An exemplary storage medium is coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium may also be a component of the processor. The processor and the storage medium may be located in an Application Specific Integrated Circuit (ASIC). Of course, the processor and the storage medium may also exist as discrete components in an electronic device or a master control device.

[0183] It should be noted that in this article, the terms "including", "comprising" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including that element.

[0184] The serial numbers of the embodiments of the present application above are only for description and do not represent the superiority or inferiority of the embodiments.

[0185] Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of the present application.

[0186] The above are only the preferred embodiments of the present application. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present application, several improvements and refinements can still be made, and these improvements and refinements should also be regarded as the protection scope of the present application.

Claims

1. A method for recommending television programs, characterized in that: include: Acquiring physiological characteristic data of a viewer in front of the TV, wherein the physiological characteristic data includes data detected by a millimeter wave radar and / or a facial image of the viewer captured by a camera; Determining the audience's mentality change index based on the physiological characteristic data; A recommended program is determined based on a currently playing program, the mentality change index and a program recommendation index of each candidate program among a plurality of candidate programs, wherein the program recommendation index is determined based on the mentality change index of historical audiences when watching key segments of the candidate programs, and the key segments are determined based on at least one or more of the mentality change index of historical audiences when watching the candidate programs, information on changes in audience size, and the moment of switching programs.

2. The method according to claim 1, characterized in that: The physiological characteristic data includes heart rate data and breathing data detected by the millimeter wave radar, and determining the audience's mentality change index based on the physiological characteristic data includes: The difference between the heart rate data and the average heart rate data of the audience in a preset time period is divided by the average heart rate data to obtain a heart rate variation index; The difference between the breathing data and the average breathing data of the audience in a preset time period is divided by the average breathing data to obtain a breathing change index; The mentality change index is determined based on the heart rate change index and the breathing change index.

3. The method according to claim 1, characterized in that The physiological characteristic data includes a facial image of the audience captured by a camera, and determining the audience's mentality change index based on the physiological characteristic data includes: determining the emotion classification and emotion intensity of the audience according to the facial image of the audience; The audience's mentality change index is determined according to the emotion classification and emotion intensity of the audience corresponding to each of the plurality of facial images of the audience.

4. The method according to claim 1, characterized in that: In the case that the audience is a plurality of audiences, determining the audience's mentality change index based on the physiological characteristic data includes: Determining a mentality change index of each audience member based on the physiological characteristic data of each audience member; Determining the age and / or gender of each viewer based on the facial image of each viewer, and determining a weight coefficient of each viewer according to the age and / or gender of each viewer; The mentality change index of each audience is weighted averaged according to the weight coefficient of each audience to obtain the mentality change index of the multiple audiences.

5. The method according to any one of claims 1 to 4, characterized in that: The step of determining a recommended program based on the currently playing program, the mentality change index, the facial image of the viewer, and the program recommendation index of each candidate program among a plurality of candidate programs comprises: Inputting the type of the currently playing program, the mentality change index and the facial image of the audience into a pre-trained recommendation model to obtain a recommended program type output by the recommendation model, wherein the recommendation model is trained based on audience sample data, the audience sample data includes program type sample data, mentality change sample data, facial image sample data and recommended program type labels, the mentality change sample data is determined based on physiological feature sample data, and the physiological feature sample data is obtained by a millimeter wave radar on a television sample; The recommended program is determined according to the recommended program type and the program recommendation index of each candidate program.

6. The method according to any one of claims 1 to 4, characterized in that: Also includes: For any of the candidate programs, obtaining the mentality change index of historical viewers in each time period during the process of watching the candidate program; If the mentality change index of historical audiences in the first time period exceeds a preset mentality change index range, the segment corresponding to the first time period is determined as the key segment.

7. The method according to any one of claims 1 to 4, characterized in that: Also includes: For any of the candidate programs, obtaining information on changes in the number of viewers in each time period during the viewing of the candidate program by historical viewers; If the audience volume change information of the second time period exceeds the preset audience volume change range, the segment corresponding to the second time period is determined as the key segment.

8. The method according to any one of claims 1 to 4, characterized in that: Also includes: For any of the candidate programs, obtaining the program switching time when the historical audience watched the candidate program, and determining the switching amount in each time period according to the program switching time; If the switching amount in the third time period exceeds a preset switching amount threshold, the segment corresponding to the third time period is determined as the key segment.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method according to any one of claims 1 to 8 when executed.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 8 through the computer program.

Images