CN112333545B - Television content recommendation method, system, storage medium and smart television - Google Patents

Television content recommendation method, system, storage medium and smart television Download PDF

Info

Publication number
CN112333545B
CN112333545B CN201910698763.0A CN201910698763A CN112333545B CN 112333545 B CN112333545 B CN 112333545B CN 201910698763 A CN201910698763 A CN 201910698763A CN 112333545 B CN112333545 B CN 112333545B
Authority
CN
China
Prior art keywords
user
user type
information
television
probability
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910698763.0A
Other languages
Chinese (zh)
Other versions
CN112333545A (en
Inventor
王鑫
赵向军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
TCL Technology Group Co Ltd
Original Assignee
TCL Technology Group Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by TCL Technology Group Co Ltd filed Critical TCL Technology Group Co Ltd
Priority to CN201910698763.0A priority Critical patent/CN112333545B/en
Publication of CN112333545A publication Critical patent/CN112333545A/en
Application granted granted Critical
Publication of CN112333545B publication Critical patent/CN112333545B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
    • H04N21/4662—Learning process for intelligent management, e.g. learning user preferences for recommending movies characterized by learning algorithms
    • H04N21/4666—Learning process for intelligent management, e.g. learning user preferences for recommending movies characterized by learning algorithms using neural networks, e.g. processing the feedback provided by the user
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
    • G06F16/73—Querying
    • G06F16/735—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/161—Detection; Localisation; Normalisation
    • G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/168—Feature extraction; Face representation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/172—Classification, e.g. identification
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/02—Feature extraction for speech recognition; Selection of recognition unit
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
    • G10L15/063—Training
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/441—Acquiring end-user identification, e.g. using personal code sent by the remote control or by inserting a card
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/442—Monitoring of processes or resources, e.g. detecting the failure of a recording device, monitoring the downstream bandwidth, the number of times a movie has been viewed, the storage space available from the internal hard disk
    • H04N21/44213—Monitoring of end-user related data
    • H04N21/44218—Detecting physical presence or behaviour of the user, e.g. using sensors to detect if the user is leaving the room or changes his face expression during a TV program
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/443—OS processes, e.g. booting an STB, implementing a Java virtual machine in an STB or power management in an STB
    • H04N21/4436—Power management, e.g. shutting down unused components of the receiver
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
    • H04N21/4667—Processing of monitored end-user data, e.g. trend analysis based on the log file of viewer selections
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
    • H04N21/4668—Learning process for intelligent management, e.g. learning user preferences for recommending movies for recommending content, e.g. movies

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Signal Processing (AREA)
  • General Health & Medical Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Computational Linguistics (AREA)
  • Acoustics & Sound (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Artificial Intelligence (AREA)
  • Social Psychology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)

Abstract

The invention provides a television content recommendation method, a system, a storage medium and an intelligent television, which comprise the following steps: acquiring user information; inputting the user information into a trained deep learning model, and determining the user type of the user; and entering a corresponding television mode according to the user type, and recommending the television content corresponding to the user type based on the television mode corresponding to the user type. The invention can help users to intelligently screen proper television content and effectively avoid children from watching unhealthy television programs.

Description

Television content recommendation method, system, storage medium and smart television
Technical Field
The invention relates to the technical field of intelligent televisions, in particular to a television content recommendation method, a television content recommendation system, a storage medium and an intelligent television.
Background
In recent years, watching television programs as the most traditional way of home entertainment has still affected the lives of most people. With the development of internet technology and playing systems, television programs are more diversified and more wonderful. However, not all tv programs are suitable for children to watch, unhealthy tv programs may also bring harm to physical and mental health of children, and many parents have forced the children to get away from the tv screen by means of forced channel change, shutdown or criticizing education. However, children have poor self-control, television can be stolen while parents are away, and the existing television does not set exclusive television programs for children specifically, and parents do not know which programs are suitable for children to watch.
Disclosure of Invention
The embodiment of the invention provides a television content recommendation method, a system, a storage medium and an intelligent television, which aim to solve the problem that in the prior art, a television does not set exclusive television content for a specific user.
A first aspect of the present application provides a television content recommendation method, including:
acquiring user information;
inputting the user information into a trained deep learning model, and determining the user type of the user;
and entering a corresponding television mode according to the user type, and recommending the television content corresponding to the user type based on the television mode corresponding to the user type.
A second aspect of the present application provides a television content recommendation system comprising:
a user information acquisition unit for acquiring user information;
the user type determining unit is used for inputting the user information into the trained deep learning model and determining the user type of the user;
and the mode switching and content recommending unit is used for entering a corresponding television mode according to the user type so as to recommend the television content corresponding to the user type based on the television mode corresponding to the user type.
A third aspect of the present application provides a smart television, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to the first aspect when executing the computer program.
A fourth aspect of the present application provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the method of the first aspect as described above.
A fifth aspect of the application provides a computer program product comprising a computer program which, when executed by one or more processors, performs the steps of the method as described in the first aspect above.
In the embodiment of the invention, the user information is acquired and input into the trained deep learning model to determine the user type of the user, and then the user enters the corresponding television mode according to the user type, so that the television content corresponding to the user type is recommended based on the television mode corresponding to the user type.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.
Fig. 1 is a flowchart of an implementation of a television content recommendation method provided by an embodiment of the present invention;
fig. 2 is a flowchart of a specific implementation of the television content recommendation method S102 according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a speech recognition network model of a television content recommendation method according to an embodiment of the present invention;
fig. 4 is a schematic diagram of a face recognition network model of a television content recommendation method according to an embodiment of the present invention;
fig. 5 is a flowchart of a specific implementation of recommending, based on a television mode corresponding to the user type, television content corresponding to the user type according to an embodiment of the present invention;
fig. 6 is a block diagram of a television content recommendation system according to an embodiment of the present invention;
fig. 7 is a schematic diagram of an intelligent television provided in an embodiment of the present invention.
Detailed Description
In order to make the objects, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the embodiments described below are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 shows an implementation flow of a television content recommendation method provided by an embodiment of the present invention, where the method flow includes steps S101 to S103. The specific realization principle of each step is as follows:
s101: and acquiring user information.
Specifically, the user information includes voice feature information and face feature information. In the embodiment of the application, the television content recommendation method is operated on the smart television, the smart television is provided with the microphone and the camera, the voice characteristic information of the user is collected through the microphone, and the face characteristic information of the user is collected through the camera.
Optionally, before the step S101, a television wake-up instruction input by a user is obtained, and user information of the user is obtained based on the television wake-up instruction. In the embodiment of the application, a user inputs a television wake-up instruction by pressing a remote controller or a smart television key. The television awakening instruction is used for awakening the intelligent television in the dormant state, after the intelligent television is awakened, the microphone collects voice characteristic information of the user in real time, and the camera collects face characteristic information of the user in real time.
S102: and inputting the user information into the trained deep learning model, and determining the user type of the user.
In the embodiment of the application, the user information is input into a trained deep learning model, the user type probability of the user is output, and the user type of the user is determined according to the user type probability. Specifically, the user types of the users comprise children and non-children, and whether the user watching the television is a child is determined through the user information and the trained deep learning model.
As an embodiment of the present invention, as shown in fig. 2, the deep learning model includes a speech recognition network model and a face recognition network model, and the S102 specifically includes:
a1: and inputting the voice characteristic information into a trained voice recognition network model, and determining a first user type probability of the user. The first user type probability is the probability of determining the user type according to the voice recognition network model.
A2: and inputting the face feature information into a trained face recognition network model, and determining the second user type probability of the user. The second user type probability is the probability of determining the user type according to the face recognition network model.
A3: and determining the user type of the user according to the first user type probability and/or the second user type probability.
Optionally, if the first user type probability reaches a first preset user type probability threshold, it is determined that the user type corresponding to the first user type probability is a child, and if the first user type probability does not reach the first preset user type probability threshold, it is determined that the user type corresponding to the first user type probability is a non-child. For example, the first preset user type probability threshold is 0.5, if the first user type probability P1 is greater than or equal to 0.5, the user type corresponding to the first user type probability is a child, and if the first user type probability P1 is less than 0.5, the user type corresponding to the first user type probability is a non-child.
Optionally, if the second user type probability reaches a second preset user type probability threshold, it is determined that the user type corresponding to the second user type probability is a child, and if the second user type probability does not reach the second preset user type probability threshold, it is determined that the user corresponding to the second user type probability is a non-child. For example, the second preset user type probability threshold is 0.5, if the second user type probability P2 is greater than or equal to 0.5, the user type corresponding to the second user type probability is a child, and if the second user type probability P2 is less than 0.5, the user type corresponding to the second user type probability is a non-child.
Optionally, the weights of the first user type probability and the second user type probability are respectively preset, the first user type probability and the second user type probability are fused according to the first user type probability and the weight thereof and the second user type probability and the weight thereof to obtain a fused classification probability, and the user type of the user is determined according to the fused classification probability. The first user type probability P1 is weighted as epsilon, the second user type probability P2 is weighted as eta, and the fusion classification probability F ═ epsilon P1+ eta P2, e.g., epsilon ═ 0.5 and eta ═ 0.5. When F is larger than or equal to 0.5, determining that the user is a child; when F < 0.5, determining that the user is a non-child. In the embodiment of the application, the user type is determined by combining the voice characteristic information and the face characteristic information of the user, so that the accuracy of determining the user type can be improved.
Further, the speech recognition network model is shown in fig. 3, and in the embodiment of the present application, the training of the speech recognition network model specifically includes:
b1: and constructing a voice recognition network model, wherein the voice recognition network model comprises a coding layer, a feedforward type sequence memory network layer and a full connection layer.
B2: obtaining sample voice information, wherein the sample voice information has a user type label; the type tag is a child or a non-child;
b3: and in the coding layer, coding the sample voice information to obtain a sample voice vector, and transmitting the sample voice vector to the feedforward type sequence memory network layer. Specifically, the input sample speech information is encoded into a vector representation, and in the embodiment of the present application, the encoding algorithm used is a word2vec (word to vector) algorithm.
B4: and in the feedforward type sequence memory network layer, extracting the characteristics of the sample voice vector to obtain a sample voice characteristic vector, and transmitting the sample voice characteristic vector to the full connection layer. Specifically, the feedforward type sequence memory network layer is an FSMN-layer with 10 layers. The input of the part is a sample which is coded by a word2vec algorithmThe output is a sample speech feature vector. The FSMN-layer structure is that a first output result h of the l layer is obtained from coding layer data by a bidirectional GRUt lThen, a second output result P of the layer is obtained through the full connection of the single layert lA1 is to Pt lStoring in Memory block, then using bidirectional GRU to obtain output result h of l +1 layer from Memory blockt l+1Wherein h ist l、Pt lAnd ht l+1Parameters for recording intermediate results.
B5: and in the full connection layer, carrying out user type probability calculation on the sample voice information according to the sample voice feature vector to obtain a first user type probability of the sample voice information. In the embodiment of the application, the fully-connected layer is a fully-connected neural network layer with a softmax activation function, and the processed sample voice information is classified through the fully-connected neural network layer. In a usage phase of the model, a first user type probability P1 of the sample speech information to be returned softmax.
B6: and optimizing and adjusting model parameters of the voice recognition network model according to a preset parameter adjustment algorithm until the user type corresponding to the first user type probability of the sample voice information is consistent with the user type label of the sample voice information. Specifically, a back propagation algorithm is used for optimizing and adjusting model parameters of the voice recognition network model.
Specifically, in the embodiment of the present application, in the training phase, the input of the speech recognition network model is sample speech information, and the output label is a non-child or child. The method mainly comprises the steps of extracting characteristics of sample voice information of a user and recognizing the sample voice information by adopting a Deep learning model structure of Deep-FSMN (Deep Forward Skip Memory network) + FCNN (full Connected Neural network).
Further, the face recognition network model is shown in fig. 4, and in the embodiment of the present application, the training of the face recognition network model specifically includes:
c1: and constructing a face recognition network model, wherein the face recognition network model comprises a convolution layer, a pooling layer and a full-connection layer.
C2: obtaining sample face information, wherein the sample face information has a user type label; the user type tag is child or non-child.
C3: and in the convolutional layer, extracting the face features according to the sample face information to obtain the sample face features of the sample face information. In the embodiment of the application, the MTCNN algorithm is used for extracting the facial features. Specifically, face key points are detected, wherein the face key points comprise a left eye, a right eye, a nose tip, a left mouth corner and a right mouth corner. And acquiring sample face features of the sample face information according to the detected face key points.
C4: and in the full connection layer, according to the sample face features, carrying out user type probability calculation on the sample face information to obtain a second user type probability of the sample face information. In the embodiment of the application, the fully-connected layer is a fully-connected neural network layer with a softmax activation function, and the processed sample face information is classified through the fully-connected neural network layer. In the usage phase of the model, a second user type probability P2 of the sample face information of softmax is to be returned.
C5: and optimizing and adjusting the model parameters of the face recognition network model according to a preset parameter adjustment algorithm until the user type corresponding to the second user type probability of the sample face information is consistent with the type label of the sample face information. Specifically, a back propagation algorithm is used for optimizing and adjusting model parameters of the face recognition network model.
Specifically, in the embodiment of the present application, the MTCNN algorithm is used to extract the facial data features, and the MTCNN algorithm includes three stages: in the first stage, a full convolution neural network, namely P-Net, is adopted to obtain a candidate window and a boundary regression vector of the candidate window, meanwhile, the candidate window is calibrated according to the boundary regression vector, and then a large number of repeated candidate windows are removed by using non-maximum suppression (NMS); in the second stage, a more complex convolutional neural network, R-Net, is employed. Training the candidate window bodies subjected to P-Net in an R-Net network by the R-Net, then training in a full connection layer mode to eliminate a large number of candidate window bodies which do not meet the requirements, then finely adjusting the candidate window bodies by using a regression value of a boundary regression vector, and then removing an overlapped window body by using NMS; in the third stage, a more complex convolutional neural network, O-Net, is used. And obtaining the candidate window and the boundary regression vector of the candidate window by the O-Net, finely adjusting the candidate window by using the regression value of the boundary regression vector, and removing the overlapped window by using NMS (network management system). And simultaneously displaying five face key point positions while removing the overlapped candidate window. And finally, inputting the five face key point characteristics and other face characteristics into a full connection layer to obtain a second user type probability of face information classification.
In the embodiment of the application, in the training stage, the input of the face recognition network model is sample face information, and the output label is non-child or child. The extraction and recognition of the facial features of the user are completed by adopting an MTCNN (Multi-task masked Connected computational Networks) + FCNN (full Connected Neural Networks) deep learning model structure, and compared with the algorithm of the prior facial detection, the MTCNN has higher precision and higher facial feature extraction efficiency.
S103: and entering a corresponding television mode according to the user type, and recommending the television content corresponding to the user type based on the television mode corresponding to the user type.
In the embodiment of the application, a television mode corresponding to the user type is entered according to the user type. The television mode comprises a child television mode and a common television mode, the user type comprises children and non-children, and specifically, if the user is a child, the child television mode is entered; and if the user is a non-child, entering a common television mode.
In the embodiment of the present application, in the child television mode, the television content has been filtered, and television programs suitable for children to watch are reserved. The child television mode of the user is preset with a continuous watching time length threshold value and a rest time length. And when the user is a child, entering a child television mode, and reminding the user to temporarily turn off the television when the continuous watching time of the user reaches the continuous watching time threshold. Further, after the user is reminded to temporarily turn off the television, a television turning-off instruction of the user is detected, if the television turning-off instruction of the user is not detected within a preset time, the current television content is quitted from being played, and the television is automatically switched to a dormant state, so that the situation that children are addicted to the television and the physical and mental development are influenced is avoided.
In the embodiment of the application, if the user is a child, recommending the television content which the child is allowed to watch in a child television mode; and if the user is a non-child, recommending the television content in a common television mode.
As an embodiment of the present invention, fig. 5 shows a specific implementation flow of recommending, based on a television mode corresponding to the user type, television content corresponding to the user type in a television content recommendation method provided in an embodiment of the present invention, and details are as follows:
d1: and acquiring the interest tag selected by the user.
D2: and obtaining the historical behavior information of the user from a play cache library.
D3: and generating an interest feature vector of the user based on the interest tag selected by the user and the historical behavior information.
D4: and selecting the television content corresponding to the interest feature vector from a candidate television content set according to the interest feature vector and a preset feature resource comparison table to generate a television content recommendation table.
D5: and recommending the television content to the user based on the television content recommendation table.
In the embodiment of the present application, a behavior portrait is first taken for a user, a plurality of interest tags are preset, and a specified number of interest tags are selected by the user, where the interest tags are exemplarily set as follows: english, mathematics, animation, comic, children's songs, nature, sports, movies, entertainment, interviews, science and education, agriculture, opera, military affairs, talent show, children, legal, economic, documentary, human history, natural geography. And then obtaining the historical behavior information of the user from a play cache library, wherein the historical behavior information of the user is stored in the play cache library. Analyzing the historical behavior information of the user, and generating an interest feature vector of the user according to the interest tag selected by the user and the historical behavior information. The preset feature resource comparison table comprises a mapping relation between interest features and resource contents. Specifically, the mapping relationship between the interest feature ID and the resource content ID is included. And selecting the television content corresponding to the interest feature vector from a candidate television content set according to the interest feature vector and a preset feature resource comparison table to generate a television content recommendation table. Further, the television content recommendation table may be generated by different recommendation engine recommendations. And if a plurality of recommendation engines exist, presetting a characteristic resource comparison table of each engine by each recommendation engine according to the interest characteristic vector and the television content corresponding to each recommendation engine.
Optionally, the specific implementation flow of step D5 is as follows:
d51: and sequencing the television contents in the television content recommendation table according to a preset sequencing rule.
D52: and recommending the television content to the user according to the sequenced television content recommendation table.
In this embodiment of the present application, the television contents in the television content recommendation table are sorted, for example, the historical behavior information of the user includes a historical score, and the television contents in the television content recommendation table are sorted according to the historical score from high to low. The historical score may be the score of the user on the tv content, or may be the average score of all users who have viewed the tv content on the tv content. Or, according to the interest tag selected by the user, the television contents in the television content recommendation table are sorted, and the television content recommendation is performed on the user according to the sorted television content recommendation table, so that the recommendation efficiency and the satisfaction degree of the user are improved, and the user experience is enhanced.
Optionally, in this embodiment of the application, before sorting the television contents in the television content recommendation table according to a preset sorting rule, filtering the television contents in the television content recommendation table is further included. Specifically, the television contents in the television content recommendation table are filtered according to the interest tag selected by the user, the television content recommendation table is updated according to the filtering result, and the television contents in the updated television content recommendation table are sorted.
In the embodiment of the invention, the user information is acquired and input into the trained deep learning model to determine the user type of the user, and then the user enters the corresponding television mode according to the user type, so that the television content corresponding to the user type is recommended based on the television mode corresponding to the user type.
It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.
Fig. 6 shows a block diagram of a television content recommendation system provided in an embodiment of the present application, and only shows portions related to the embodiment of the present application for convenience of description.
Referring to fig. 6, the television content recommendation system includes: a user information obtaining unit 61, a user type determining unit 62, a mode switching and content recommending unit 63, wherein:
a user information acquisition unit 61 for acquiring user information;
a user type determining unit 62, configured to input the user information into the trained deep learning model, and determine a user type of the user;
and a mode switching and content recommending unit 63, configured to enter a corresponding television mode according to the user type, and recommend television content corresponding to the user type based on the television mode corresponding to the user type.
Optionally, the user information includes speech feature information and face feature information, the deep learning model includes a speech recognition network model and a face recognition network model, and the user type determining unit 62 includes:
a first probability determination module, configured to input the speech feature information into a trained speech recognition network model, and determine a first user type probability of the user;
the second probability determination module is used for inputting the face feature information into a trained face recognition network model and determining the second user type probability of the user;
and the user type determining module is used for determining the user type of the user according to the first user type probability and/or the second user type probability.
Optionally, the television content recommendation system further includes:
the first model building unit is used for building a voice recognition network model, and the voice recognition network model comprises a coding layer, a feedforward type sequence memory network layer and a full connection layer;
the voice information acquisition unit is used for acquiring sample voice information, and the sample voice information is provided with a user type label; the user type tag is a child or a non-child;
the information coding unit is used for coding the sample voice information in the coding layer to obtain a sample voice vector and transmitting the sample voice vector to the feedforward type sequence memory network layer;
a voice feature extraction unit, configured to perform feature extraction on the sample voice vector in the feedforward type sequence memory network layer to obtain a sample voice feature vector, and transmit the sample voice feature vector to the full connection layer;
a first classification unit, configured to perform user type probability calculation on the sample voice information according to the sample voice feature vector in the full connection layer, so as to obtain a first user type probability of the sample voice information;
and the first parameter optimization unit is used for optimizing and adjusting the model parameters of the voice recognition network model according to a preset parameter adjustment algorithm until the user type corresponding to the first user type probability of the sample voice information is consistent with the user type label of the sample voice information.
Optionally, the television content recommendation system further includes:
the second model building unit is used for building a face recognition network model, and the face recognition network model comprises a convolution layer, a pooling layer and a full-connection layer;
the face information acquisition unit is used for acquiring sample face information, and the sample face information is provided with a user type label; the type tag is a child or a non-child;
a face feature extraction unit, configured to perform face feature extraction according to the sample face information in the convolutional layer, to obtain a sample face feature of the sample face information;
the second classification unit is used for carrying out user type probability calculation on the sample face information in the full connection layer according to the sample face characteristics to obtain a second user type probability of the sample face information;
and the second parameter optimization unit is used for optimizing and adjusting the model parameters of the face recognition network model according to a preset parameter adjustment algorithm until the user type corresponding to the second user type probability of the sample face information is consistent with the type label of the sample face information.
Optionally, the mode switching and content recommending unit 63 includes:
the interest tag acquisition module is used for acquiring the interest tag selected by the user;
the historical information acquisition module is used for acquiring historical behavior information of the user;
the feature vector generation module is used for generating an interest feature vector of the user based on the interest tag selected by the user and the historical behavior information;
the content recommendation table generation module is used for selecting the television content corresponding to the interest feature vector from the candidate television content set according to the interest feature vector and a preset feature resource comparison table to generate a television content recommendation table;
and the television content recommending module is used for recommending the television content to the user based on the television content recommending table.
Optionally, the television content recommendation module includes:
the content sequencing submodule is used for sequencing the television contents in the television content recommendation table according to a preset sequencing rule;
and the content recommendation submodule is used for recommending the television content to the user according to the sequenced television content recommendation table.
Optionally, the television modes include a child television mode and a general television mode, the user types include children and non-children, and the mode switching and content recommending unit 63 includes:
the first mode switching module is used for entering a child television mode if the user is a child;
and the second mode switching module is used for entering a common television mode if the user is a non-child.
In the embodiment of the invention, the user information is acquired and input into the trained deep learning model to determine the user type of the user, then the corresponding television mode is entered according to the user type, and the television content corresponding to the user type is recommended based on the television mode corresponding to the user type.
Fig. 7 is a schematic diagram of a smart television according to an embodiment of the present invention. As shown in fig. 7, the smart tv 7 of this embodiment includes: a processor 70, a memory 71 and a computer program 72, such as a television content recommendation program, stored in said memory 71 and operable on said processor 70. The processor 70, when executing the computer program 72, implements the steps in the various television content recommendation method embodiments described above, such as the steps 101 to 103 shown in fig. 1. Alternatively, the processor 70, when executing the computer program 72, implements the functions of each module/unit in each system embodiment described above, for example, the functions of the units 61 to 63 shown in fig. 6.
Illustratively, the computer program 72 may be partitioned into one or more modules/units that are stored in the memory 71 and executed by the processor 70 to implement the present invention. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution process of the computer program 72 in the smart tv 7. For example, the computer program 72 may be divided into a user information obtaining unit, a user type determining unit, a television mode switching unit, and a television content recommending unit, and the specific functions of each unit are as follows:
a user information acquisition unit for acquiring user information;
the user type determining unit is used for inputting the user information into the trained deep learning model and determining the user type of the user;
the television mode switching unit is used for entering a corresponding television mode according to the user type;
and the television content recommending unit is used for recommending the television content corresponding to the user type based on the television mode corresponding to the user type.
The smart television 7 may include, but is not limited to, a processor 70 and a memory 71. Those skilled in the art will appreciate that fig. 7 is only an example of the smart tv 7, does not constitute a limitation to the smart tv 7, and may include more or less components than those shown, or combine some components, or different components, for example, the smart tv may further include an input-output device, a network access device, a bus, etc.
It should be understood that in the embodiments of the present Application, the Processor 702 may be a Central Processing Unit (CPU), and the Processor may be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 701 may include both read-only memory and random access memory, and provides instructions and data to the processor 702. Some or all of memory 701 may also include non-volatile random access memory. For example, memory 701 may also store information of device types.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of each functional unit or module is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units or modules as required, that is, the internal structure of the apparatus (or system) is divided into different functional units or modules so as to perform all or part of the above-mentioned functions. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the system may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described system embodiments are merely illustrative, and for example, the division of the modules or units is only one logical division, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments may be implemented. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable medium may contain other components which may be suitably increased or decreased as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media which may not include electrical carrier signals and telecommunications signals in accordance with legislation and patent practice.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.

Claims (9)

1. A method for recommending television content, comprising:
acquiring user information, wherein the user information comprises voice characteristic information and face characteristic information;
inputting the user information into a trained deep learning model, and determining the user type of the user; the deep learning model comprises a voice recognition network model and a face recognition network model, the user information is input into the trained deep learning model, and the user type of the user is determined, wherein the method comprises the following steps: inputting the voice characteristic information into a trained voice recognition network model, and determining a first user type probability of the user; inputting the face feature information into a trained face recognition network model, and determining a second user type probability of the user; determining the user type of the user according to the first user type probability and the second user type probability, specifically, respectively presetting the weight of the first user type probability and the second user type probability, fusing the first user type probability and the second user type probability according to the first user type probability and the weight thereof, obtaining a fused classification probability, and determining the user type of the user according to the fused classification probability;
and entering a corresponding television mode according to the user type, and recommending the television content corresponding to the user type based on the television mode corresponding to the user type.
2. The television content recommendation method according to claim 1, further comprising training the speech recognition network model, specifically comprising:
constructing a voice recognition network model, wherein the voice recognition network model comprises a coding layer, a feedforward type sequence memory network layer and a full connection layer;
obtaining sample voice information, wherein the sample voice information has a user type label;
in the coding layer, coding the sample voice information to obtain a sample voice vector, and transmitting the sample voice vector to the feedforward type sequence memory network layer;
in the feedforward type sequence memory network layer, extracting the characteristics of the sample voice vector to obtain a sample voice characteristic vector, and transmitting the sample voice characteristic vector to the full connection layer;
in the full connection layer, according to the sample voice feature vector, carrying out user type probability calculation on the sample voice information to obtain a first user type probability of the sample voice information;
and adjusting the model parameters of the voice recognition network model according to a preset parameter adjustment algorithm until the user type corresponding to the first user type probability of the sample voice information is consistent with the user type label of the sample voice information.
3. The television content recommendation method according to claim 1, further comprising training the face recognition network model, specifically comprising:
constructing a face recognition network model, wherein the face recognition network model comprises a convolution layer, a pooling layer and a full-connection layer;
obtaining sample face information, wherein the sample face information has a user type label;
in the convolutional layer, extracting face features according to the sample face information to obtain sample face features of the sample face information;
in the full connection layer, according to the sample face features, carrying out user type probability calculation on the sample face information to obtain a second user type probability of the sample face information;
and optimizing and adjusting the model parameters of the face recognition network model according to a preset parameter adjustment algorithm until the user type corresponding to the second user type probability of the sample face information is consistent with the type label of the sample face information.
4. The method according to claim 1, wherein recommending the tv content corresponding to the user type based on the tv mode corresponding to the user type comprises:
obtaining the interest tag selected by the user;
acquiring historical behavior information of the user;
generating an interest feature vector of the user based on the interest tag selected by the user and the historical behavior information;
selecting television contents corresponding to the interest feature vector from a candidate television content set according to the interest feature vector and a preset feature resource comparison table, and generating a television content recommendation table;
and recommending the television content to the user based on the television content recommendation table.
5. The method of claim 4, wherein the recommending television content to the user based on the television content recommendation table comprises:
sequencing the television contents in the television content recommendation table according to a preset sequencing rule;
and recommending the television content to the user according to the sequenced television content recommendation table.
6. The method according to any one of claims 1 to 5, wherein the TV modes include a child TV mode and a normal TV mode, the user types include children and non-children, and the entering into the corresponding TV mode according to the user types includes:
if the user is a child, entering a child television mode;
and if the user is a non-child, entering a common television mode.
7. A television content recommendation system, characterized in that the television content recommendation system comprises:
the system comprises a user information acquisition unit, a processing unit and a processing unit, wherein the user information acquisition unit is used for acquiring user information which comprises voice characteristic information and face characteristic information;
the user type determining unit is used for inputting the user information into the trained deep learning model and determining the user type of the user; the deep learning model comprises a voice recognition network model and a face recognition network model, and the user type determining unit comprises: a first probability determination module, configured to input the speech feature information into a trained speech recognition network model, and determine a first user type probability of the user; the second probability determination module is used for inputting the face feature information into a trained face recognition network model and determining the second user type probability of the user; the user type determining module is used for determining the user type of the user according to the first user type probability and the second user type probability; the user type determination module is specifically configured to: respectively presetting the weight of the first user type probability and the weight of the second user type probability, fusing the first user type probability and the second user type probability according to the first user type probability and the weight thereof and the second user type probability and the weight thereof to obtain a fused classification probability, and determining the user type of the user according to the fused classification probability;
and the mode switching and content recommending unit is used for entering a corresponding television mode according to the user type so as to recommend the television content corresponding to the user type based on the television mode corresponding to the user type.
8. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the television content recommendation method according to any one of claims 1 to 6.
9. An intelligent television comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the television content recommendation method according to any one of claims 1 to 6 when executing the computer program.
CN201910698763.0A 2019-07-31 2019-07-31 Television content recommendation method, system, storage medium and smart television Active CN112333545B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910698763.0A CN112333545B (en) 2019-07-31 2019-07-31 Television content recommendation method, system, storage medium and smart television

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910698763.0A CN112333545B (en) 2019-07-31 2019-07-31 Television content recommendation method, system, storage medium and smart television

Publications (2)

Publication Number Publication Date
CN112333545A CN112333545A (en) 2021-02-05
CN112333545B true CN112333545B (en) 2022-03-22

Family

ID=74319460

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910698763.0A Active CN112333545B (en) 2019-07-31 2019-07-31 Television content recommendation method, system, storage medium and smart television

Country Status (1)

Country Link
CN (1) CN112333545B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113378045B (en) * 2021-06-08 2024-02-09 深圳Tcl新技术有限公司 Digital content distribution method, device and storage medium
CN117177014B (en) * 2023-11-03 2024-01-30 帕科视讯科技(杭州)股份有限公司 IPTV-based children anti-addiction early warning method, storage medium and server

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI540894B (en) * 2014-06-13 2016-07-01 國立臺灣大學 Display system and program selection method thereof
CN105426723A (en) * 2015-11-20 2016-03-23 北京得意音通技术有限责任公司 Voiceprint identification, face identification and synchronous in-vivo detection-based identity authentication method and system
CN105681920B (en) * 2015-12-30 2017-03-15 深圳市鹰硕音频科技有限公司 A kind of Network teaching method and system with speech identifying function
CN106529402B (en) * 2016-09-27 2019-05-28 中国科学院自动化研究所 The face character analysis method of convolutional neural networks based on multi-task learning
US10743063B2 (en) * 2017-11-20 2020-08-11 Qualcomm Incorporated Intelligent routing of notifications related to media programming
CN108447490B (en) * 2018-02-12 2020-08-18 阿里巴巴集团控股有限公司 Voiceprint recognition method and device based on memorability bottleneck characteristics
CN108509619B (en) * 2018-04-04 2021-05-04 科大讯飞股份有限公司 Voice interaction method and device
CN109121007B (en) * 2018-09-18 2020-11-03 深圳市酷开网络科技有限公司 Movie content recommendation method based on multi-face recognition, smart television and system
CN109547854A (en) * 2019-01-23 2019-03-29 深圳市酷开网络科技有限公司 A kind of TV method for pushing, smart television and storage medium based on Application on Voiceprint Recognition

Also Published As

Publication number Publication date
CN112333545A (en) 2021-02-05

Similar Documents

Publication Publication Date Title
Ma et al. Joint embedding VQA model based on dynamic word vector
CN111400591B (en) Information recommendation method and device, electronic equipment and storage medium
Zhou et al. A lightweight convolutional neural network for real-time facial expression detection
WO2021042826A1 (en) Video playback completeness prediction method and apparatus
CN112270196B (en) Entity relationship identification method and device and electronic equipment
CN108920648B (en) Cross-modal matching method based on music-image semantic relation
CN113033450B (en) Multi-mode continuous emotion recognition method, service inference method and system
CN111368075A (en) Article quality prediction method and device, electronic equipment and storage medium
CN110035302B (en) Information recommendation method and device, model training method and device, computing equipment and storage medium
CN111652357A (en) Method and system for solving video question-answer problem by using specific target network based on graph
CN112138403B (en) Interactive behavior recognition method and device, storage medium and electronic equipment
CN112333545B (en) Television content recommendation method, system, storage medium and smart television
US20230353828A1 (en) Model-based data processing method and apparatus
CN113283238B (en) Text data processing method and device, electronic equipment and storage medium
CN112699310A (en) Cold start cross-domain hybrid recommendation method and system based on deep neural network
CN111144344B (en) Method, device, equipment and storage medium for determining person age
CN116958324A (en) Training method, device, equipment and storage medium of image generation model
CN115795156A (en) Material recall and neural network training method, device, equipment and storage medium
CN113254788B (en) Big data based recommendation method and system and readable storage medium
CN115438246A (en) Content evaluation method and device, storage medium and electronic equipment
CN114626480A (en) Multi-source heterogeneous data feature extraction device and method, storage medium and electronic equipment
CN116361508A (en) Video recommendation method and related equipment
CN111310460B (en) Statement adjusting method and device
CN113610080A (en) Cross-modal perception-based sensitive image identification method, device, equipment and medium
CN117556149B (en) Resource pushing method, device, electronic equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
CB02 Change of applicant information

Address after: 516006 TCL science and technology building, No. 17, Huifeng Third Road, Zhongkai high tech Zone, Huizhou City, Guangdong Province

Applicant after: TCL Technology Group Co.,Ltd.

Address before: 516006 Guangdong province Huizhou Zhongkai hi tech Development Zone No. nineteen District

Applicant before: TCL Corp.

CB02 Change of applicant information
GR01 Patent grant
GR01 Patent grant