Electronic program guide generation method and device, equipment and medium

Abstract

The invention relates to an electronic program guide generation method and device, equipment and a medium, and belongs to the technical field of electronic program guides, and the method comprises the steps: obtaining historical watching program records, real-time playing program data and a program database; analyzing the historical program watching records to obtain watching preference characteristics; analyzing the real-time playing program data to obtain real-time program characteristics; analyzing the program database to obtain program characteristics of each program in the program database; calculating a first similarity between the watching preference feature and the real-time program feature, and generating a reference score according to the first similarity; calculating a second similarity between the watching preference feature and the program feature of each program in the program database, and generating a comparison score according to the second similarity; screening the programs of which the comparison scores are greater than the reference score in a program database, and setting the programs as to-be-played programs; and sorting the comparison scores of the to-be-played programs, and generating an electronic program guide according to a sorting result.

Description

Technical Field

[0001] The present invention belongs to the technical field of electronic program guides, and in particular relates to an electronic program guide generation method, device, equipment and medium. Background Art

[0002] With the explosive growth of digital media content and the diversification of user consumption habits, the generation of electronic program guides and user experience have become the focus of television and streaming service providers. Compared with traditional passive program playback, electronic program guides can flexibly adjust the programs presented to users.

[0003] Existing methods for generating electronic program guides usually identify users' interests based on their historical viewing records, then filter out programs with high similarity from a database based on their interests, and then make an electronic program guide; however, this method ignores the possibility that users' interests may change in the short term, and has the technical problem of poor dynamic performance. Summary of the invention

[0004] In view of the above analysis, the embodiments of the present invention aim to provide an electronic program guide generation method, apparatus, device and medium to solve one or more of the above problems existing in the prior art.

[0005] The object of the present invention is achieved in that:

[0006] A first aspect of the present invention provides an electronic program guide generation method, comprising:

[0007] Obtain historical viewing records, real-time broadcast program data and program database;

[0008] Analyze the historical program viewing records to obtain viewing preference characteristics; analyze the real-time broadcast program data to obtain real-time program characteristics; analyze the program database to obtain program characteristics of each program in the program database;

[0009] Calculating a first similarity between the viewing preference feature and the real-time program feature, and generating a benchmark score according to the first similarity; calculating a second similarity between the viewing preference feature and the program features of each program in the program database, and generating a comparison score according to the second similarity;

[0010] Filtering the programs in the program database whose comparison scores are greater than the benchmark scores and setting them as programs to be played;

[0011] The comparison scores of the programs to be played are sorted, and an electronic program guide is generated according to the sorting result.

[0012] Furthermore, the history program viewing record includes the type of the history program viewing, viewing duration, number of views, and viewing time period.

[0013] Furthermore, analyzing the historical program viewing records to obtain viewing preference features includes: using a Word2vec model to convert the historical program viewing records into feature vectors; calculating the weights of program type, viewing duration, number of views, and viewing time period, and adjusting the feature vectors through a weighted strategy to obtain the viewing preference features.

[0014] Furthermore, the analyzing the real-time program data to obtain real-time program features includes: capturing image frames from the video stream of the real-time program and extracting image features; performing spectrum analysis on the audio stream of the real-time program to extract audio features; extracting text features in program subtitles or dialogues; and performing multimodal fusion of the image features, audio features, and text features to obtain a program feature vector.

[0015] Furthermore, the program database is analyzed to obtain program features of each program in the program database, including: extracting metadata of each program in the program database, including information such as program type, duration, actors, directors, tags and introduction; and using a keyword-based text analysis model to extract and classify features of the program metadata to generate program features.

[0016] Further, the calculating a first similarity between the viewing preference feature and the real-time program feature and generating a benchmark score according to the first similarity includes: calculating a cosine similarity between the viewing preference feature and the real-time program feature as the first similarity, expressed as: v u represents the viewing preference feature, v r Indicates the real-time program characteristics; the benchmark score is expressed as: score s =α×sim1+β, α, β∈[0,1] represent parameters used to adjust the benchmark score.

[0017] Further, the calculating a second similarity between the viewing preference feature and the program feature of each program in the program database, and generating a comparison score according to the second similarity, includes: calculating a cosine similarity between the viewing preference feature and the program feature of each program in the program database as the second similarity, expressed as: v pi Represents the program characteristics of the i-th program in the program database; the comparison score is expressed as: score c =γ×sim2+δ, γ,δ∈[0,1] represent the parameters used to adjust the contrast score.

[0018] A second aspect of the present invention provides an electronic program guide generating device, comprising:

[0019] The acquisition module is used to obtain historical viewing program records, real-time broadcast program data and program database;

[0020] An analysis module, used to analyze the historical program viewing records to obtain viewing preference characteristics; analyze the real-time broadcast program data to obtain real-time program characteristics; analyze the program database to obtain program characteristics of each program in the program database;

[0021] a calculation module, configured to calculate a first similarity between the viewing preference feature and the real-time program feature, and generate a benchmark score according to the first similarity; calculate a second similarity between the viewing preference feature and the program features of each program in the program database, and generate a comparison score according to the second similarity;

[0022] A screening module, used for screening programs in a program database whose comparison scores are greater than the reference scores, and setting them as programs to be played;

[0023] The generating module is used to sort the comparison scores of the programs to be played and generate an electronic program guide according to the sorting result.

[0024] A third aspect of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the electronic program guide generation method described in any embodiment is implemented.

[0025] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for generating an electronic program guide described in any one of the embodiments is implemented.

[0026] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0027] The electronic program guide generation method provided by the present invention obtains viewing preference features reflecting the user's long-term interests and real-time program features reflecting the user's short-term interests by analyzing the user's historical program viewing records, real-time program broadcast data and program database, calculates a benchmark score based on the viewing preference features and the real-time program features, sets standards for the electronic program guide, calculates a comparison score based on the viewing preference features and the program features, compares the comparison score with the benchmark score, and mines programs that meet the user's long-term interests and short-term interests from the program database, thereby generating an electronic program guide, which can ensure that the guide meets the user's long-term interests with high accuracy, and can also ensure that it meets the user's short-term interests with good dynamic performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of this specification. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0029] Figure 1 A flowchart of an electronic program guide generation method provided in Embodiment 1 of the present invention;

[0030] Figure 2 A schematic diagram of an electronic program guide generating device provided in Embodiment 2 of the present invention;

[0031] Figure 3 This is a schematic diagram of the electronic device architecture provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. It should be noted that, in the absence of conflict, the embodiments and features in the embodiments disclosed in this disclosure can be combined, separated, interchanged and / or rearranged with each other. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0033] Example 1

[0034] A specific embodiment of the present invention, as Figure 1 As shown, a method for generating an electronic program guide is disclosed, comprising the following steps:

[0035] S1. Obtain historical viewing records, real-time broadcast program data and program database.

[0036] Exemplarily, the historical program viewing record includes the historical program viewing type, viewing duration, viewing times, and viewing time period.

[0037] In some embodiments, the historical program viewing records also include user interaction data, such as search queries, program ratings, favorite records, etc.

[0038] The real-time broadcast program data includes video stream, audio stream, text data, metadata, etc.

[0039] The program database includes programs to be broadcast and their metadata.

[0040] In some embodiments, the historical program viewing records need to be pre-processed, such as data cleaning, formatting, time slicing, etc.

[0041] S2. Analyze the historical program viewing records to obtain viewing preference characteristics; analyze the real-time broadcast program data to obtain real-time program characteristics; analyze the program database to obtain program characteristics of each program in the program database.

[0042] In this embodiment, step S2 specifically includes:

[0043] S201, using a Word2vec model to convert the historical program viewing record into a feature vector;

[0044] For example, a pre-trained Skip-gram model is used to first extract keywords from the historical program viewing records, and then the keyword list is mapped to the vector space through serialization technology to obtain feature vectors, including the program type vector vector type (p i ), viewing time vector vector duration (p i ), number of views vector vector frequency (p i ), viewing time vector vector time (p i ).

[0045] S202, calculating the weights of program type, viewing duration, number of views, and viewing period, and adjusting the feature vector by a weighted strategy to obtain the viewing preference feature;

[0046] Specifically, the weight of program type is expressed as: Indicates program p i The weight of the type; the weight of the viewing time is expressed as: duration(p i ) indicates that the user watches program p i The length of the video, N represents the total number of programs in the historical viewing record; the weight of the number of views is expressed as: count(p i ) indicates that the user watches program p i The weight of the viewing period is expressed as: Indicates watching program p i time from now, λ represents the attenuation coefficient;

[0047] The viewing preference feature is expressed as follows:

[0048]

[0049] In some embodiments, a real-time update mechanism is also introduced, and the steps are as follows:

[0050] (I) regularly adding the program data newly watched by the user to the historical program viewing record;

[0051] (II) Gradually reduce the weight of old viewing records through a decay mechanism.

[0052] S203, collecting image frames from the video stream of the real-time broadcast program, and extracting image features;

[0053] Specifically, real-time program content feature extraction is a very important part, including image features, audio features and text features. Its purpose is to identify and extract features by analyzing the content of the currently playing program when the user is watching the program, so as to match it with the user's viewing preferences; the steps for extracting image features are as follows:

[0054] (I) Image frame sampling: Extract static image frames from the video stream at a certain frequency, sampling one frame every second to obtain the visual features of the video, and then identify and extract the key frames in the video. Determine which frames are key frames by detecting scene changes or calculating the difference between frames;

[0055] (II) Image feature representation: Use the pre-trained CNN model to represent the extracted image frames and convert the image into a feature vector; use the object detection algorithm to identify the specific objects or scenes in the image, thereby extracting more fine-grained features and identifying the characters, objects, backgrounds, etc. in the program; generate image descriptions through visual semantic analysis to convert visual information into readable text for further analysis;

[0056] (III) Video content aggregation: Aggregate the features of continuous image frames into video clip features to capture the dynamic changes of the video; identify the types of different scenes in the video and annotate the categories of the scenes, such as indoor, outdoor, action scenes, etc.

[0057] S204, performing spectrum analysis on the audio stream of the real-time broadcast program to extract audio features;

[0058] The steps for extracting audio features are as follows:

[0059] (I) Audio signal processing: Analyze the frequency characteristics of audio signals through short-time Fourier transform, extract MFCC (Mel-frequency cepstral coefficients), Chroma features, pitch, rhythm and other features to describe the acoustic properties of audio;

[0060] (II) Audio event detection: Use speech recognition models such as DeepSpeech to convert speech in the audio into text and extract the dialogue and speech content in the program; detect and identify specific sound effects in the audio, such as laughter, applause, and music, and identify the frequency and location of the sound effects in the program; identify the emotional type of the program, such as joy, sadness, tension, etc., by analyzing the tone and emotional clues in the audio.

[0061] S205, extracting text features in program subtitles or dialogues;

[0062] The steps for extracting text features are as follows:

[0063] (I) Subtitle and dialogue processing: Extract subtitle information from the video. You can use OCR technology to recognize text from image frames, or you can obtain it from the program's own subtitle files such as SRT. Perform word segmentation, part-of-speech tagging, entity recognition and other processing on the extracted text to obtain the text's structured information;

[0064] (II) Topic and keyword extraction: Use topic modeling algorithms such as LDA to identify themes in the text and extract the core topics discussed in the program; use algorithms such as TextRank to extract keywords from the text and identify important concepts in the program;

[0065] (III) Text Sentiment Analysis: Use sentiment analysis models to classify the sentiment of text and identify the emotional attitudes expressed in the text, such as positive, negative, and neutral; calculate the sentiment intensity in the text and evaluate the strength of the sentiment, providing more sophisticated analysis for sentiment recognition.

[0066] S206, performing multimodal fusion on the image features, audio features and text features to obtain a program feature vector;

[0067] Specifically, the multimodal fusion of image features, audio features, and text features includes the following steps:

[0068] (I) Feature Alignment

[0069] Feature alignment: Synchronize features from different modalities to ensure that they correspond to the same time period on the timeline, which usually involves processing timestamps between audio, video frames, and subtitles;

[0070] Spatial alignment: Mapping features of different modalities to the same or similar dimensions in feature space to facilitate fusion;

[0071] (II) Feature Fusion Strategy

[0072] Feature concatenation: simply concatenate multimodal features together to form a high-dimensional feature vector;

[0073] Weighted feature fusion: Assign weights to features of different modalities and perform weighted summation or weighted average according to their importance. The weights can be determined through experiments or learning mechanisms.

[0074] Deep fusion: Use deep neural networks for feature fusion and automatically learn complex relationships between modalities. This can be achieved through multi-layer perceptrons, self-attention mechanisms, etc.

[0075] (III) Representation and storage of fusion features

[0076] Feature normalization: Normalize the fused feature vectors to ensure that the numerical ranges of different feature dimensions are consistent.

[0077] Feature storage and indexing: The fused program feature vectors are stored in the database and indexed for quick retrieval and matching of user preferences.

[0078] S207, extracting metadata of each program in the program database, including program type, duration, actors, directors, tags, introduction and other information;

[0079] S208: Use a keyword-based text analysis model to extract and classify features of the program metadata to generate program features.

[0080] Specifically, keyword-based text analysis models, such as TF-IDF, TextRank, etc., are used to identify and extract the keywords that best represent the program from the metadata, classify the extracted keywords and metadata, and classify the program into appropriate types or themes; by combining the extracted keywords and classification information, the model will generate program features, including structured metadata and features obtained from text analysis, to ensure that the generated program features can accurately reflect the main information of the program and the user's potential interests.

[0081] S3. Calculate a first similarity between the viewing preference feature and the real-time program feature, and generate a benchmark score according to the first similarity; calculate a second similarity between the viewing preference feature and the program features of each program in the program database, and generate a comparison score according to the second similarity.

[0082] In this embodiment, step S3 specifically includes:

[0083] S301, calculating the cosine similarity between the viewing preference feature and the real-time program feature as a first similarity, expressed as: v u represents the viewing preference feature, v r Indicates real-time program characteristics;

[0084] S302, the benchmark score is expressed as: score s=α×sim1+β, α, β∈[0,1] represent the parameters used to adjust the benchmark score;

[0085] S303: Calculate the cosine similarity between the viewing preference feature and the program features of each program in the program database as the second similarity, expressed as: v pi represents the program features of the i-th program in the program database;

[0086] S304, the comparison score is expressed as: score c =γ×sim2+δ, γ,δ∈[0,1] represent the parameters used to adjust the contrast score.

[0087] In some embodiments, a neural network may be used to optimize α, β and γ, δ so that the benchmark score and the comparison score are more accurate.

[0088] S4, filtering the programs in the program database whose comparison scores are greater than the benchmark scores and setting them as programs to be played;

[0089] Specifically, step S4 includes the following steps:

[0090] S401, traverse all programs in the program database, and calculate the comparison score of each program one by one according to steps S303 and S304;

[0091] S402. For each program, compare the comparison score with the benchmark score obtained in step S302; if the comparison score is greater than the benchmark score, it means that the program is consistent with the user's viewing preference and is set as the program to be played; if the comparison score is less than the benchmark score, it means that the program does not meet the user's viewing preference and needs to be excluded when generating the electronic program guide.

[0092] S5. Sort the comparison scores of the programs to be played, and generate an electronic program guide according to the sorting result.

[0093] Specifically, after all the programs to be played are obtained in step S4, in order to further consider the viewing preferences of users, the comparison scores of all the programs to be played are sorted by a sorting algorithm, and all the programs to be played are made into an electronic program guide according to the comparison scores from high to low.

[0094] Compared with the prior art, the electronic program guide generation method provided in this embodiment obtains viewing preference features reflecting the user's long-term interests and real-time program features reflecting the user's short-term interests by analyzing the user's historical program viewing records, real-time program broadcast data and program database, calculates a benchmark score based on the viewing preference features and the real-time program features, sets standards for the electronic program guide, calculates a comparison score based on the viewing preference features and the program features, and compares the comparison score with the benchmark score to mine programs that meet the user's long-term interests and short-term interests from the program database, thereby generating an electronic program guide, which can ensure that the guide meets the user's long-term interests with high accuracy, and at the same time can ensure that it meets the user's short-term interests with good dynamic performance.

[0095] Example 2

[0096] This embodiment provides an electronic program guide generating device, such as Figure 2 As shown, including:

[0097] The acquisition module is used to obtain historical viewing program records, real-time broadcast program data and program database;

[0098] An analysis module, used to analyze the historical program viewing records to obtain viewing preference characteristics; analyze the real-time broadcast program data to obtain real-time program characteristics; analyze the program database to obtain program characteristics of each program in the program database;

[0099] a calculation module, configured to calculate a first similarity between the viewing preference feature and the real-time program feature, and generate a benchmark score according to the first similarity; calculate a second similarity between the viewing preference feature and the program features of each program in the program database, and generate a comparison score according to the second similarity;

[0100] A screening module, used for screening programs in a program database whose comparison scores are greater than the reference scores, and setting them as programs to be played;

[0101] The generating module is used to sort the comparison scores of the programs to be played and generate an electronic program guide according to the sorting result.

[0102] Example 3

[0103] This embodiment provides an electronic device, such as Figure 3 As shown, it includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the electronic program guide generation method as described in any of the above embodiments is implemented.

[0104] Example 4

[0105] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for generating an electronic program guide as described in any of the above embodiments is implemented.

[0106] Computer readable storage media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0107] The professionals should further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0108] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0109] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for generating an electronic program guide, characterized in that: include: Obtain historical viewing records, real-time broadcast program data and program database; Analyze the historical program viewing records to obtain viewing preference characteristics; Analyzing the real-time broadcast program data to obtain real-time program features; Analyzing the program database to obtain program features of each program in the program database; calculating a first similarity between the viewing preference feature and the real-time program feature, and generating a benchmark score according to the first similarity; calculating a second similarity between the viewing preference feature and a program feature of each program in the program database, and generating a comparison score according to the second similarity; Filtering the programs in the program database whose comparison scores are greater than the benchmark scores and setting them as programs to be played; The comparison scores of the programs to be played are sorted, and an electronic program guide is generated according to the sorting result.

2. The method for generating an electronic program guide according to claim 1, characterized in that: The historical program viewing records include the historical program viewing type, viewing duration, viewing times, and viewing time period.

3. The method for generating an electronic program guide according to claim 2, characterized in that: The analyzing the historical program viewing records to obtain viewing preference characteristics includes: Using the Word2vec model to convert the historical program viewing records into feature vectors; The weights of program type, viewing duration, number of views, and viewing period are calculated, and the feature vector is adjusted through a weighted strategy to obtain the viewing preference feature.

4. The method for generating an electronic program guide according to claim 1, characterized in that: The analyzing the real-time broadcast program data to obtain real-time program features includes: Capture image frames from the video stream of the real-time broadcast program and extract image features; Perform spectrum analysis on the audio stream of real-time broadcast programs to extract audio features; Extract text features from program subtitles or dialogues; The image features, audio features and text features are multimodally fused to obtain a program feature vector.

5. The method for generating an electronic program guide according to claim 1, characterized in that: The analyzing the program database to obtain program features of each program in the program database includes: Extract metadata of each program in the program database, including program type, duration, actors, directors, tags, introduction and other information; A keyword-based text analysis model is used to extract and classify the program metadata to generate program features.

6. The method for generating an electronic program guide according to claim 1, characterized in that: The calculating a first similarity between the viewing preference feature and the real-time program feature, and generating a benchmark score according to the first similarity, comprises: The cosine similarity between the viewing preference feature and the real-time program feature is calculated as the first similarity, which is expressed as: v u represents the viewing preference feature, v r Indicates real-time program characteristics; The benchmark score is expressed as: score s =α×sim1+β, α, β∈[0,1] represent parameters used to adjust the benchmark score.

7. The method for generating an electronic program guide according to claim 1, characterized in that: The calculating a second similarity between the viewing preference feature and the program features of each program in the program database, and generating a comparison score according to the second similarity, comprises: The cosine similarity between the viewing preference feature and the program features of each program in the program database is calculated as the second similarity, which is expressed as: v pi represents the program features of the i-th program in the program database; The comparison score is expressed as: score c =γ×sim2+δ, γ,δ∈[0,1] represent the parameters used to adjust the contrast score.

8. An electronic program guide generating device, characterized in that: The device comprises: The acquisition module is used to obtain historical viewing program records, real-time broadcast program data and program database; An analysis module, used to analyze the historical program viewing records to obtain viewing preference characteristics; analyze the real-time broadcast program data to obtain real-time program characteristics; analyze the program database to obtain program characteristics of each program in the program database; a calculation module, configured to calculate a first similarity between the viewing preference feature and the real-time program feature, and generate a benchmark score according to the first similarity; calculate a second similarity between the viewing preference feature and the program features of each program in the program database, and generate a comparison score according to the second similarity; A screening module, used for screening programs in a program database whose comparison scores are greater than the reference scores, and setting them as programs to be played; The generating module is used to sort the comparison scores of the programs to be played and generate an electronic program guide according to the sorting result.

9. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the method for generating an electronic program guide as claimed in any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the method for generating an electronic program guide as claimed in any one of claims 1 to 7 is implemented.

Images