Cloud big data real-time intelligent flow casting system and method based on AI analysis

Abstract

The invention provides a cloud big data real-time intelligent streaming system based on AI analysis. The cloud big data real-time intelligent streaming system comprises a video preprocessing module, a user behavior analysis module, a video generation module and a delivery and feedback module. The video preprocessing module is used for acquiring original video materials, performing automatic segmentation and classification according to the original video materials, and establishing a video clip material library; the user behavior analysis module is used for collecting and analyzing video watching behavior data of a user in real time, constructing a user interest model, and updating user attention information in real time according to the user interest model; the video generation module is used for acquiring user attention information, selecting video clips from a video clip material library, and combining the selected video materials according to a combination strategy; and the delivery and feedback module is used for delivering the edited video to a target platform, collecting feedback information in real time, introducing an AI feedback model, and generating a delivery strategy based on data feedback.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic projection systems, and more specifically to a cloud-based big data real-time intelligent traffic projection system and method based on AI analysis. Background Art

[0002] With the development of the Internet, inserting advertising information into video content is a common marketing method, which attracts the audience's attention through pictures and audio and effectively conveys product information. Traditional video advertising methods mainly use manual screening of video materials, analysis of user preferences, and manual screening and editing, which have problems such as low efficiency, high cost and slow response; traditional video streaming methods mainly rely on manual judgment and experience accumulation, and cannot achieve accurate and personalized video delivery. Summary of the invention

[0003] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a cloud-based big data real-time intelligent streaming system and method based on AI analysis to overcome the above-mentioned defects in the prior art.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A cloud-based big data real-time intelligent streaming system based on AI analysis, including a video preprocessing module, a user behavior analysis module, a video generation module, and a streaming and feedback module;

[0006] The video preprocessing module is used to obtain original video materials, automatically segment and classify the original video materials, and establish a video clip material library;

[0007] The user behavior analysis module is used to collect and analyze the user's video viewing behavior data in real time, and to construct a user interest model, and to update the user's attention information in real time according to the user interest model;

[0008] The video generation module is used to obtain user attention information, select video clips from the video clip material library, and combine the selected video materials according to the combination strategy;

[0009] The delivery and feedback module is used to deliver the edited video to the target platform, collect feedback information in real time, introduce an AI feedback model, and generate a delivery strategy based on data feedback.

[0010] Preferably, the combination strategy includes:

[0011] Combination preset steps: Based on historical video data analysis, a series of effective material combination templates are preset;

[0012] Steps for adjusting elements: using the material combination template, obtaining corresponding video materials, and creating videos based on the video materials, and obtaining corresponding video element requirements based on user attention information, and adjusting the materials in the material combination template;

[0013] Slicing and reassembly steps: Slice the selected video material, extract the parts that are more closely related to the video and reassemble them.

[0014] Preferably, the AI ​​feedback model is used to analyze and evaluate based on the feedback information and provide corresponding feedback results. The combination strategy also includes a video template adjustment step, which is used to obtain a material combination template and adjust the preset combination template based on the visual feedback information. The adjustment includes the content, sequence and duration of the material.

[0015] Preferably, a classification unit is provided in the video preprocessing module, and the classification unit is provided with different classification rules, and the classification rules include technical features, application scenarios, etc. According to different classification rules, the content of the video clips is analyzed and the video clips are classified, and the video material components are labeled according to the classification results.

[0016] Preferably, the technical features include editable video length, divisible material, material relevance, etc.

[0017] Preferably, the duration can be edited including classification principles including classification according to video content,

[0018] Preferably, the delivery strategy includes: obtaining audience type according to the user interest model, performing targeted delivery according to the target audience's attention information, and continuously optimizing the delivery strategy according to data feedback results.

[0019] Preferably, the feedback information includes the number of views, the number of likes, the completion rate, the number of fans gained, etc.

[0020] Preferably, the attention information includes video duration, user's personal preferences and interests, consumption habits, sensitive words, etc.

[0021] A cloud-based big data real-time intelligent streaming method based on AI analysis.

[0022] S1: Obtain video material and clipping information of the video material, the clipping information includes clipping time information and clipping content information, and obtain material slices according to the clipping information;

[0023] S2: Obtain material slices and use AI to classify video materials;

[0024] S3: Obtaining user attention information, which includes video length preferences, keyword sensitivity, eye-catching features, etc.;

[0025] S4: acquiring video materials according to the user's attention information, and combining the selected video materials using a combination strategy;

[0026] S5: Deliver the edited video to the target platform according to the delivery strategy.

[0027] The beneficial effects of the present invention are as follows: through user behavior analysis, it is possible to collect user viewing behavior data in real time, construct a user interest model, and update user key information in real time, which can accurately capture the user's personalized interests and provide precise guidance for the designation of food content. By utilizing the video generation module, according to the user's attention information, only video clips can be selected and combined from the video clip material library, and video content that meets the user's interests can be generated; by utilizing the delivery and feedback module, the edited video can be delivered to the target platform, and the delivery strategy can be adjusted in real time according to user feedback; the automatic processing of video content is realized, the video content production effect is improved, and the burden of manual processing is reduced. At the same time, through AI analysis and big data technology, the production and delivery of video content can be optimized to attract user attention. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is the overall structure diagram of the present invention;

[0029] Figure 2 It is a flow chart of the present invention;. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] It should be noted that when a component is referred to as being "fixed to" another component, it may be directly on the other component or there may also be a component centered. When a component is considered to be "connected to" another component, it may be directly connected to the other component or there may also be a component centered. When a component is considered to be "set on" another component, it may be directly set on the other component or there may also be a component centered. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0033] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings:

[0034] like Figure 1-2 As shown, the present invention provides a cloud-based big data real-time intelligent streaming system based on AI analysis, including a video preprocessing module, a user behavior analysis module, a video generation module, and a streaming and feedback module;

[0035] The video preprocessing module is used to obtain the original video material, and automatically segment and classify the original video material, and establish a video clip material library. It is used to obtain the original video material, and automatically segment the video material based on scene changes and key frames, classify the video material, and create a video clip material library. A dynamically updated label system is established for the video material, allowing multi-dimensional annotation and retrieval of the video material;

[0036] The user behavior analysis module is used to collect and analyze the user's video viewing behavior data in real time, and construct a user interest model. According to the user interest model, the user's attention information is updated in real time. Through the video playback platform, the user's viewing behavior data such as viewing time (the total time the user watches the video, the dwell time of the video clip), click-through rate (the user's clicks on the video), interactive behavior (such as likes, comments, shares, skips, pauses, etc.) are captured in real time. The collected viewing behavior data is cleaned, denoised and formatted, and valuable information is extracted. In-depth analysis of user behavior is performed, and users are divided into different interest groups through cluster analysis. The user's future viewing trend is predicted through time series. The potential connection between user behaviors is discovered through association rule mining. Analysis based on viewing behavior data As a result, a user interest model is constructed, which is an abstract representation of user preferences, interests, needs and other information. It can reflect the unique characteristics and rules of users in video content consumption. The user interest model can be in various forms, such as vector space model, subject model, neural network model, etc. Different models are selected to reflect the user's attention information through the user interest model; as the user's viewing behavior continues to occur, the user interest model needs to be continuously updated to reflect the user's interest changes. By implementing the detection of user behavior data, the parameters and weights of the user interest model are dynamically adjusted to ensure that the model can accurately reflect the user's latest concerns and needs. In the updating process, the time decay effect needs to be considered. The influence of earlier prosperous behaviors on current interests gradually weakens, while the most recent behaviors have stronger timeliness.

[0037] The video generation module is used to obtain user attention information, and use the user attention information to obtain the user's interest preferences, segmentation tendencies and possible content needs, and select video clips from the video clip library. Based on user data analysis, the video clips that match the user's preferences are automatically selected from the library, and the selected video materials are combined according to the combination strategy. The selected video clips are intelligently sorted and combined using the combination strategy, which requires content understanding of the video materials, rhythm control, transition effect selection, etc., and personalized editing of the combined video based on user information, such as Tianjie background music, subtitles, filters, special effects, etc., to enhance the attractiveness and expressiveness of the video;

[0038] The combined strategy includes:

[0039] Combination preset steps: Based on historical video data analysis, preset a series of effective material combination templates. By combining historical video data, such as viewing time, number of likes, number of comments, sharing rate, etc., identify which video material combination patterns are more popular with the audience. Based on these data, preset a series of funny video combination templates; Based on big data analysis tools, conduct statistical analysis on various indicators of historical videos to find out the arch features of successful videos. According to the analysis results, design a variety of different types of material combination templates, including but not limited to opening, transition, climax, ending and other parts, and test the template effect in a small range, and optimize and adjust according to feedback;

[0040] Steps for adjusting elements: using the material combination template, obtain the corresponding video material, create a video based on the video material, and obtain the corresponding video element requirements based on the user's attention information, and adjust the material in the material combination template; after obtaining the video material, combine it with the preset material combination template for preliminary creation; at the same time, according to the user's attention information and behavior data, make personalized adjustments to the material in the template to meet the needs of different user groups; according to the template requirements, select appropriate video, picture, audio and other materials from the material library, and construct a user portrait through the user's historical viewing records, interactive behavior and other information to understand the user's preferences. According to the user portrait, make personalized adjustments to the material in the template, such as changing the background music, adjusting the video rhythm, adding feature tags or elements, etc.;

[0041] Slicing and reorganization steps: Slice the selected video material, extract the parts that are more closely related to the video and reorganize them; refine the selected video material, extract the parts that are more closely related to the video theme through slicing, and recombine these clips to form a more compact and attractive video content; use video editing software to accurately cut the video material information, extract key frames or clips, and reorder and recombine the extracted clips according to the theme and rhythm of the video to ensure the continuity and attractiveness of the video. In the process of reorganization, add transition effects, subtitles and other elements to enhance the visual effect and viewing experience of the video.

[0042] The delivery and feedback module is used to deliver the edited video to the target platform and collect feedback information in real time. It also introduces an AI feedback model to generate delivery strategies based on data feedback. It supports delivering edited video content to multiple target platforms and uses API interfaces or crawler technology to collect user feedback information on each platform in real time, including but not limited to playback volume, likes, completion rate, and increase in fans. It also introduces AI technology to construct a feedback analysis model to mine and analyze the collected feedback information, identify user behavior patterns, emotional tendencies, and potential needs, and provide data results for the AI ​​feedback model to automatically adjust and optimize delivery strategies, such as delivery time, target audience positioning, content recommendation algorithms, etc., to increase food exposure and user interaction rate, and generate a detailed delivery effect evaluation report, including the changing trends of key indicators, summary analysis of user feedback, and optimization suggestions for delivery strategies, to provide data support for video delivery and editing.

[0043] The AI ​​feedback model is used to analyze and evaluate feedback information and provide corresponding feedback results. The combination strategy also includes a video template adjustment step. The video template adjustment step is used to obtain the material combination template and adjust the preset combination template according to the visual feedback information, including the content, sequence and duration of the material; based on user feedback, the video content, material sequence, video duration and other aspects are optimized to ensure that the video can accurately convey the dressing information and attract the target audience; the AI ​​feedback model is used to collect and analyze user feedback on the published videos, which is the key information in the user's behavior, such as which parts attract the audience and which parts cause user loss, and the collected feedback data is mined to identify It can identify user preferences, interests, and potential dissatisfaction, and use machine learning algorithms to analyze the relationship between user feedback and video content, sequence, duration, etc., to find potential correlations. Based on the analysis results, it can adjust the preset material combination template, including content adjustment, sequence adjustment, and duration adjustment. Content adjustment includes replacing or adding materials that are more in line with user interests, such as changing background music, changing background images, etc. Sequence adjustment is to rearrange the order of materials in the video to better guide user attention flow and improve viewing experience. Duration adjustment: According to the user viewing market and skip rate, the duration of each slice in the video is adjusted to ensure that important information is displayed in a moderate way while avoiding lengthy and boring parts.

[0044] A classification unit is provided in the video preprocessing module. The classification unit is provided with different classification rules. The classification rules include technical features, application scenarios, etc. According to different classification rules, the content of the video clips is analyzed and the video clips are classified. According to the classification results, the video material components are labeled.

[0045] Technical features include editable video length, divisible materials, and material relevance.

[0046] Duration can be edited, including classification principles, including classification according to video content,

[0047] The delivery strategies include: obtaining audience types based on user interest models, conducting targeted delivery based on the target audience's attention information, and continuously optimizing the delivery strategies based on data feedback results; dividing users into different audience types based on user interest models, and distinguishing audience types based on common interests, hobbies, needs, consumption habits and other factors to facilitate subsequent targeted delivery; summarizing the delivery process, collecting key data such as ad display and click-through conversion in real time, and entering into promotional analysis to understand the effectiveness of advertising delivery and changes in user behavior after exposure to advertising; and adjusting and optimizing the delivery strategies based on data feedback results.

[0048] The information to be paid attention to includes video length, user’s personal preferences and interests, consumption habits, sensitive words, etc.

[0049] A cloud-based big data real-time intelligent streaming method based on AI analysis.

[0050] S1: Obtain video material and editing information. Obtain original video material from various sources (such as video library, live stream, user upload, etc.), and obtain editing information of these materials. This information records in detail the start time, duration and specific content description of different segments in each material, providing basic data for subsequent video editing, analysis and combination.

[0051] S2: Use AI to classify video materials and input the acquired material slices into the AI ​​model. The model is trained to recognize video content and classify the video based on the content, thereby improving the efficiency of video processing and ensuring that subsequent combination and delivery can be based on accurate classification results.

[0052] S3: Obtain user attention information and collect users' video viewing preferences through user behavior analysis, historical data, etc., including their preferred video length, sensitivity to specific keywords, and video features that attract their attention such as special effects, soundtracks, storylines, etc., to ensure that the final video content can accurately match the interests and needs of the target audience.

[0053] S4: Combination strategy of video materials: According to the information that users pay attention to, select appropriate clips from the classified video materials, and use the combination strategy to splice these clips into a complete video. The combination strategy may include considering factors such as the coherence, emotional progression, and rhythm of the clips; through intelligent combination, generate video content that is both in line with user interests and attractive.

[0054] S5: Deliver videos according to delivery strategies; formulate delivery strategies, including determining target platforms, such as social media, video websites, apps, delivery time, target audiences, etc. Then, deliver the edited and assembled videos to the corresponding platforms according to the delivery strategies; ensure that the videos can reach the target audience at the most appropriate time and in the most effective way to achieve the best communication effect and advertising conversion rate.

[0055] The above are only preferred embodiments of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A cloud-based big data real-time intelligent streaming system based on AI analysis, characterized in that: It includes video preprocessing module, user behavior analysis module, video generation module, and delivery and feedback module; The video preprocessing module is used to obtain original video materials, automatically segment and classify the original video materials, and establish a video clip material library; The user behavior analysis module is used to collect and analyze the user's video viewing behavior data in real time, and to construct a user interest model, and to update the user's attention information in real time according to the user interest model; The video generation module is used to obtain user attention information, select video clips from the video clip material library, and combine the selected video materials according to the combination strategy; The delivery and feedback module is used to deliver the edited video to the target platform, collect feedback information in real time, introduce an AI feedback model, and generate a delivery strategy based on data feedback.

2. According to the cloud-based big data real-time intelligent streaming system based on AI analysis according to claim 1, it is characterized in that: The combined strategy includes: Combination preset steps: Based on historical video data analysis, a series of effective material combination templates are preset; Steps for adjusting elements: using the material combination template, obtaining corresponding video materials, and creating videos based on the video materials, and obtaining corresponding video element requirements based on user attention information, and adjusting the materials in the material combination template; Slicing and reassembly steps: Slice the selected video material, extract the parts that are more closely related to the video and reassemble them.

3. According to claim 2, a cloud-based big data real-time intelligent streaming system based on AI analysis is characterized in that: The AI ​​feedback model is used to analyze and evaluate based on the feedback information and provide corresponding feedback results. The combination strategy also includes a video template adjustment step. The video template adjustment step is used to obtain a material combination template and adjust the preset combination template based on the visual feedback information. The adjustment includes the content, sequence, and duration of the material.

4. According to the cloud-based big data real-time intelligent streaming system based on AI analysis as described in claim 1, it is characterized in that: The video preprocessing module is provided with a classification unit, and the classification unit is provided with different classification rules. The classification rules include technical features, application scenarios, etc. According to different classification rules, the content of the video clips is analyzed and the video clips are classified, and the video material components are labeled according to the classification results.

5. According to the cloud-based big data real-time intelligent streaming system based on AI analysis as described in claim 1, it is characterized in that: The technical features include editable video length, divisible materials, and material relevance.

6. The cloud-based big data real-time intelligent streaming system based on AI analysis according to claim 1 is characterized in that: The delivery strategy includes: obtaining the audience type according to the user interest model, conducting targeted delivery according to the target audience's attention information, and continuously optimizing the delivery strategy according to the data feedback results.

7. The cloud-based big data real-time intelligent streaming system based on AI analysis according to claim 1 is characterized in that: The feedback information includes the number of views, number of likes, completion rate, number of fans gained, etc.

8. The cloud-based big data real-time intelligent streaming system based on AI analysis according to claim 1 is characterized in that: The attention information includes video length, user's personal preferences and interests, consumption habits, sensitive words, etc.

9. A cloud-based big data real-time intelligent streaming method based on AI analysis, characterized in that: Said S1: Obtain video material and clipping information of the video material, the clipping information includes clipping time information and clipping content information, and obtain material slices according to the clipping information; S2: Obtain material slices and use AI to classify video materials; S3: Obtaining user attention information, which includes video length preferences, keyword sensitivity, eye-catching features, etc.; S4: acquiring video materials according to the user's attention information, and combining the selected video materials using a combination strategy; S5: Deliver the edited video to the target platform according to the delivery strategy.

Images