A multimedia data analysis method and a client

Abstract

The application discloses a kind of multimedia data analysis method and client, the method includes the following steps: obtaining multimedia file to be analyzed, according to preset frequency, the screenshot is carried out to the multimedia file, and the subject analysis is carried out to the screenshot;Each frame screenshot, main part, the total number of each main part and the multimedia file are sent to server, so that the server uses model to analyze the data received, and output analysis result.The application is processed by frame extraction and screenshot to multimedia file, and the subject analysis is carried out to the screenshot, identifies main part and sets up identification for main part, sends analysis result and multimedia file to server, so that the server uses model to analyze the data received and outputs analysis result, reduces the data amount of server, avoids occupying server resource, to reduce server calculation and analysis pressure, further improve the overall efficiency of system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, in particular to a multimedia data analysis method and client. BACKGROUND

[0002] At present, there are many picture recognition functions or video stream key information recognition functions on the market, such as video scanning content, video stream related identification and key annotation content functions. The main principle is that the client generates a video stream and transmits it to the server, and the server uses a large model to analyze the data of the video stream. The server needs to analyze all video streams or pictures, and the server pressure is multiplied. Therefore, there is an urgent need for a multimedia data analysis method and client that can solve the above problems. SUMMARY

[0003] The technical problem to be solved by the present application is to provide a multimedia data analysis method and client that can avoid occupying server resources, relieve server pressure, speed up server calculation and analysis, and further improve the overall efficiency of the system.

[0004] To solve the above technical problems, the technical scheme adopted by the present application is:

[0005] A multimedia data analysis method, comprising the steps of:

[0006] S1, obtaining a multimedia file to be analyzed, taking screenshots of the multimedia file according to a preset frequency, and performing subject analysis on the screenshots;

[0007] S2, sending each frame of screenshot, the subject part, the total number of each subject part and the multimedia file to the server, so that the server uses a model to analyze the received data and outputs the analysis result.

[0008] To solve the above technical problems, another technical scheme adopted by the present application is:

[0009] A multimedia data analysis client, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to realize each step of the above multimedia data analysis method.

[0010] The application has the beneficial effects that the application provides a multimedia data analysis method and client, acquires multimedia files to be analyzed, takes screenshots of the multimedia files according to a preset frequency, and performs subject analysis on the screenshots, identifies subject regions of the screenshots and sets an identifier for the identified subject parts, and sums up the total number of each identified subject part, which helps to improve the accuracy of subsequent server model identification and analysis, and sends the analysis results and the multimedia files to the server as input parameters of the server model, further analyzes by using the server model, fully utilizes the resources of the server, improves the depth and breadth of analysis, and thus outputs more accurate analysis results, effectively reduces the amount of data transmitted to the server by performing identification and preliminary analysis on the client side, avoids occupying the resources of the server, reduces the pressure of the server, speeds up the calculation and analysis of the server, and thus further improves the overall efficiency of the system. BRIEF DESCRIPTION OF DRAWINGS

[0011] Figure 1 A flowchart of a multimedia data analysis method according to an embodiment of the application;

[0012] Figure 2 A schematic diagram of a multimedia data analysis client according to an embodiment of the application;

[0013] Label explanation:

[0014] 1. A multimedia data analysis client; 2. A memory; 3. A processor. DETAILED DESCRIPTION

[0015] To describe the technical content, purposes and effects of the application in detail, the following will be described in combination with the embodiments and the accompanying drawings.

[0016] Please refer to Figure 1 The application provides a multimedia data analysis method, which includes the following steps:

[0017] S1. Acquire multimedia files to be analyzed, take screenshots of the multimedia files according to a preset frequency, and perform subject analysis on the screenshots.

[0018] S2. Send each frame of the screenshots, the subject parts, the total number of each subject part, and the multimedia files to the server, so that the server uses a model to analyze and output analysis results.

[0019] From the above description, the beneficial effects of the present application are that: the multimedia file to be analyzed is obtained, the multimedia file is screenshot according to the preset frequency, and the subject analysis is performed on the screenshot, the subject area of the screenshot is identified, and the identified subject part is labeled, which helps to improve the accuracy of subsequent server model identification and analysis, and the analysis result and the multimedia file are sent to the server as the input parameters of the server model, and further analysis is performed by using the server model, the resources of the server are fully utilized, the depth and breadth of analysis are improved, and more accurate analysis results are output, the identification and preliminary analysis are performed on the client side, the data amount transmitted to the server is effectively reduced, the server resources are avoided to be occupied, the server pressure is reduced, the server calculation and analysis are accelerated, and thus the overall efficiency of the system is further improved.

[0020] Further, the screenshot of the multimedia file according to the preset frequency and the subject analysis of the screenshot include:

[0021] The type of the multimedia file is judged, if the type of the multimedia file is a video stream, the video stream is screenshot according to the preset frequency, and the screenshot is subject analyzed.

[0022] From the above description, different processing methods are adopted according to different types of multimedia files, for the multimedia file of the video type, the video is converted into a picture for identification and analysis by screenshot according to the preset frequency, which not only ensures the integrity of the multimedia data, but also reduces the data amount for processing and improves the processing efficiency, in addition, the conversion can clearly show the key information in the video, reduce the analysis error caused by the video playing speed or frame rate, and help to improve the accuracy of data analysis.

[0023] Further, the preset frequency is dynamically set according to the video duration, if the video duration does not exceed the preset duration, the video stream is screenshot according to the preset frequency, if the video duration exceeds the preset duration, the exceeding proportion of the duration is calculated using the video duration and the preset duration, the preset frequency is reduced according to the exceeding proportion, and the video stream is screenshot according to the reduced frequency.

[0024] From the above description, the preset frequency is dynamically set according to the video duration, if the video duration does not exceed the preset duration, the preset frequency is used for frame extraction, otherwise the preset frequency is reduced and frame extraction is performed again, the dynamic setting is used to adapt to the uncertainty of the video duration, for long video, the preset frequency is reduced, the screenshot interval time is improved, the screenshot times are reduced, the data amount for processing is reduced, and the data analysis efficiency is improved.

[0025] Further, the operation of subject analysis on the screenshot specifically includes: identifying a subject area of the screenshot and setting an identifier for the identified subject part, and aggregating the total number of each identified subject part.

[0026] From the above description, by identifying the subject area of the screenshot, it is ensured that the subject part can be accurately and quickly identified, the identified subject part is set with an identifier to facilitate subsequent analysis and statistics, the total number of each identified subject part is aggregated, and the subject of the picture is finally determined, so that the identification and subject analysis of the multimedia file are realized, and a foundation is laid for subsequent server analysis using a large model and output of more accurate results.

[0027] Further, the operation of subject analysis on the screenshot specifically includes: identifying a subject area of the screenshot and setting an identifier for the identified subject part, and aggregating the total number of each identified subject part.

[0028] If the type of the multimedia file is a picture, the picture is directly subjected to subject analysis.

[0029] From the above description, according to different types of multimedia files, different processing methods are correspondingly adopted, for the multimedia file of the picture type, data analysis is directly performed, unnecessary frame extraction and screenshot processing are avoided, resource waste is reduced, and fast identification and subject analysis of the picture can be realized to obtain an analysis result.

[0030] Please refer to Figure 2 Another embodiment of the present application provides a multimedia data analysis client, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements each step of the above-mentioned multimedia data analysis method when executing the computer program.

[0031] The above-mentioned multimedia data analysis method and terminal of the present application are suitable for multimedia data analysis, and the following will be described through specific embodiments:

[0032] Embodiment one

[0033] Please refer to Figure 1 A multimedia data analysis method includes the following steps:

[0034] S1, a multimedia file to be analyzed is acquired, screenshots of the multimedia file are taken according to a preset frequency, and subject analysis is performed on the screenshots.

[0035] In the embodiment, the multimedia file is screenshotted at a preset frequency, effectively reducing the amount of data to be processed, significantly improving the processing speed, reducing the consumption of computing resources, taking the screenshots as key frames of the multimedia content, facilitating subsequent subject analysis and content identification, and making the entire analysis process more efficient.

[0036] Further, in the embodiment, the screenshotting of the multimedia file at a preset frequency and the subject analysis of the screenshots include: judging the type of the multimedia file, if the type of the multimedia file is a video stream, then frame extraction screenshotting of the video stream is performed at a preset frequency, and subject analysis of the screenshots is performed, if the type of the multimedia file is a picture, then subject analysis of the picture is directly performed, different processing methods are correspondingly adopted according to different types of multimedia files, for multimedia files of video type, the frequency is dynamically adjusted according to the video duration, frame extraction screenshotting is performed, the video is converted into a picture for identification and analysis, which not only ensures the integrity of the multimedia data, but also reduces the amount of data to be processed, improves the processing efficiency, and helps to improve the accuracy of data analysis, for multimedia files of picture type, data analysis is directly performed, unnecessary frame extraction screenshotting processing is avoided, resource waste is reduced, and rapid identification and preliminary analysis of the picture can be achieved to obtain the analysis result.

[0037] Further, in the embodiment, the preset frequency is dynamically set according to the video duration, if the video duration does not exceed the preset duration, then frame extraction screenshotting of the video stream is performed at the preset frequency, if the video duration exceeds the preset duration, then the exceeding proportion of the duration is calculated using the video duration and the preset duration, the frequency is reduced according to the exceeding proportion, and frame extraction screenshotting of the video stream is performed at the reduced frequency, through this dynamic setting of the video screenshot frequency to adapt to the uncertainty of the video duration, for long videos, the preset frequency is reduced, the screenshot interval time is increased, the screenshot times are reduced, the amount of data to be processed is reduced, and the data analysis efficiency is improved.

[0038] Further, in the embodiment, the subject analysis of the screenshots specifically includes: identifying the subject area of the screenshot and setting an identifier for the identified subject part, and summarizing the total number of each identified subject part, so as to accurately and quickly identify the subject part for subsequent analysis and statistics, and finally determine the subject of the picture, thereby realizing the identification and preliminary analysis of the multimedia file, and laying a foundation for subsequent analysis by the server using a large model and outputting more accurate results.

[0039] S2, send each frame of screenshot, subject part, total number of each subject part and the multimedia file to the server, so that the server uses a model to analyze the received data and outputs an analysis result.

[0040] In the embodiment, the analysis result and the multimedia file are sent to the server as input parameters of the server model for further analysis. For example, a video about cat playing with a duration of 100 seconds is frame-sampled at an interval of 5 seconds, and 20 frame-sampled pictures are obtained. The 20 pictures are analyzed for subjects. The subject parts of each picture are different, and the subject parts include cat, grassland, table and house. The identified subject parts are marked and summarized to obtain the total number of times each subject part is identified: cat = 20, table = 3, house = 2, and grassland = 5. Finally, the 20 pictures, the subject parts, the total number of times of each subject part, and the video are sent to the server. The resources of the server are fully utilized to improve the depth and breadth of analysis, so as to output more accurate analysis results. The recognition and subject analysis are performed on the client side, the amount of data transmitted to the server is effectively reduced, the server resources are avoided to be occupied, the server pressure is reduced, the server calculation and analysis are accelerated, and thus the overall efficiency of the system is further improved.

[0041] Embodiment Two

[0042] Please refer to Figure 2 A multimedia data analysis client 1 includes a memory 2, a processor 3, and a computer program stored in the memory 2 and executable on the processor 3. When the processor 3 executes the computer program, each step of the multimedia data analysis method of embodiment one is implemented.

[0043] In summary, the application provides a multimedia data analysis method and client, acquires a multimedia file to be analyzed, takes screenshots of the multimedia file according to a preset frequency, judges the type of the multimedia file, if the type of the multimedia file is a video stream, the preset frequency is dynamically set according to the video duration, when the video duration does not exceed the preset duration, the video is frame-extracted and screenshot according to the preset frequency, when the video duration exceeds the preset duration, the exceeding proportion of the duration is calculated using the video duration and the preset duration, the frequency is reduced according to the exceeding proportion, the video stream is frame-extracted and screenshot according to the reduced frequency, and the screenshots are subject analyzed, the video screenshot frequency is dynamically set to adapt to the uncertainty of the video duration, the data analysis efficiency is improved, if the type of the multimedia file is a picture, the picture is directly subject analyzed, different processing methods are taken according to different types of multimedia files, resource waste can be avoided, the subject area of the screenshots can be quickly identified and preliminarily analyzed, the identified subject parts are labeled, the total number of the identified subject parts is summarized, each frame of screenshot, subject part, total number of each subject part and the multimedia file are sent to a server, so that the server uses a model to analyze the received data and outputs an analysis result, the screenshot data is identified and subject analyzed on the client side, the amount of data transmitted to the server is effectively reduced, the server resources are avoided to be occupied, the server pressure is reduced, the server calculation and analysis are accelerated, and thus the overall system efficiency is further improved.

[0044] The above is only an embodiment of the application, and does not limit the patent scope of the application, any equivalent transformation or direct or indirect application in the related technical field using the content of the application specification and drawings is also included in the patent protection scope of the application.

Claims

1. A method of analyzing multimedia data, characterized by, The method comprises the steps of: S1, acquiring a multimedia file to be analyzed, taking screenshots of the multimedia file according to a preset frequency, and performing subject analysis on the screenshots; S2, sending each frame of screenshot, a subject part, a total number of each subject part, and the multimedia file to a server, so that the server uses a model to analyze the received data and outputs an analysis result; The step of taking screenshots of the multimedia file according to a preset frequency and performing subject analysis on the screenshots comprises: judging the type of the multimedia file, and if the type of the multimedia file is a video stream, taking screenshots of the video stream according to a preset frequency and performing subject analysis on the screenshots; The preset frequency is dynamically set according to the video duration, if the video duration does not exceed a preset duration, the video stream is taken screenshots according to the preset frequency, if the video duration exceeds the preset duration, the exceeding proportion of the duration is calculated using the video duration and the preset duration, the preset frequency is reduced according to the exceeding proportion, and the video stream is taken screenshots according to the reduced frequency; The operation of performing subject analysis on the screenshots specifically comprises: identifying the subject area of the screenshot and setting an identifier for the identified subject part, and summarizing the total number of each identified subject part.

2. The method of claim 1, wherein, The step of taking screenshots of the multimedia file according to a preset frequency and performing subject analysis on the screenshots further comprises: judging the type of the multimedia file, and if the type of the multimedia file is a picture, directly performing subject analysis on the picture. 3.An analytics client for multimedia data, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, The processor executes the computer program to implement the following steps: S1, acquiring a multimedia file to be analyzed, taking screenshots of the multimedia file according to a preset frequency, and performing subject analysis on the screenshots; S2, sending each frame of screenshot, a subject part, a total number of each subject part, and the multimedia file to a server, so that the server uses a model to analyze the received data and outputs an analysis result; The step of taking screenshots of the multimedia file according to a preset frequency and performing subject analysis on the screenshots comprises: judging the type of the multimedia file, and if the type of the multimedia file is a video stream, taking screenshots of the video stream according to a preset frequency and performing subject analysis on the screenshots; The preset frequency is dynamically set according to the video duration, if the video duration does not exceed a preset duration, the video stream is taken screenshots according to the preset frequency, if the video duration exceeds the preset duration, the exceeding proportion of the duration is calculated using the video duration and the preset duration, the preset frequency is reduced according to the exceeding proportion, and the video stream is taken screenshots according to the reduced frequency; The operation of performing subject analysis on the screenshots specifically comprises: identifying the subject area of the screenshot and setting an identifier for the identified subject part, and summarizing the total number of each identified subject part.

4. The multimedia data analysis client of claim 3, wherein, The step of taking screenshots of the multimedia file according to a preset frequency and performing subject analysis on the screenshots further comprises: judging the type of the multimedia file, and if the type of the multimedia file is a picture, directly performing subject analysis on the picture.

Images