A method for constructing a multi-platform video sample library

By building a multi-platform video sample library, extracting and downloading video information for different video platforms and streaming media transmission protocols, the platform limitation problem of video transmission status measurement research in the existing technology is solved, and the multi-platform and multi-protocol Video data acquisition and management.

CN114925222BActive Publication Date: 2024-11-08SOUTHEAST UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210552440.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-05-20
Publication Date
2024-11-08
Estimated Expiration
2042-05-20

AI Technical Summary

Technical Problem

The lack of a multi-platform video sample library in the existing technology has led to platform limitations in the research on video transmission status measurement and management, which cannot meet the diversity needs of multi-video platforms and streaming media transmission protocols.

Method used

Build a multi-platform video sample library, and through the video information extraction module and the download module, it is responsible for the extraction and download of video information, and personalized construction for different video platforms and streaming media transmission protocols, obtain the description information, attribute information and transmission information of the video, and form Video sample library.

Benefits of technology

The video information database and video library construction in multi-video platforms and multi-streaming protocol scenarios is realized, providing complete data support for video recognition, with good independence and expansion, and adapting to changes in different video platforms and protocols.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114925222B_ABST
    Figure CN114925222B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for constructing a multi-platform video sample library. According to the characteristics of different video platforms and streaming media transmission protocols, video information extraction modules and download modules in different scenarios are constructed, which are responsible for video information extraction and video downloading respectively. On a client that can access the Internet, video URLs are processed in turn, and attribute information and transmission information of the video are obtained to construct a video sample library. The two processing modules in the present invention are personalized according to the actual video platform and streaming media protocol, and have good independence and extensibility. The video sample library in the present invention is composed of a video information database and a video database. The video information database stores data such as description information, attribute information, and transmission information of all videos, and the video database stores all video files. The present invention can be used for the construction of video information libraries and video libraries in scenarios of multiple video platforms and multiple streaming media protocols, and provide a basis for video recognition.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a method for constructing a multi-platform video sample library, and belongs to the technical field of network measurement. Background Art

[0002] With the rapid development of Internet technology and multimedia technology, anyone can become a video content creator. Massive amounts of videos are uploaded and played on the Internet every day. The number of clicks on major video platforms such as YouTube, FaceBook, iQiyi, and Tencent Video continues to rise. Videos have penetrated into every aspect of people's lives and have become an indispensable and important data object in the era of big data.

[0003] As Internet video data explodes exponentially, the demand for measuring and managing network videos is also increasing. On the one hand, network managers need to understand the status of user video playback in order to provide a good network access environment for video users. On the other hand, the distribution status of different types and platforms of videos in the network is also the basis for pipeline service providers to provide customized services for the platform. Therefore, the measurement of video transmission status has become an important requirement for network management.

[0004] Some studies have studied the video distribution and transmission mechanisms of specific video platforms, such as identifying the video resolution, playback jams, and the platform to which the video belongs by analyzing network traffic. These studies are all focused on videos on specific platforms. Because there are a large number of videos and platforms on the Internet, different video platforms use different transmission protocols. Most videos are transmitted in encrypted form, and the encryption protocols used are also different. The container formats of video files are also different. These differences require researchers to first collect data from the corresponding platform before conducting research, and data collection requires a lot of manpower and resources. Due to the lack of sample data from multiple platforms, existing studies have certain platform limitations.

[0005] Establishing a multi-platform video sample library is a solution to this problem. In the field of network measurement research, there are already many other authoritative and highly credible public measurement sample data sets, such as active measurement data sets "IPv6 ASLinks" and "AS Rank", wireless network data sets "buffalo / phonelab-wifi" and "cambridge / haggle", malicious code data sets "VirusShare" and "MalwareBazaar", etc.; in the field of video behavior recognition, the newly released "Kinetics" and "Something-Something" data sets in recent years have also been widely used. As an important data object in network measurement research, video traffic also needs a public sample library, but there is currently no multi-platform video sample library. Therefore, it is urgent to build a multi-platform video sample library that can be used for video measurement research.

[0006] Although there are some video libraries now, most of these video libraries are built for user groups, and their main purpose is to optimize the quality of video retrieval and improve the efficiency of video retrieval. Most of them are built for video resources on a single platform or in a specific field, such as the construction of course libraries for MOOC platforms and the construction of university physics experiment video resource libraries. From the perspective of video network measurement research, these video libraries are not applicable, and their content is not sufficient to support the measurement and management research of video transmission in the field of network measurement.

[0007] The multi-platform video sample library used for measuring and studying videos on the Internet needs to include not only the information of the video itself, but also comprehensive video-related data such as video description information, attribute information, and transmission information to provide clues for related research. Among the existing related patents, most of them are aimed at a single streaming media transmission protocol, and the actual application scenarios are relatively fixed, and cannot meet the needs of various video platform downloads; their purpose is mainly to download, store and play videos, and they do not pay attention to the description information, attribute information and transmission information of the video, etc., and there is less information for video measurement research.

[0008] The application of the multi-platform video sample library construction method provided by the present invention can solve the difficulty of lacking sufficient video research sample data in the current network measurement field. Summary of the invention

[0009] In order to solve the above problems, the present invention discloses a method for constructing a multi-platform video sample library. According to the characteristics of different video platforms and streaming media transmission protocols, video information extraction modules and download modules are constructed in different scenarios, which are responsible for video information extraction and video downloading respectively. On the client that can access the Internet, the video URL is processed in turn, and the attribute information and transmission information of the video are obtained to construct a video sample library. The two processing modules in the present invention are personalized according to the actual video platform and streaming media protocol, and have good independence and scalability; the video sample library in the present invention is composed of a video information database and a video database, and the video information database stores data such as description information, attribute information, and transmission information of all videos, and the video database stores all video files. The present invention can be used for the construction of video information libraries and video libraries in scenarios of multiple video platforms and multiple streaming media protocols, and provide a basis for video recognition work.

[0010] In order to achieve the purpose of the present invention, the specific technical steps of this scheme are as follows: A method for constructing a multi-platform video sample library, the method comprising the following steps:

[0011] Step (1) constructing a personalized video platform extractor and downloader for different video platforms and streaming media transmission protocols, obtaining an extractor list and a downloader list, wherein the extractor is responsible for extracting valid information from the video web page, and the downloader is responsible for downloading the encrypted video to the local computer according to the characteristics of the streaming media transmission protocol;

[0012] Step (2) According to the architecture and layout characteristics of the video platform, formulate the retrieval and crawling strategy of each platform, write the URL crawling program, and build the URL list;

[0013] Step (3) according to the URL list, select a suitable extractor from the extractor list to obtain description information related to the video, and extract information such as the download address of the video resource and the streaming media transmission protocol;

[0014] Step (4) based on the download address of the video resource and the streaming media transmission protocol information, select a suitable video downloader. If it is the DASH protocol, proceed to step (5); if it is the HLS protocol, proceed to step (6);

[0015] Step (5) uses a DASH downloader to download the video file, parse the video file, and extract the attribute information of the video, wherein the segment information indirectly reflects the characteristics of the DASH protocol video transmission;

[0016] Step (6) Use the HLS downloader to download the video file, parse the video file to obtain the attribute information of the video, and during the download process, directly obtain relevant information reflecting the video transmission characteristics of the HLS protocol;

[0017] Step (7) constructs a video sample library, which consists of a video information database and a video database. The video information database includes description information, attribute information and transmission information of all video files, and the video database includes all video files.

[0018] Furthermore, in step (1), according to the different video platforms and streaming media transmission protocols, the steps for constructing the extractor and downloader are as follows:

[0019] (1.1) Different video platforms have different web page information, and the effective information related to the video is different. For specific video platforms, a personalized extractor is built to extract video information that can be processed by the downloader, and an extractor list is built based on the extractor;

[0020] (1.2) Different video platforms use different streaming media transmission protocols. Build personalized downloaders for different protocols, use the downloaders to download videos, and build a downloader list based on the downloaders.

[0021] Furthermore, in step (2), the crawling strategy of the video URL is as follows:

[0022] (2.1) For video platforms that use static loading, the video web page HTML is processed through the parsing library to obtain the required target information and extract the video URL;

[0023] (2.2) For dynamically loaded video platforms, you can use browser developer tools or automated testing libraries to simulate user browsing behavior, and use third-party libraries to parse the response data returned by the web page to extract the video URL;

[0024] (2.3) Through steps (2.1) and (2.2), the URLs of the current mainstream video platforms can be captured, and all video URLs can be aggregated to obtain a video URL list.

[0025] Furthermore, in step (3), the steps of extracting video information are as follows:

[0026] (3.1) Determine the video platform to which the URL belongs, select the corresponding extractor from the extractor list, download the webpage source code of the current URL, and extract the video description information as comprehensively as possible according to the architecture and layout of the video platform, including but not limited to the video ID, title, duration, format and other information;

[0027] (3.2) Analyze and process the extracted information to obtain different format information of the same video, mainly including the video resource download address and streaming media transmission protocol, etc. You can select several video formats according to your needs and pass them to the downloader.

[0028] Furthermore, in step (4), the streaming media transmission protocol used by the video platform is determined based on the video download address and protocol information passed to the downloader, and the corresponding video downloader is selected. The difference in video downloaders determines the way of obtaining video transmission information. If it is the DASH protocol, the information reflecting the video transmission can be obtained through the video parser after the video download is completed. If it is the HLS protocol, the information related to the video transmission can be obtained during the video download process. The downloaders are independent of each other and can be dynamically expanded to cover the current mainstream transmission protocols.

[0029] Furthermore, in step (5), the video downloaded through the DASH downloader can be used by a video parser, such as an fmp4 parser to extract the encoding format, bit rate, frame rate and other attribute information of the mp4 file, wherein the "sidx box" records the relevant information of the segments of each moof+mdat combination, including size and duration, which indirectly reflects the transmission characteristics of the video.

[0030] Furthermore, in step (6), the HLS downloader obtains the complete video by requesting the link of each fragment in the m3u8 file, and obtains the video's encoding format, bit rate, frame rate and other attribute information by parsing the video file. During the download process, the size of each fragment is saved in sequence, and information related to video transmission can be directly obtained. The video download and transmission information can be obtained synchronously, saving time and resources.

[0031] Furthermore, in step (7), the video description information, video attribute information and video transmission information are collated and merged, and all video related information is stored in the information library to complete the construction of the video information database; all video files (in formats such as mp4, webm, etc.) are stored in the video library to complete the construction of the video database.

[0032] Compared with the prior art, the technical solution of the present invention has the following beneficial technical effects.

[0033] (1) The present invention proposes a method for constructing a video sample library in a scenario of multiple video platforms and multiple streaming media transmission protocols. Due to the diversity of video platforms and streaming media transmission protocols, the present invention achieves the purpose of video information extraction and video download by customizing a video information extraction module and a download module for a specific video platform and the streaming media transmission protocol used therein. The two functional modules and different video platforms have a certain degree of independence. When a version of a certain platform is updated, it does not affect the information extraction and video download of other platforms. The feasibility of its practical application is relatively high.

[0034] (2) Through the method proposed in the present invention, while extracting video information and establishing a video database, the transmission information of the video can be further obtained. Videos downloaded by the DASH protocol can obtain relevant information that indirectly reflects video transmission by parsing video files, while videos downloaded by the HLS protocol can directly obtain information related to video transmission by saving the size of each segment during the video download process. Therefore, a video information library including video description information, attribute information, and transmission information can be synchronously established. The video information database and the video database together constitute a video sample library, which stores more comprehensive video-related data including video description information, attribute information, transmission information, and video files;

[0035] (3) The video information extraction module proposed in the present invention is constructed for a specific video platform, and the video download module is constructed for a specific streaming media transmission protocol. In practice, the number of video platforms is much larger than the number of streaming media transmission protocols. When a video platform needs to be added, only the extraction module of the video platform needs to be added. If the streaming media transmission protocol used by the platform is already included in the list of video download modules, it can be used directly. Otherwise, an additional download module needs to be constructed. Therefore, the method proposed in the present invention has good scalability and can meet the situation where the scene is constantly changing in actual applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic diagram of the system structure of a method for constructing a multi-platform video sample library;

[0037] Figure 2 Download flowchart for video;

[0038] Figure 3 Obtaining flow chart for video transmission information;

[0039] Figure 4 This is a schematic diagram of the video sample library structure. DETAILED DESCRIPTION

[0040] The technical solution provided by the present invention will be described in detail below in conjunction with specific embodiments. It should be understood that the following specific implementation methods are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0041] Specific embodiment: The present invention provides a method for constructing a multi-platform video sample library, and its overall system structure is as follows: Figure 1 As shown, the following steps are included:

[0042] Step (1) constructing a personalized video platform extractor and downloader for different video platforms and streaming media transmission protocols, obtaining an extractor list and a downloader list, wherein the extractor is responsible for extracting valid information from the video web page, and the downloader is responsible for downloading the encrypted video to the local computer according to the characteristics of the streaming media transmission protocol;

[0043] In one embodiment of the present invention, according to different video platforms and streaming media transmission protocols, the steps of constructing the extractor and the downloader are as follows:

[0044] (1.1) Different video platforms have different web page information, and the effective information related to the video is different. For specific video platforms, a personalized extractor is built to extract video information that can be processed by the downloader, and an extractor list is built based on the extractor. Table 1 is an example of a list of domestic and foreign video platforms, and extractors are built for these platforms respectively.

[0045] Table 1: List of domestic and foreign video platforms

[0046]

[0047]

[0048] (1.2) Different video platforms use different streaming media transmission protocols. We build personalized downloaders for different protocols, use the downloaders to download videos, and build a downloader list based on the downloaders. Table 2 is a list of streaming media transmission protocols, and we build downloaders for these protocols.

[0049] Table 2: List of streaming media transmission protocols

[0050]

[0051] Step (2) According to the architecture and layout characteristics of the video platform, formulate the retrieval and crawling strategy of each platform, write the URL crawling program, and build the URL list;

[0052] In one embodiment of the present invention, the crawling strategy of the video URL is as follows:

[0053] (2.1) For video platforms that use static loading, the video web page HTML is processed through the parsing library to obtain the required target information and extract the video URL;

[0054] (2.2) For dynamically loaded video platforms, you can use browser developer tools or automated testing libraries to simulate user browsing behavior, and use third-party libraries to parse the response data returned by the web page to extract the video URL;

[0055] (2.3) Through steps (2.1) and (2.2), the URLs of the current mainstream video platforms can be captured, and all video URLs can be aggregated to obtain a video URL list.

[0056] Step (3) According to the URL list, select a suitable extractor from the extractor list to obtain the description information related to the video, and extract the download address of the video resource and the streaming media transmission protocol and other information. Table 3 shows some video description information fields and descriptions:

[0057] Table 3: Some video description information fields and descriptions

[0058]

[0059]

[0060] In one embodiment of the present invention, the steps of extracting video information are as follows:

[0061] (3.1) Determine the video platform to which the URL belongs, select the corresponding extractor from the extractor list, download the webpage source code of the current URL, and extract the video description information as comprehensively as possible according to the architecture and layout of the video platform, including but not limited to video ID, title, duration, format and other information; Table 4 shows some video description information:

[0062] Table 4: Partial video description information

[0063]

[0064]

[0065] (3.2) Analyze and process the extracted information to obtain different format information of the same video, mainly including the video resource download address and streaming media transmission protocol, etc. You can select several video formats according to your needs and pass them to the downloader.

[0066] Step (4) Select a suitable video downloader based on the download address of the video resource and the streaming media transmission protocol information. If it is the DASH protocol, proceed to step (5); if it is the HLS protocol, proceed to step (6). Figure 2 Indicates the process of video downloading;

[0067] In one embodiment of the present invention, the streaming media transmission protocol used by the video platform is determined based on the video download address and protocol information passed to the downloader, and the corresponding video downloader is selected. The difference in video downloaders determines the way of obtaining video transmission information. If it is the DASH protocol, the information reflecting the video transmission can be obtained through the video parser after the video download is completed. If it is the HLS protocol, the information related to the video transmission can be obtained during the video download process. The downloaders are independent of each other and can be dynamically expanded to cover the current mainstream transmission protocols. Figure 3 Represents the process of obtaining video transmission information.

[0068] Step (5) uses a DASH downloader to download the video file, parse the video file, and extract the attribute information of the video, wherein the segment information indirectly reflects the characteristics of the DASH protocol video transmission;

[0069] In one embodiment of the present invention, the video downloaded through the DASH downloader can be used by a video parser, such as an fmp4 parser to extract the encoding format, bit rate, frame rate and other attribute information of the mp4 file, wherein the "sidx box" records the relevant information of the fragments of each moof+mdat combination, including size and duration, etc., which indirectly reflects the transmission characteristics of the video.

[0070] Step (6) uses the HLS downloader to download the video file, parses the video file to obtain the attribute information of the video, and during the download process, the relevant information reflecting the HLS protocol video transmission characteristics can be directly obtained. Table 5 shows the fields and descriptions of some video attribute information and transmission information:

[0071] Table 5: Some video attributes, transmission information fields and descriptions

[0072]

[0073]

[0074] In one embodiment of the present invention, the HLS downloader obtains the complete video by requesting the link of each fragment in the m3u8 file, and obtains the attribute information of the video such as the encoding format, bit rate, frame rate, etc. by parsing the video file. The size of each fragment is saved in sequence during the download process, and the information related to the video transmission can be directly obtained. The video download and the acquisition of the transmission information can be obtained synchronously, saving time and resources. Table 6 shows some video attribute information and transmission information:

[0075] Table 6: Some video attributes and transmission information

[0076]

[0077]

[0078] Step (7) constructs a video sample library, which consists of a video information database and a video database. The video information database includes description information, attribute information and transmission information of all video files, and the video database includes all video files.

[0079] In one embodiment of the present invention, video description information, video attribute information and video transmission information are collated and merged, and all video related information is stored in an information library to complete the construction of a video information database; all video files (in formats such as mp4, webm, etc.) are stored in a video library to complete the construction of a video database. Figure 4 It is a structural diagram of the video sample library. There is a one-to-one correspondence between the data in the information library and the video library.

[0080] The technical means disclosed in the scheme of the present invention are not limited to the technical means disclosed in the above-mentioned implementation mode, but also include technical schemes composed of any combination of the above-mentioned technical features. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications are also regarded as the protection scope of the present invention.

Claims

1. A method for constructing a multi-platform video sample library. It is characterized in that The method comprises the following steps: Step (1) constructing a personalized video platform extractor and downloader for different video platforms and streaming media transmission protocols, obtaining an extractor list and a downloader list, wherein the extractor is responsible for extracting valid information from the video web page, and the downloader is responsible for downloading the encrypted video to the local computer according to the characteristics of the streaming media transmission protocol; Step (2) According to the architecture and layout characteristics of the video platform, formulate the retrieval and crawling strategy of each platform, write the URL crawling program, and build the URL list; Step (3) according to the URL list, select a suitable extractor from the extractor list to obtain description information related to the video, and extract the download address and streaming media transmission protocol information of the video resource; Step (4) based on the download address of the video resource and the streaming media transmission protocol information, select a suitable video downloader. If it is the DASH protocol, proceed to step (5); if it is the HLS protocol, proceed to step (6); Step (5) uses a DASH downloader to download the video file, parse the video file, and extract the attribute information of the video, wherein the segment information indirectly reflects the characteristics of the DASH protocol video transmission; Step (6) using an HLS downloader to download the video file, parsing the video file to obtain the attribute information of the video, and directly obtaining relevant information reflecting the video transmission characteristics of the HLS protocol during the download process; Step (7) constructs a video sample library, which consists of a video information database and a video database. The video information database includes description information, attribute information and transmission information of all video files, and the video database includes all video files.

2. A method for constructing a multi-platform video sample library according to claim 1, It is characterized in that The step (1) specifically comprises the following sub-steps: (1.1) Different video platforms have different web page information, and the effective information related to the video is different. For specific video platforms, a personalized extractor is built to extract the video information processed by the downloader, and an extractor list is built based on the extractor; (1.2) Different video platforms use different streaming media transmission protocols. Build personalized downloaders for different protocols, use the downloaders to download videos, and build a downloader list based on the downloaders.

3. The method for constructing a multi-platform video sample library according to claim 1, It is characterized in that The step (2) specifically comprises the following sub-steps: (2.1) For video platforms that use static loading, the video web page HTML is processed through the parsing library to obtain the required target information and extract the video URL; (2.2) For dynamically loaded video platforms, use browser developer tools or automated testing libraries to simulate user browsing behavior, and use third-party libraries to parse the response data returned by the web page to extract the video URL; (2.3) Through steps (2.1) and (2.2), the URLs of the current mainstream video platforms can be captured, and all video URLs can be aggregated to obtain a video URL list.

4. The method for constructing a multi-platform video sample library according to claim 1, It is characterized in that The step (3) specifically comprises the following sub-steps: (3.1) Determine the video platform to which the URL belongs, select the corresponding extractor from the extractor list, download the webpage source code of the current URL, and extract the video description information as comprehensively as possible according to the architecture and layout of the video platform, including but not limited to the video ID, title, duration, format and other information; (3.2) Analyze and process the extracted information to obtain different format information of the same video, mainly including the video resource download address and streaming media transmission protocol, etc. You can select several video formats according to your needs and pass them to the downloader.

5. The method for constructing a multi-platform video sample library according to claim 1, It is characterized in that In the step (4), the streaming media transmission protocol used by the video platform is determined based on the video download address and protocol information passed to the downloader, and the corresponding video downloader is selected. The difference in video downloaders determines the way of obtaining video transmission information. If it is the DASH protocol, the information reflecting the video transmission is obtained through the video parser after the video download is completed. If it is the HLS protocol, the information related to the video transmission is obtained during the video download process. The downloaders are independent of each other and can be dynamically expanded to cover the current mainstream transmission protocols.

6. The method for constructing a multi-platform video sample library according to claim 1, It is characterized in that In the step (5), the video downloaded by the DASH downloader is used to extract the encoding format, bit rate, and frame rate attribute information of the mp4 file through the video parser and the fmp4 parser, wherein the "sidx box" records the relevant information of the fragments of each moof+mdat combination, including size and duration, which indirectly reflects the transmission characteristics of the video.

7. The method for constructing a multi-platform video sample library according to claim 1, It is characterized in that In step (6), the HLS downloader obtains the complete video by requesting the link of each segment in the m3u8 file, obtains the encoding format, bit rate, and frame rate attribute information of the video by parsing the video file, and directly obtains the information related to the video transmission by sequentially saving the size of each segment during the download process. The video download and the acquisition of the transmission information can be obtained synchronously, saving time and resources.

8. The method for constructing a multi-platform video sample library according to claim 1, It is characterized in that In the step (7), the video description information, video attribute information and video transmission information are sorted and merged, and all video related information is stored in the information library to complete the construction of the video information database; all video files are stored in the video library to complete the construction of the video database.