JP2003288361A - Content distribution system and content distribution method - Google Patents

Abstract

(57) [Summary] [Problem] There is a demand for distribution of streaming media suitable for a user. A Web server includes a communication unit, an application unit, a user management unit, a streaming data encoding unit, and a content holding unit.
including. The communication unit 30 communicates with the user terminal 12 and the Web server 14 via the Internet 18. The application unit 50 includes a registration unit 32, a list display unit 34,
Database interface 38, SMIL output unit 6
0 and a SMIL generation unit 62.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a streaming media distribution technique for developing a multi-story for each user.

[0002]

2. Description of the Related Art In recent years, the Internet has spread rapidly,
A wide variety of information is flooding the WEB. In order for the user to obtain the information he / she wants, he / she has to select and sort out the flooded information. Also, in the future, broadband will be accelerated in the development of the Internet infrastructure. Under such circumstances, broadband content, which is represented by video media content such as streaming data, is attracting attention both in electronic commerce and in private. As the network speeds up, users will demand higher quality content, but if the quality of the media is only improved, it is similar to reproducing the current terrestrial TV on the Internet. It can be said that it is not tied to quality improvement in the sense of meaning.

The streaming distribution server is basically designed only for flowing data to the user. Therefore, the management form of the streaming media need only upload the streaming media to the streaming distribution server. Therefore, it is not possible to specify the content information such as the owner, producer, genre, etc. of the streaming media. Further, even when managing streaming media, it is inconvenient if there is no various content information.

[0004]

Present streaming distributions are live and on-demand. Live
Although this is a distribution method that takes advantage of the characteristics of globality and scalability of the Internet, it is difficult for the distribution side to instantly grasp what kind of content the user is currently browsing. On-demand, as well as existing storage media on the Internet, may even promote the flood of information. One of these factors is that, unlike the conventional text-based contents, the distribution server and the user cannot grasp the information itself included in the streaming media. In addition, the distribution method has a one-source / one-use relationship as in the case of the conventional contents.

The present invention has been made in view of such circumstances, and an object thereof is to provide a technique for distributing streaming media suitable for users. In particular, the present invention aims to provide a streaming media distribution technology that pays attention to user preference information, personalizes and optimizes it based on individual preferences, and develops it in multi-story for each user.

[0006]

One aspect of the present invention relates to a content distribution system. The content distribution system has versatility by incorporating moving image contents that are likely to be used by various users, and holds a template described in the markup language format by the user as a general-purpose scenario. And a user attribute acquisition unit that acquires a user attribute, a content specifying unit that specifies a plurality of moving image contents suitable for the user to be distributed from the general-purpose scenario based on the acquired user attribute, and the acquired user attribute Based on, the user acquires the specified moving image content,
A scenario editing unit that rewrites general-purpose scenarios so that they can be played back in a predetermined order and creates individual scenarios for each user,
A scenario distribution unit that distributes the rewritten individual scenario to the user.

As a markup language, XML (eXtens
ible Markup Language) is exemplified. This XML is H
The document structure can be described in a simple format such as TML (Hypertext Markup Language), and tags can be defined independently. Therefore, by defining a conditional statement such as “if conditional statement” and providing an interpreter function, a description of a scenario having a branch structure in XML format is realized.

In the versatile template, link information to all moving image contents is described in a scenario format. Only the link information suitable for the user is specified from the link information according to the attribute of the user, and the individual scenario customized for each user is generated. This customized personalized scenario is sent to the user,
The user acquires the moving image content as streaming data based on this individual scenario. That is, the multi-story is developed for each user.

As the user attributes, age, sex,
Examples include occupations. For example, if the user is a man,
Content that is commercial for women's cosmetics is generally unnecessary.

Further, the scenario editing unit leaves the moving image content suitable for the user among the moving image contents described in the general-purpose scenario, and physically deletes the other moving-image contents to set the general-purpose scenario. Individual scenarios may be generated by making them compact. For example, in the general-purpose scenario described in the XML format and having the branch structure as described above, the selected moving image content remains and the conditional statement branching the scenario is deleted. X
In the program described in the ML format, link information and tags for including the link information are described line by line. Therefore, unnecessary descriptions can be deleted line by line.

A plurality of possible scenarios may be prepared in advance, but if there are many branches, the number of scenarios to be prepared will increase, and management may become complicated. Therefore, it is effective to dynamically generate a scenario according to the user's attribute.

The content distribution system further includes a session management unit that manages a history of moving image contents accessed by the user based on the distributed individual scenario, and the content specifying unit is a moving image managed by the session management unit. The moving image content suitable for the user may be estimated by referring to the attribute and history of the image content. Here, examples of the attribute relating to the moving image content include a length of time of the moving image content and a field to which sports or the like belong.

Further, the markup language is XML, and in the general-purpose scenario, the scenario may be described in a branch structure using if conditional statements, and the condition for branching the scenario may be an attribute of the user. As a result, the generated individual scenario is a scenario in which the if conditional statement and the link information that has not been specified are deleted and only the specified link information is described. However, the conditional statement that realizes the branch structure is not limited to the if conditional statement, and it is sufficient that the branch structure can be described in XML.

[0014] Another aspect of the present invention relates to a content distribution method. This content distribution method was acquired based on a general-purpose scenario, a step of retaining a general-purpose scenario in which link information to a wide range of moving image content is described in a markup language format, a step of acquiring a user attribute. A step of identifying moving image content suitable for the user according to the attribute of the user, leaving the moving image content identified as suitable, and physically deleting the description of the link information to other content. And generating a customized secondary scenario for each user.

It should be noted that any combination or rearrangement of the above constituent elements, and the expression of the present invention as a method, a computer program or the like is also effective as an aspect of the present invention.

[0016]

BEST MODE FOR CARRYING OUT THE INVENTION In the present embodiment, in streaming media distribution on the Internet, personal preference information that develops on a time axis is acquired and analyzed, and further, for media data, a meta data specialized for multimedia contents is provided. Add data.

Further, in distributing streaming media on the Internet, the streaming media distributed for each user are personalized and optimized. Specifically, description data for streaming distribution is prepared in advance for all users,
It analyzes user preference data to be distributed and user interactive sessions, and dynamically generates description data for streaming distribution.

In addition, multi-use of media data,
Content information such as the owner, creator, and genre of media data is added to the media data as metadata for the purpose of improving the function of media management. The metadata is described in XML format. Since XML is a markup language, data can be easily detected and extracted. XML can be used in many applications as long as the data structure is fixed.

In this embodiment, SMIL (Synchronized Multimedia Integration Lang) is used for streaming distribution.
uage) is used. SMIL can synchronize streaming data such as moving images and sounds on the Web with text. In addition, since SMIL conforms to XML, it is easy to maintain data compatibility.

The description data for streaming media distribution is described by a scenario description method to form a story. The story is first divided into a plurality of chapters, and each chapter has one or a plurality of scenarios, and the optimum scenario is specified in each chapter according to the user's attribute. This scenario is the minimum unit of moving image content, and finally specified scenarios are compactly put together and sent to the user. The user acquires the moving image content described in the scenario acquired using a predetermined application.

First, a basic platform for realizing the streaming distribution system will be described. This platform aims to analyze personal information and content information and dynamically personalize and optimize streaming data to be distributed. The main functions of the platform are six types of streaming data distribution server, web server, user session monitoring, personal preference analysis, media data management, and automatic reconstruction of distribution data. By using these,
Streaming data is personalized and optimized for individual users.

A series of sessions is a HTTP (Hypertext Transfer Protocol) for the user's Web server.
It is triggered by access using the protocol. In this embodiment, one session in streaming media distribution is handled in units of projects. A project is composed of one or more contents, and the flow of this project is defined as a project story. In other words, personalization of distribution data means that the story changes dynamically, and optimization means adjustment of the size of the project.

The Web server delivers the project in response to the request while monitoring the session. Content within the project is roughly classified and defined into two types. One is a streaming media that is media data having a time axis, which is characteristic of this embodiment. The other is H, which has been used conventionally and has no time axis.
The contents are TML and still images. The latter is used for navigation when the user selects streaming media and for supplementing information in conjunction with the streaming media.

The above-mentioned six types of functions will be briefly described. The streaming data delivery server delivers streaming media. On this platform,
RTP (Real-Time Transport Protocol), RTSP
(Real-Time Streaming Protocol) protocol can be handled,
The server also supports SMIL.

The Web server supplements information in conjunction with navigation to the user and streaming data. The user session monitoring function monitors a series of actions of the user after accessing the Web server, analyzes actions in the session, and accumulates action history.
In addition, by including project information and its progress as session information, streaming delivery with multi-story characteristics is realized in sessionless HTTP communication.

The personal taste analysis function analyzes the user's taste and situation based on the active information obtained from the session and the private information such as hobbies and characteristics acquired by the questionnaire at the time of the first access, and the story of the project. Generate preference information for personalization in generation.

The media data management function manages media data by defining metadata that describes the content of multimedia content such as streaming media and images. A content management system described later is used for this management. Also,
By defining the position of each content in the project, the story can be personalized. By using these two data, media data can be comprehensively managed and distributed. The distribution data automatic reconstruction function composes a project story based on user session information, preference information, and media data information.

Next, media management will be described. The scene data is described in XML format, and the metadata of the content file used in each scene is described. It is composed of a clip information section that describes link information to a content file used in one scene, and a layout section that describes its layout information.

One project is formed from scenario data. The scenario data is described in XML format and includes a plurality of scene data. The scenario data consists of a layout section that describes the definition of the viewport and a flow section that describes the flow of the story. The scene XML path is described in the flow portion.

In this system, a content management system is used to manage media data. The content management system is composed of a three-tiered architecture which is a general enterprise application architecture. That is, the content management system is composed of a client layer, an application layer, and a database layer.

The client layer realizes registration of content data such as scenario XML, scene XML, and media data using a WEB browser. The application layer performs management processing such as registration and modification of content data in response to a request from a client. The database layer is
An XML database system, which is highly compatible with a highly versatile programming language, is used to manage data processed in the application layer in the XML format.

This content management system has three functions. Media data management, scenario XM
L generation and SMIL format project generation. In the management of media data, scenario data, scene data, metadata of streaming media, and streaming media can be added / changed / deleted as necessary. In the scenario generation, meta XML, scene XML,
A scenario XML suitable for the user is generated from the user information. In the project generation, a SMIL format project for streaming media is generated from the scenario XML. The content management system collectively manages these processes to synchronize the data in the present system and smoothly operate the system.

A user preference analysis method will be described. In this system, different streaming media are distributed for each user for the purpose of multi-story streaming media. The following information is used to realize this distribution. There are three types of information, and profile information such as age, sex, occupation, favorite color, etc. acquired at the first login of the user, bandwidth at the time of login of the user, user's network environment such as type of terminal used, user's own situation, For example, "normal", "urgent", "entertainment",
Status information for acquiring “survey” and the like in a questionnaire format and history preference information for acquiring potential preference information of the user by analyzing the behavior history of the user. Data mining is to analyze a plurality of data groups using a technique such as a decision tree, a neural network, and a correlation, and normalize the mutual relation of data and the relation of data in a certain phenomenon. As a result, the streaming media suitable for the user can be automatically delivered.

A method of recommending a project to a user will be described. In this system, a user recommends multiple projects when browsing streaming media. The user selects a plurality of desired projects from the recommended projects.
The log data of the selected project is always saved and managed. The contents of the log data include the genre of the project selected by the user, the length of the project, the digestion rate of the story, and the contents used in the project. This log data is analyzed by data mining.

A market basket analysis is used as the analysis method. The market basket analysis statistically analyzes what projects are simultaneously selected from the user's past history. The ratio of projects that appeared at the same time is calculated from the analysis results, and when that ratio exceeds a certain threshold, that project has a new rule that if one project is recommended in the future, the other project is also recommended. Becomes higher. By doing so, it is possible to automatically recommend a project that reflects the user's taste.

As the above process is repeated, the bias of project recommendation is inevitably generated. User management
Centrally managed on the server side. Therefore, the users having similar tastes are grouped and the taste rules are shared.

Social mining is used for sharing preference rules. By using social mining, it is possible to derive a rule that, when the users have similar tastes, the rule established by one user also applies to the other user. This is to complement rules that do not exist in each other when the patterns of rules are similar among the users having similar tastes. As a result, the tendency of the preference among the categorized users is analyzed and interpolated to reduce the bias of recommendation and suppress unnecessary information.

In the scenario branch, different selection items are set for each user. This depends on the user's profile information and status information. The profile information limits the information, and the status information allows the user's current information to be grasped, so it is possible to acquire time information from scene data and media data and generate shortest to longest scenarios. .

An outline of story development of a scenario will be described based on the flowchart shown in FIG. This story is
A fixed prologue chapter that is common to all users, a first main chapter that is divided into three types of routes depending on the age of the user, two types of routes that are divided according to gender, and one route that is further divided depends on how busy the user is. It is composed of four types of chapters, a second main chapter divided into two types of routes, and a fixed epilogue chapter common to all users.

In the chapter of the prologue, only the prologue scene that is fixed in common to all users is included (S10). Next, in the first main chapter, one of the three types of scenes is selected according to the age condition (S12). If the user's age is less than 20 years old (a in S12), scene 1A is selected (S1).
14), 20 years or older and under 70 years old (b of S12),
The scene 1B is selected (S16), and when the person is 70 years old or older (c in S12), the scene 1C is selected (S18).

Next, in the second main chapter, first, a scene is selected according to the sex of the user (S20). When the user is a male (M in S20), the scene 2A is selected (S22), and when the user is a female (W in S20), the selected scene is further subdivided depending on the busyness of the user in the sub chapter. (S24). If the user is a person in a hurry (S24B), scene 2B
Is selected (S26), and if the user is a "non-hurried person" (F in S24), scene 2C and scene 2D are selected in that order (S28, S30). Finally, the chapter moves to the epilogue, and the epilogue common to all users is selected (S32).

In order to realize the story expansion of this scenario, the descriptive data are scenario XML and chapter XM.
It is composed of four types of files: L, scene XML, and content file.

The scenario XML is the source of the scenario, and includes a layout section that describes the definition of the viewpoint.
It consists of a flow part that describes the story flow. The flow is generated by the link information to the chapter XML described later. 2 to 5 show examples of chapter XML corresponding to the four types of chapters shown in FIG.

FIG. 2 shows the contents of the scenario XML. <Layout> and </ layout> on lines 4-6
The part sandwiched by the> tags is the layout part, and defines the size of the frame when the streaming media is displayed to the user. The 7th to 12th lines correspond to the flow part, and the link information to each chapter is described.
For example, <chapter src = "" alt = "main chapter-part 1"/> on the 9th line is link information to the first main chapter,
Specifically, the link information to the chapter XML of the first main chapter shown in FIG. 4 described later is described.

The chapter XML is composed of a description of the case of the story according to the preference parameter and link information to the scene XML in each case. Also, link information to another chapter XML can be described, and cases can be combined. FIG. 3 shows the chapter XML of the prologue chapter, and describes the link information to the scene “Introductory scene” which is the scene 1A in FIG. Figure 4 shows the chapter XML of the first main chapter.
Is shown.

FIG. 5 shows the chapter XML of the second main chapter. FIG. 5A relates to the selection of the scene according to the sex shown in FIG. The sixth line shows
The link information of the scene “scene for men” corresponding to the scene 2A when the user shown in FIG. In the ninth line, a link to the chapter XML regarding the subchapter when the user is a woman is described, and the subchapter XML is shown in FIG. 5B. In the sixth line of FIG. 5 (b), the link information to the scene "scene for people in a hurry" corresponding to the scene 2A is shown, and in the ninth and tenth lines, "hurry" corresponding to the scenes 2C and 2D. "Scene 1 for non-human" and "Scene 2 for non-hurried" are described.

The scene XML is composed of a clip information section which describes link information to a content file used in one scene, and a layout section which describes its layout information. FIG. 6 is a diagram showing a scene XML of the prologue. The third to sixth lines in FIG. 6 are the layout section, and 7
The 11th to 11th lines correspond to the clip information section.

The content file refers to a media file itself such as voice, image, text data.

FIG. 7 shows a compact chapter XML generated by specifying a scene suitable for the user. Here, as attributes of the user, it is assumed that the age is “under 20”, the gender is “female”, and the busy state is “not in a hurry”. At this time, in the flowchart shown in FIG. 1, the scene A1 is changed to the second scene in the first main chapter.
Scene 2C and Scene 2 in the main chapter
D is identified. In the chapter XML shown in FIG. 7, only necessary link information of FIGS. 3 to 5 is described.

Further, the content of the media file is registered by the management tool, and its ID (Identification Numbe
The scene XML in which r) has been written is registered, and then chapter XML and scenario XML are registered in order to make them usable by the SMIL generation module. Of course, it is also possible to use the registered chapter XML, scene XML, and content file used in another story. By calling the SMIL generation module by specifying the ID of the registered scenario XML and the preference parameter, the cases are removed, the individual XML files are combined, and one SMI is created.
An L file will be generated.

The configuration of the above system for realizing streaming distribution will be described. FIG. 8 shows a network system 10 including the content distribution system 9 according to this embodiment. In the network system 10, the user terminal 12 and the content distribution system 9 are connected via the Internet 18. The content distribution system 9 further includes a Web server 14, a scenario holding unit 20, and a streaming server 16. The streaming server 16 includes a video streaming holding unit 22 and an audio streaming holding unit 24.

The streaming server 16 delivers the streaming media to the user terminal 12. The Web server 14 supplements information in cooperation with user navigation and streaming data. User terminal 1
2 includes a web browser. In addition, the user terminal 12
A user who registers content by using is called a client.

FIG. 9 is a configuration diagram of the Web server 14 and the scenario holding unit 20. The web server 14 includes a communication unit 30, an application unit 50, a user management unit 66,
The streaming data encoding unit 68, the database interface 38, and the content holding unit 64 are included. The communication unit 30 communicates with the user terminal 12 and the web server 14 via the Internet 18. The application unit 50 includes a registration unit 32, a list display unit 34, and S.
The MIL output unit 60 and the SMIL generation unit 62 are included.

The user management unit 66 includes an authentication unit 70, a preference analysis unit 72, a session management unit 74, and a user information holding unit 76. The session management unit 74 allows the user to select W
After accessing the eb server 14, a series of actions of the user is monitored to analyze the action in the session and accumulate the action history. In addition, by including project information and its progress in the session information, streaming delivery with multi-story characteristics is realized in sessionless HTTP communication.

The taste analysis unit 72 analyzes the user's taste and situation based on the active information obtained from the session and the user's private information such as hobbies and characteristics acquired by the questionnaire at the first access. Generate preference information for personalization in project story generation.

The user information holding unit 76 includes a preference analysis unit 72.
The user's preference information and private information analyzed in (4) are held. The database interface 38 is an interface for connecting to the scenario holding unit 20, and is defined by an API (Application Programming Interface).

The registration unit 32 performs management processing such as registration and modification of content data in response to a request from the client. The list display unit 34 displays a list of registered contents to the client who registered the contents. The scenario content management unit 36 manages scenarios and contents. The SMIL generation unit 62 creates a SMIL format file to be delivered to the user, and
The L output unit 60 delivers the created SMIL format file.

Streaming data encoding unit 68
Encodes the moving image content and the audio content registered by the client and records them in the video streaming holding unit 22 and the audio streaming holding unit 24, respectively.

The scenario holding unit 20 includes a list HTML management unit 80, a list XSL holding unit 92, and a list XML holding unit 94. The list HTML management unit 80 is the registration unit 3
XML scenario, scene XML registered in 2
And manages a list of content and presents the list to the user upon request of the client. The list of contents registered by the client is held in the list XSL holding unit 92 and the list XML holding unit 94.

The scenario holding unit 20 further includes the scenario S.
MIL generation unit 82, scenario XML holding unit 84, scene XML holding unit 86, and SMIL generation XSL holding unit 8
Including 8. The scenario SMIL generation unit 82 uses the scenario X
Based on the ML holding unit 84 and the SMIL generation XSL holding unit 88, SMIL to be delivered to the user is generated.

The present invention has been described above based on the embodiments. This embodiment is merely an example, and it will be understood by those skilled in the art that various modifications can be made to the combinations of the respective constituent elements and the respective processing processes, and such modifications are also within the scope of the present invention. .

[0062]

According to the present invention, it is possible to expect the delivery of streaming media suitable for users.

[Brief description of drawings]

FIG. 1 is a flowchart showing an outline of scenario development of a scenario.

FIG. 2 is a diagram showing a scenario XML.

[Figure 3] Chapter XML of the prologue chapter
FIG.

FIG. 4 is a diagram showing a chapter XML of a first main chapter.

FIG. 5 is a diagram showing a chapter XML of a second main chapter.

FIG. 6 is a diagram showing a scene XML of a prologue.

FIG. 7 is a diagram showing chapter XML in which a scene suitable for a user is selected.

FIG. 8 is a diagram showing a network system including a content distribution system according to the present embodiment.

FIG. 9 is a configuration diagram of a Web server and a scenario holding unit.

[Explanation of symbols]

9 content distribution system, 12 user terminals,
14 Web servers, 16 streaming servers,
18 Internet, 20 Scenario holding unit, 2
2 video streaming holding unit, 32 registration unit, 3
6 scenario content management unit, 60 SMIL output unit, 62 SMIL generation unit, 64 content holding unit, 72 preference analysis unit, 74 session management unit,
76 user information storage unit, 82 scenario SMIL generation unit, 84 scenario XML storage unit, 86 scene XML storage unit, 88 SMIL generation XSL storage unit.

   ─────────────────────────────────────────────────── ─── Continued front page    F-term (reference) 5B075 ND12 PR03 PR08 UU40                 5C064 BA01 BA07 BB07 BC07 BC18                       BC23 BC25 BC27 BD02 BD07

Claims (5)

[Claims] 1. A scenario holding unit which has versatility by incorporating moving image contents that are likely to be used by various users, and holds a template described in the markup language format by the user as a general-purpose scenario. A user attribute acquisition unit that acquires a user attribute, a content specification unit that specifies a plurality of moving image contents suitable for a user to be delivered from the general-purpose scenario based on the acquired user attribute, A scenario editing unit that acquires the specified moving image content and rewrites the general-purpose scenario so that it can be reproduced in a predetermined order, and generates an individual scenario for each user; and a scenario distribution unit that distributes the generated individual scenario to the user. A content distribution system comprising: 2. The scenario editing unit leaves a moving image content suitable for a user among the moving image contents described in the general-purpose scenario, and physically deletes other moving image contents, The content distribution system according to claim 1, wherein the individual scenario is generated by compacting a general-purpose scenario. 3. A session management unit that manages a history of moving image content accessed by a user based on the delivered individual scenario, wherein the content specifying unit is a moving image content managed by the session management unit. The content distribution system according to claim 1 or 2, wherein a moving image content suitable for a user is estimated by referring to an attribute relating to the attribute and the history. 4. The markup language is XML, and in the general-purpose scenario, a scenario is described in a branch structure using if conditional statements, and a condition for branching the scenario is set as an attribute of the user. Claims 1 to 3
Content distribution system according to any one of. 5. A step of holding a general-purpose scenario in which link information to a wide range of moving image contents is described in a markup language format, a step of acquiring a user attribute, and the acquisition based on the general-purpose scenario. Identifying a moving image content suitable for the user according to the attribute of the user, and leaving the moving image content identified as suitable,
And a step of physically deleting the description of the link information to the other contents to generate a secondary scenario customized for each user.