JP5391431B2 - Content processing apparatus, content processing program, and content processing method - Google Patents

Description

  The present invention relates to a content processing apparatus, a content processing program, and a content processing method.

  The network Web environment is evolving day by day, and the amount of information contained therein is increasing. On the other hand, the form of information is not only text, but multimedia such as still images, moving images, and voices is further advanced. In order for Web viewers to make effective use of such a large amount of information on the network, it is most desirable that the information according to the purpose of use can be efficiently searched. Therefore, in recent years, information can be efficiently searched by using various search engines.

  By the way, as shown in FIG. 21, such a search engine basically only presents a list of a part of search results (several lines of text including keywords or image thumbnails near the keywords) To obtain more information, open each linked Web page one by one and check the contents. If you cannot obtain the desired information, return to the search result list screen again. Become.

  On the other hand, when searching for information using a keyword in a field that is not familiar, it is not always easy to come up with an appropriate keyword, and search results often contain information that the viewer does not expect (unnecessary). . In such a case, there is a method of narrowing down the search target while adding keywords by trial and error, but it takes time and effort to repeat the search many times. Therefore, when the search results are categorized and displayed, the viewer can easily grasp the entire contents of the target field, so that it is easy to find out only the information of the desired category. Such categorized display of search results is described in Non-Patent Document 1, for example.

  However, it is very cumbersome to repeat the process of opening each link destination from the search result list, and if the search result and the link destination are reciprocated many times in this way, the information browsing context is lost and the search is performed. There is a problem that efficiency decreases. On the other hand, when a lot of information is presented for each search result Web page, the overall listability is deteriorated, and the search efficiency is also lowered. In recent years, methods for reducing the display of linked Web pages in the list screen have also appeared. However, these can grasp the atmosphere of the entire link destination, but the details of text and images are detailed. There is a problem of not knowing.

  In addition, according to the technique for categorizing and displaying the search results described in Non-Patent Document 1, the list screen after selecting a category is an extracted text similar to the normal keyword search result screen of the search engine To get more information, you must open the linked page.

  In addition, the following problems are known for displaying search results. That is, when a result of searching with a certain keyword is classified into a plurality of clusters, it may be difficult to find a desired item because they are ranked together. For example, as a result of searching with the keyword “Japan”, items related to “Honshu”, “Tokyo”, “Kyushu”, “Osaka”, “Kyoto” are obtained. Here, items related to two different clusters, “island” and “city” are mixed (“Honshu” and “Kyushu” are related to “Island”, and “City” is related to “Island”. "Tokyo", "Osaka", "Kyoto"). “Tokyo”, “Osaka” and “Kyoto” are not necessary for users who want to know information about “islands”, and “Honshu” and “Kyushu” are unnecessary for users who want to know information about “city”. . As described above, as a result of a mixture of items related to different clusters, unnecessary information may be included in the search results, and it may be difficult to find items belonging to a desired cluster.

  The present invention has been made in view of the above, and is capable of efficiently performing an information search while maintaining the browsing context, eliminating the need to search many times, and in particular a plurality of different clusters. It is an object of the present invention to provide a content processing apparatus, a content processing program, and a content processing method that can make it easier to find an item that belongs to a desired cluster when items belonging to are mixed.

In order to solve the above-described problems and achieve the object, the content processing apparatus of the present invention changes the level of detail within the same screen for the content belonging to one cluster among the contents clustered into a plurality of clusters. And at least one of the display media transitions is continuously realized in units of the one cluster , and the content belonging to another cluster is linked to at least one of the detail level change of the content belonging to the one cluster and the display media transition. A zooming description generating unit that describes a zooming operation / behavior that continuously realizes at least one of a change in detail level and display media transition in the same screen in units of other clusters , and predetermined content from a user An operation accepting unit that accepts zoom-in / out operations on Depending on the zoom in / out operation accepted by the Department, in accordance with the generated described by the zooming description generation unit, the clustered content, at least one of the details of change and display media transition in one cluster units belonging operation content was carried out, and a clustering display unit Ru is displayed by performing at least one of the details of change and display media transitions in other cluster unit in conjunction with this, the clustering display unit, the user selects the content When zooming in, content belonging to a cluster different from the content selected by the user is hidden .

Further, the content processing program of the present invention allows a computer to perform at least one of a detail level change and a display media transition on the same screen among the contents clustered into a plurality of clusters. Details of the content that is continuously realized in one cluster unit and that belongs to another cluster in the same screen in conjunction with at least one of the change in the level of detail of the content that belongs to the one cluster and display media transition Zooming description generation means for describing a zooming operation / behavior that continuously realizes at least one of the degree change and display media transition in units of the other clusters, and an operation reception for receiving a zoom-in / out operation for a predetermined content from the user and it means, by the operation receiving means Depending on only with the zoom in / out operation, in accordance with the generated described by the zooming description generating unit, the clustered content, at least one of the details of change and display media transition in one cluster units belonging operation content performs a clustering display means Ru is displayed by performing at least one of the details of change and display media transitions in other cluster unit in conjunction with this, to run, the clustering display means, the user selects the content When zooming in, content belonging to a cluster different from the content selected by the user is hidden .

The content processing method of the present invention is a content processing method in the content processing apparatus for generating and outputting a search result of the content, among the clustered content into a plurality of clusters, for the content belonging to one cluster , At least one of detail level change and display media transition within the same screen is continuously realized in units of the one cluster, and for the content belonging to another cluster, the detail level change of the content belonging to the one cluster Zooming description generation that describes a zooming operation / behavior that continuously realizes at least one of detail level change and display media transition in the same cluster unit in conjunction with at least one of display media transition a step,'s for predetermined content from the user An operation accepting step of accepting a Muin / out operation, depending on the zoom-in / out operation accepted by the operation reception process, in accordance with the description generated by the zooming description generation process, one of the clustered content, the operation content belongs at least one of the done, including a clustering display step of Ru is displayed by performing at least one of the details of change and display media transitions in other cluster unit in conjunction with this, the in cluster units of the detail of change, and display media transition In the clustering display step, when the user selects content and zooms in, content belonging to a cluster different from the content selected by the user is hidden .

  According to the present invention, while maintaining the browsing context, it is possible to efficiently perform an information search without the need to search many times, and particularly when items belonging to a plurality of different clusters are mixed, There is an effect that an item belonging to a desired cluster can be easily found.

FIG. 1 is a front view showing an example of multimedia content displayed on a display for explaining an outline of screen transition in content browsing in a Web environment as a premise of the present invention. FIG. 2 is a front view showing the result of zooming in on the multimedia content object A shown in FIG. FIG. 3 is a front view showing a result of a zoom-in operation performed on the multimedia content object A shown in FIG. FIG. 4 is a front view showing a result of a zoom-in operation performed on the multimedia content object B shown in FIG. FIG. 5 is an explanatory diagram showing an outline of screen transition in browsing content in a Web environment by a simple specific example. FIG. 6 is a schematic diagram showing a system construction example including the content processing apparatus according to the embodiment of the present invention. FIG. 7 is a block diagram illustrating a configuration of a server that is a content processing apparatus. FIG. 8 is a block diagram showing a schematic configuration of a system including a server which is a content processing apparatus. FIG. 9 is a block diagram illustrating a configuration of the ranking / clustering unit. FIG. 10 is a block diagram illustrating a configuration of the indexing unit. FIG. 11 is a block diagram illustrating a configuration of the ranking unit. FIG. 12 is a block diagram illustrating a configuration of the clustering unit. FIG. 13 is an explanatory diagram showing calculation formulas related to index generation and similarity calculation. FIG. 14 is an explanatory diagram showing calculation formulas related to index generation and similarity calculation. FIG. 15 is an explanatory diagram showing calculation formulas related to index generation and similarity calculation. FIG. 16 is an explanatory diagram showing calculation formulas related to index generation and similarity calculation. FIG. 17 is an explanatory diagram showing calculation formulas related to index generation and similarity calculation. FIG. 18 is a block diagram illustrating a configuration of the display unit. FIG. 19 is an explanatory diagram showing clustered search results. FIG. 20 is an explanatory diagram illustrating an example of a zooming operation on a search result list of clustered web pages. FIG. 21 is an explanatory diagram showing an operation example for a search result list of a conventional web page.

  Exemplary embodiments of a content processing apparatus, a content processing program, and a content processing method according to the present invention are explained in detail below with reference to the accompanying drawings.

  An embodiment of the present invention will be described with reference to FIGS.

[1. Overview of screen transitions when browsing content in the Web environment]
First, an outline of screen transition in content browsing in a Web environment that is a premise of the present invention will be described.

  In the present embodiment, for the purpose of enhancing the content browsing effect in the Web environment, “zooming” that can realize not only the change in the detail level but also the transition of the display media using the zooming for the content in the Web environment.・ Cross media ”is applied.

  “Zooming cross media” is to position the zoom operation as a third operation for scrolling and anchor clicking. While scrolling moves the point of interest continuously within the same screen and anchor clicks fly to different screens, zooming / cross media continuously realizes changes in detail and display media transitions within the same screen. Display media transition refers to media objects that are related to each other, and according to the zoom operation, the display is continuously changed from one to the other, or the other is continuously displayed and the other is continuously enlarged. It is to make additional display while vice versa (or vice versa). For example, when subtitle text associated with a moving image is displayed, it is gradually changed from subtitle text display to subtitle text with a zooming operation.

Such zooming operation / behavior is modeled on the basis of three basic concepts of zoom target, zoom rate, and zoom propagation.
1. Zoom Target “Zoom Target” is a unit that is a target of zoom operation, and is composed of “media objects” of text, still images, moving images, and audio that constitute the content. A viewer designates a desired zoom target and performs a zoom operation.
2. Zoom rate “Zoom rate” is a value uniquely determined for a zoom target, and is a continuous value between 0.0 and 1.0 that represents the degree of zoom. Zooming in the direction of increasing the zoom rate is called zooming in and zooming out in the direction of decreasing the zoom rate. For each media object, the zoom rate is associated with specific attributes (font size, image size, moving image or audio playback section, etc.).
3. Zoom Propagation “Zoom Propagation” means that when a zoom target is zoomed, other zoom targets are also zoomed together. In zoom propagation, the zoom ratio of the propagation destination is calculated from the zoom ratio of the propagation source zoom object based on a predetermined “zoom ratio relationship”, and the display state of the propagation destination zoom object is controlled. The zoom rate relationship is the ratio of change in the zoom rate between the propagation source and the propagation destination.

  Here, an example of zoom propagation when the zoom ratio relationship is a positive value will be described. 1 to 3 are explanatory diagrams showing examples of screen transition in content browsing in a Web environment. FIG. 1 is a front view showing an example of content in a Web environment displayed on a display. The example shown in FIG. 1 is an example in which “business trip report” is applied as content in the Web environment. The content in the Web environment shown in FIG. 1 is mainly composed of text as an object.

  FIG. 2 shows the result of zooming in on the content object A in the Web environment shown in FIG. As shown in FIG. 2, when the zoom-in operation is performed on the object A of the content in the Web environment, the object B below the object A is newly displayed, and the new object C is the lower part of the object B. Is displayed. Here, the new object C is a still image of a moving image in which the discussion scene at the time when the discussion related to the object A was made.

  Further, FIG. 3 shows the result of performing a zoom-in operation on the content object A in the Web environment shown in FIG. As shown in FIG. 3, when the zoom-in operation is performed on the object A of the content in the Web environment, the font sizes of the objects A and B are enlarged. In addition, as shown in FIG. 3, when the zoom-in operation is performed on the object A of the content in the Web environment, the detail level of the still image that is the object C is increased (the image size is increased). Furthermore, when the object C is enlarged to a certain size, reproduction of a moving image (and sound) is started.

  The screen transition in content browsing in such a Web environment is described in the following display description (structure description and zooming description). The structure description and the zooming description are described in an XML (Extensible Markup Language) based zooming description language (ZDL).

  Here, the zooming description will be described below with a simple example.

1. Designation of media object Each of text, still image, moving image, and audio is represented by tags of <text />, <img />, <video />, and <audio />. This tag specifies the specific text content or source file.
(Example 1)
<text id = text1> This is text. </ text>
<img id = img1 src = "img1.jpg"/>
<video id = video1 src = "video1.mpg"/>
<audio id = audio1 src = "audio1.wav"/>

2. Zoom rate range
In zoomFactor attribute in the <div /> tag, specify the maximum and minimum values of the zoom factor and represent the range of the zoom factor as a continuous parameter.
(Example 2)
<div zoomFactor = "0.0 0.5"/>

3. Display of media object Each text, still image, moving image, and audio is represented by <dispText>, <dispImg>, <dispVideo>, and <dispAudio> tags. It is always included in the <div /> tag, and represents the display state of each media object at the zoom ratio in the range specified by the <div /> tag.
(Example 3-1)
<div zoomFactor = "0.0 0.5">
<dispText id = "disptext1" textId = "text1" fontSize = "10pt 14pt"/>
</ div>
The above means that the text specified by text1 in the <text /> tag is displayed as 10pt when the zoom factor is 0.0 and 14pt when the zoom factor is 0.5, and the font size is as smooth as possible during that time. Is to change.
(Example 3-2)
<div zoomFactor = "0.0 0.5">
<dispImg id = "dispimg1" imgId = "img1" imgSize = "10% 50%"/>
</ div>
The above means that the still image specified by img1 in the <img /> tag is displayed at 10% of the maximum size when the zoom factor is 0.0, and 50% when the zoom factor is 0.5. Is to change the size as smoothly as possible.
(Example 3-3)
<div zoomFactor = "0.0 0.5">
<dispVideo id = "dispvideo1" videoId = "video1" start = "0" end = "300" videoSize = "10% 50%"/>
</ div>
The above means that the 0th to 300th frame of the video specified by video1 in the <video /> tag is 10% of the maximum size when the zoom ratio is 0.0, and when the zoom ratio is 0.5 Each is displayed at 50%, and the size is changed as smoothly as possible.
(Example 3-4)
<div zoomFactor = "0.0 0.5">
<dispAudio id = "dispaudio1" audioId = "audio1" start = "0s" end = "30s"/>
</ div>
The above meaning is that the first 30 seconds of the audio specified by audio1 in the <audio /> tag is reproduced during the zoom ratio of 0.0 to 0.5.

4). Zoom target Describe the zoom target with the <zoomObject /> tag. The initial zoom rate, the zoom target of the zoom propagation destination, and the zoom rate relationship are also described with attributes in the tag. Be sure to include either <div /> or other <zoomObject />.

The structure description and zooming description for the screen transition examples shown in FIGS. 1 to 3 are shown below. In the following description, the structure description of the content and the zooming description regarding the zooming operation / behavior are described separately (onlyHierarchy = "on" indicates that the description is only for the structure).

(Content structure description)
<zoomObject id = "zoom1" onlyHierarchy = "on">
<text id = "text1"> 1. Improvements of new teaching materials </ text>
<text id = "text2"> Ask participants to give feedback on the content of the material. </ text>
<zoomObject id = "zoom2" onlyHierarchy = "on">
<text id = "text3"> How can I get new things out one after another,
How can I get it used over and over again (problem) </ text>
<zoomObject id = "zoom3" onlyHierarchy = "on">
<text id = "text4"> Development needs to be made in a slightly shorter span. </ text>
<text id = "text5"> I will update every 3 months. </ text>
.......
</ zoomObject>
<zoomObject id = "zoom4" onlyHierarchy = "on">
<video id = "video1" src = "meeting 11.mpg"/>
</ zoomObject>
</ zoomObject>
<zoomObject id = "zoom5" onlyHierarchy = "on">
<text id = "text10"> First time with interpreter memo in step 6 .... </ text>
.......
</ zoomObject>
.......
</ zoomObject>

(Zooming description for zooming operation / behavior)
<zoomObject id = "zoom1" initZoomFactor = "0.0" relatedTo = "children: 0.8">
<div zoomFactor = "0.0 1.0">
<dispText id = "disptext1" textId = "text1" textSize = "12pt 14pt"/>
<dispText id = "disptext2" textId = "text2" textSize = "12pt 14pt"/>
</ div>
</ zoomObject>

<zoomObject id = "zoom2" initZoomFactor = "0.0" relatedTo = "children: 0.8">
<div zoomFactor = "0.0 1.0">
<dispText id = "disptext5" textId = "text3" textSize = "11pt"/>
</ div>
</ zoomObject>

<zoomObject id = "zoom3" initZoomFactor = "0.0" relatedTo = "brother: -0.5">
<div zoomFactor = "0.1 1.0">
<dispText id = "disptext6" textId = "text4" textSize = "8pt 10pt"/>
<dispText id = "disptext7" textId = "text5" textSize = "8pt 10pt"/>
.......
</ div>
</ zoomObject>

<zoomObject id = "zoom4" initZoomFactor = "0.0" relatedTo = "brother: -0.5">
<div zoomFactor = "0.1 0.4">
<dispVideo id = "dispvideo1" videoId = "video1" start = "0" end = "0" videoSize = "20% 50%"/>
</ div>
<div zoomFactor = "0.4 1.0">
<dispVideo id = "dispvideo2" videoId = "video1" start = "0" end = "900" videoSize = "50% 100%"/>
</ div>
</ zoomObject>

  According to the structure description of this content, zoom1 is at the top, zoom2, zoom5, etc. are below it, and zoom3, zoom4 are further below zoom2. Zoom3 is composed of a plurality of texts (text4, text5, etc.), and zoom4 is composed of one moving image (video1).

  Also, according to the zooming description regarding zooming operation / behavior, zoom1, zoom2, zoom3, and zoom4 all have an initial zoom ratio of 0.0, but the first two display text (text1, text2, text3) from the beginning. On the other hand, nothing is displayed for the latter two (this corresponds to the initial state of FIG. 1).

  Here, when zoom2 is zoomed in, it propagates to the zoom target of the child hierarchy (relatedTo = "children: 0.8" indicates that the zoom propagates to all the child hierarchies with a zoom ratio of 0.8). When zoom2 zoom ratio increases, zoom3 and zoom4 zoom ratios increase by 0.8 (when zoom2 zoom ratio changes +0.5, zoom3 and zoom4 change +0.4, respectively) and reach 0.1 Then, text4, text5, etc. and video1 are displayed (see FIG. 2). This is an example of media expansion because moving image information is added to text information. However, for video1, only frame 0 is still displayed here.

  If zoom2 is zoomed in further from this state, the zoom rate of zoom3 and zoom4 will also increase, and the font size of text4, text5 and other will increase. At the same time, the display size of video1 is enlarged, and when zoom4 of zoom4 reaches 0.4, frames 0 to 900 of video1 are reproduced (FIG. 3, example of change in detail).

  If zoom2 is zoomed out in the state shown in FIG. 3, zoom3 and zoom4 are zoomed out together to the state shown in FIG. 2 contrary to the case of zooming in. When each zoom ratio becomes 0.1 or less, text4, text5, etc. And video1 disappear from the screen and return to the state of FIG. Alternatively, if the operation is forcibly returned to the initial state, the state shown in FIG. 1 is restored from any display state.

  Next, an example of zoom propagation when the zoom ratio relationship is a negative value will be described.

  FIG. 4 is an explanatory diagram showing an example of screen transition in content browsing in the Web environment. FIG. 4 shows the result of zooming in on the content object B in the Web environment shown in FIG. As shown in FIG. 4, when the zoom-in operation is performed on the object B of the content in the Web environment, the font size of the object B is enlarged, but the display size of the object C is reduced.

In the zooming description regarding the zooming operation / behavior described above, both zoom3 and zoom4
relatedTo = "brother: -0.5"
There is a description. This indicates that zoom propagation occurs between zoom targets having a common parent. A negative zoom ratio relationship indicates that zooming in on one zooms out the other. Therefore, if only zoom3 is selected from the state of FIG. 3 and zoomed in, zoom4 is zoomed out (text4, text5, etc. are further enlarged, and video1 is conversely reduced (see FIG. 4)).

  The above is the outline of the screen transition in content browsing in the Web environment. FIG. 5 shows such a relationship in a simple specific example. In the example of FIG. 5, A, B, and C are zoom targets, and the parentheses indicate the zoom ratio in the initial state. An arrow represents a zoom propagation relationship, and a value attached to the arrow indicates a zoom rate relationship. Since the zoom ratio relationship of AB is 0.5, when zooming in from A to B and zooming in A and changing the zoom ratio from 0.0 to 0.8, the zoom ratio of B changes from 0.0 to 0.4. B and C propagate in zoom with each other but the rate of change is negative (-1.0), so zooming in on one zooms out on the other. For example, in the initial state, only C (moving image) is displayed, and when B (text + sound) is zoomed in and enlarged, C (moving image) is gradually reduced. As in this example, the feature of this modeling is that transition of display media can be easily realized by introducing zoom propagation.

  According to the zooming description as described above, when one media object (text, still image, moving image, audio, etc.) is zoomed, other media objects (text, still image, moving image, audio, etc.) are linked with it. Since it is possible to describe zoom propagation between zooming objects to be zoomed, natural zooming from one zoom object to another zoom object can be performed.

[2. Description of content processing apparatus]
In this embodiment, a search result list of a Web page by a search engine (for example, a robot search engine) is applied as content in the Web environment, and the search result list of the Web page is output in a zooming description language (ZDL) format. The present invention is applied to a content processing apparatus that can

  Here, a web page is a collection of data displayed at once in a web browser, text data, layout information by HTML (or CSS (Cascading Style Sheet), XML, etc.), images embedded in a document, It consists of audio, video, etc.

  The reason why zooming / cross-media is applied to information retrieval in this way is that the big utility of zooming / cross-media is to increase the degree of freedom of content browsing. The change in the level of detail and the transition of display media can be controlled according to the user's preference. That is, on the search result list screen of the Web page, the detailed level of information and the display medium necessary for determining whether the link destination Web page is the desired information without opening the link destination Web page This is because it is considered to vary depending on the knowledge of the person, the situation at the time of browsing, the form of information to be searched, and the like. For example, a viewer who is familiar with the field to be searched can be determined by the title alone, but an uninformed viewer may want image data in addition to a more detailed text description. There are cases where there is room to open each link destination over time, and conversely, there is a case where it is desired to briefly list the overall trends. If the link destination is text-based, it is appropriate that the excerpt (or summary) is displayed on the result list screen, but if the link destination contains a lot of images, it is more convenient to display a few main images. There is also. It is not realistic to generate a plurality of fixed search result list screens assuming such various situations in advance. Here, the features of zooming / cross-media that can display a plurality of media in a mixed / transition manner and can flexibly control the level of detail and the display media according to the situation and preference of the viewer will come into play.

[2-1. Example of system construction including content processing apparatus]
FIG. 6 is a schematic diagram illustrating a system configuration example including a content processing apparatus. As shown in FIG. 6, the system includes a server computer (hereinafter referred to as a server) 1 as a content processing apparatus 1 and a plurality of client computers (hereinafter referred to as client terminals) C as terminal devices via a network N as the Internet. A system connected to a stand is assumed. Needless to say, the present invention is not limited to the client-server system involving such network communication.

  The server 1 that is a content processing apparatus generally performs a Web page search based on a keyword for a general Web page on the network N that is the Internet, and stores the Web page in a zooming description language (ZDL) format. A list of search results is output.

  The client terminal C is a general personal computer or the like, and can transmit and receive data to and from the server 1 by using a Web browser. This client terminal C accepts a Web page keyword search request from a viewer and outputs it to the server 1, and at the same time, a search result list of Web pages for the viewer output in the zooming description language (ZDL) format on the server 1. Is interpreted and displayed. That is, the client terminal C is an input / output device for the server 1 that is a content processing apparatus.

  FIG. 7 is a block diagram showing the configuration of the server 1. A server 1 which is a content processing apparatus according to the present embodiment is mainly composed of a personal computer, and a CPU (Central Processing Unit) 2 for performing information processing, a ROM (Basic Input Output System) for storing a BIOS (Basic Input Output System) and the like. Read Only Memory (3), RAM (Random Access Memory) 4 that temporarily stores information during processing, HDD (Hard Disk Drive) 5 that is a storage unit that stores application programs, processing results, and the like. For connecting to a drive 6a of a removable medium 6 which is a storage medium for storing or distributing outside or obtaining information (application program etc.) from the outside, and a network N for communicating with other external electronic devices A network interface 8; a display 9 that is a display unit for displaying a process progress, a process result, and the like; It consists of an input unit such as a keyboard 10 and a mouse 11 for the author to input commands, information, etc. to the server 1 which is a content processing device. The bus controller 12 arbitrates the data communication between them and operates. Yes.

  The removable medium 6 includes a magnetic storage medium such as a flexible disk, a hard disk, and a magnetic tape, a magneto-optical storage medium such as an MO, a CD, a CD-ROM, a CD-R, a CD-RW, Various storage media such as optical storage media such as DVD-ROM, DVD-R, DVD-RAM, DVD-RW, DVD + RW, and semiconductor memory can be applied.

  In data transmission, the network interface 8 outputs a signal to the network N by sending data to the network interface 8 of the server 1 which is a content processing apparatus. Further, regarding the signal received by the network interface 8, it is determined whether or not it is necessary in the network interface 8, and processing such as fetching if necessary data and discarding if unnecessary is performed. That is, all data transmission / reception is performed via the network interface 8.

  In general, when the user turns on the power of the server 1 which is a content processing apparatus, the CPU 2 activates a program called a loader included in the BIOS in the ROM 3 and the hardware of the server 1 called the operating system (OS) from the HDD 5. And a program for managing software are read into the RAM 4. When the OS is activated, it supports activation of application programs, reading of information, storage, and the like according to user operations. As typical OS, Windows (registered trademark of Microsoft Corporation), UNIX (registered trademark of X / Open, Inc.) and the like are known. Programs that run on these OSs are called application programs. The application program is not limited to one that runs on a predetermined OS, and may be one that causes the OS to execute some of the various processes described below, or constitutes predetermined application software, an OS, or the like. It may be included as part of a group of program files.

  In general, an application program installed in the HDD 5 of the server 1 that is a content processing apparatus is recorded in the removable medium 6, and the application program recorded in the removable medium 6 is installed in the HDD 5. For this reason, the removable medium 6 can also be a storage medium for storing application programs. Furthermore, the application program may be taken in from the outside via, for example, the network interface 8 and installed in the HDD 5.

  When an application program that runs on the OS is activated, the server 1 that is a content processing apparatus performs various arithmetic processes according to the application program, and centrally controls each unit.

[2-2. Processing in content processing apparatus]
Next, characteristic processing of the present embodiment that is executed according to a content processing program that is one of application programs among various arithmetic processing executed by the CPU 2 of the server 1 that is a content processing apparatus will be described below.

  FIG. 8 is a block diagram illustrating a schematic configuration of a system including the server 1. As shown in FIG. 8, the server 1 that is a content processing apparatus has a database (web page DB) 21 that holds a web page that is the first content in the web environment by the CPU 2 controlling each unit according to the content processing program. The ranking / clustering unit 22 that receives a user query output from the client terminal C, ranks the web pages held in the web page DB 21 based on the query, and clusters the ranked web pages. And the display unit 23 that continuously displays the ranking (search result) of the clustered web pages using the zooming cross media in the zooming description language (ZDL) format.

  That is, with such a configuration, on the client terminal C side, when a viewer starts a keyword search on a Web browser, a ZDL viewer that interprets and displays a zoom description language (ZDL) is activated at the same time. Then, the search request (query) input at the client terminal C is transmitted to the server 1. On the server 1 side, which is a content processing apparatus, when a search request (query) input at the client terminal C is received, a database search based on the keyword is executed. The server 1 generates a search result list in a zooming description language (ZDL) format. The generated search result list in the zooming description language (ZDL) format is transmitted from the server 1 which is a content processing apparatus to the client terminal C and displayed on the ZDL viewer. The viewer appropriately performs a zooming operation on the displayed search result list.

  First, the configuration of the ranking / clustering unit 22 will be described in detail. FIG. 9 is a block diagram showing the configuration of the ranking / clustering unit 22. The ranking / clustering unit 22 generates an index based on the keyword of the web page for the text of the web page and stores it in the database (index DB) 30, and the index generated by the indexing unit 31 A ranking unit 32 that ranks web pages in response to a user query using an index stored in the DB 30 and a clustering unit that clusters web pages using an index generated by the indexing unit 31 and stored in the index DB 30 33.

  As shown in FIG. 10, the indexing unit 31 uses an existing natural language processing technique (for example, morphological analysis) to extract a representative keyword from the text of a web page, and a keyword extracting unit 41. And an index generation unit 42 that generates an index of a web page using an existing text search technique (for example, a vector space model) based on the keyword extracted in (1). The keyword extraction unit 41 can use a manual method or a semi-automatic method in addition to an automatic method (natural language processing technology) when extracting a representative keyword. The index generation unit 42 uses the extracted representative keyword to convert the contents of the web page into the same format (for example, a document vector) so that the relative similarity between them can be calculated.

  As shown in FIG. 11, the ranking unit 32 includes a query input unit 51 that allows the user to select one or more keywords from the extracted representative keywords, and a keyword selected by the user (user's A query conversion unit 52 that converts (query) into the same format as the index, and a first similarity calculation unit that calculates the similarity of each web page to the user's query using the index stored in the index DB 30. A similarity calculation unit 53 and a ranking generation unit 54 for ranking web pages with respect to user queries based on the similarity calculated by the similarity calculation unit 53 are provided. The query input unit 51 allows the user to select keywords from a one-dimensional keyword list or a two-dimensional keyword map (or otherwise). The query conversion unit 52 uses the extracted representative keyword, converts the keyword (user query) selected by the user into a format (for example, a query vector) handled by the index stored in the index DB 30, and To calculate the similarity between a query and a web page. The similarity calculation unit 53 uses a query or web page expression format (for example, a query vector or a document vector) handled by the index, and uses the similarity (0 and A value between 100%) is calculated. Based on the calculated similarity, the ranking generation unit 54 ranks web pages in descending order of similarity with respect to the user's query.

  As shown in FIG. 12, the clustering unit 33 is a second similarity calculation unit that calculates the relative similarity between web pages using the index stored in the index DB 30 that is the output of the indexing unit 31. A similarity calculation unit 61 and a cluster generation unit 62 that performs clustering on the ranking (search result) of each web page using the similarity calculated by the similarity calculation unit 61 are provided. The similarity calculation unit 61 calculates the relative similarity of each web page pair using the web page expression format handled by the index. Based on the relative similarity calculated by the similarity calculation unit 61, the cluster generation unit 62 clusters existing web pages using the existing clustering technology (eg, “hierarchical clustering method”) (as a result, the relative similarity is calculated). Web pages with high relatives belong to the same cluster, and web pages with low relative similarity belong to different clusters).

  13 to 17 show calculation formulas relating to index generation and similarity calculation in the ranking and clustering unit 22. This calculation formula is a standard calculation formula used in the “vector space model” in the existing text search technology.

  Next, the display unit 23 will be described in detail. Here, FIG. 18 is a block diagram showing the configuration of the display unit 24. As shown in FIG. 18, the display unit 23 includes a ranking display unit 71 that displays a list of web pages as an initial search result, and the same cluster as the web page selected by the user when the user selects a web page and zooms in. And a clustering display unit 72 that enlarges and displays the web pages belonging to the selected web page and displays the web pages belonging to different clusters in a reduced manner.

  Here, “clustered search results (web page ranking)” in the ranking unit 32 and the clustering unit 33 will be described with reference to FIG.

The ranking unit 32 arranges the web pages in descending order of similarity to the user query ("Japan") (in this example, the web pages are expressed by one keyword: "Honshu", "Tokyo", "Kyushu", "Osaka", "Kyoto"). On the other hand, the clustering unit 33 collects web pages having similar “meanings” (high similarity) from the web pages arranged in order, and represents the “meanings” for the different “meanings”. Create a cluster. As a result, a cluster representing "island" containing pages "Honshu" and "Kyushu" and a cluster representing "city" containing pages "Tokyo", "Osaka", and "Kyoto" were created. . Based on the clustering result, the following description is generated (zooming description generation unit).

(Content structure description)
<zoomObject id = "zoom1" onlyHierarchy = "on">
<zoomObject id = "zoom2" onlyHierarchy = "on">
<text id = "title2"> Items related to "Honshu"</text>
…………
</ zoomObject>
<zoomObject id = "zoom3" onlyHierarchy = "on">
<text id = "title3"> Items related to “Tokyo” </ text>
…………
</ zoomObject>
<zoomObject id = "zoom4" onlyHierarchy = "on">
<text id = "title4"> Items related to "Kyushu"</text>
…………
</ zoomObject>
<zoomObject id = "zoom5" onlyHierarchy = "on">
<text id = "title5"> Items related to "Osaka"</text>
…………
</ zoomObject>
<zoomObject id = "zoom6" onlyHierarchy = "on">
<text id = "title6"> Items related to “Kyoto” </ text>
…………
</ zoomObject>
…………
</ zoomObject>

(Zooming operation / behavior description)
<zoomObject id = "zoom1" initZoomFactor = "0.0" relatedTo = "children: 1.0">
</ zoomObject>
<zoomObject id = "zoom2" initZoomFactor = "0.0" relatedTo = "brother: -1.0, zoom4: 1.0">
<div zoomFactor = "0.0 1.0">
<dispText id = "disptext2" textId = "title2" textSize = "8pt 14pt"/>
…………
</ div>
</ zoomObject>
<zoomObject id = "zoom3" initZoomFactor = "0.0" relatedTo = "brother: -1.0, zoom5: 1.0, z
oom6: 1.0 ">
<div zoomFactor = "0.0 1.0">
<dispText id = "disptext3" textId = "title3" textSize = "8pt 14pt"/>
…………
</ div>
</ zoomObject>
<zoomObject id = "zoom4" initZoomFactor = "0.0" relatedTo = "brother: -1.0, zoom2: 1.0">
<div zoomFactor = "0.0 1.0">
<dispText id = "disptext4" textId = "title4" textSize = "8pt 14pt"/>
…………
</ div>
</ zoomObject>
<zoomObject id = "zoom5" initZoomFactor = "0.0" relatedTo = "brother: -1.0, zoom3: 1.0, z
oom6: 1.0 ">
<div zoomFactor = "0.0 1.0">
<dispText id = "disptext5" textId = "title5" textSize = "8pt 14pt"/>
…………
</ div>
</ zoomObject>
<zoomObject id = "zoom6" initZoomFactor = "0.0" relatedTo = "brother: -1.0, zoom3: 1.0, z
oom5: 1.0 ">
<div zoomFactor = "0.0 1.0">
<dispText id = "disptext6" textId = "title6" textSize = "8pt 14pt"/>
…………
</ div>
</ zoomObject>
…………

According to the above description, when zoom2 (item related to “Honshu”) is zoomed in / out (operation receiving unit), zoom4 (item related to “Kyushu”) is also zoomed in / out in conjunction with it. On the other hand, zoom3 (items related to “Tokyo”), zoom5 (items related to “Osaka”), and zoom6 (items related to “Kyoto”) are zoomed out / in on the contrary. this is,
<zoomObject id = "zoom2" initZoomFactor = "0.0" relatedTo = "brother: -1.0, zoom4: 1.0">
"RelatedTo =" brother: -1.0, zoom4: 1.0 "" in the description.

  Similarly, when zoom4 is zoomed (operation reception unit), zoom2 is linked and the other parts perform the opposite behavior. On the other hand, when zoom3 is zoomed (operation reception unit), zoom5 and zoom6 are linked and the other moves in the opposite direction. The same applies to zoom5 and zoom6. In this way, even if items related to “island” and items related to “city” are mixed, they are clustered, plus zoom propagation is applied to items belonging to the same cluster, and items belonging to different clusters are By performing negative zoom propagation, it is possible to obtain the same effect as browsing the search results classified for each cluster.

  FIG. 20 is an explanatory diagram illustrating an example of a zooming operation on a search result list of clustered web pages. Here, “continuous display using zooming of clustered search results” will be described. As shown in FIG. 20, when a user selects a web page (“3. web page C”) from the list of web pages and zooms in (operation reception unit), the user belongs to the same cluster as the web page selected by the user. A web page (“1. web page G”) is enlarged and displayed in conjunction with the selected web page and belongs to a different cluster (“2. web page E”, “4. web page F”). , “5. Web page B”) is reduced and displayed in conjunction.

  As described above, according to the present embodiment, the index stored in the index database generated based on the keyword of the web page is used to generate the search result of the web page for the user query, and the searched web page is Clustering and displaying the search results of clustered web pages continuously using the zooming description for zooming operation / behavior that continuously realizes change in detail and display media transition within the same screen in the cluster unit to which it belongs Let As a result, while maintaining the browsing context, it is possible to efficiently perform an information search without having to search many times, and in particular when a plurality of items belonging to different clusters are mixed, the desired cluster This makes it easier to find items belonging to.

  Further, according to the present embodiment, when items belonging to a plurality of clusters are mixed in the search result, if an item selected by the user is zoomed in / out, the items belonging to the same cluster are linked and zoomed in / out. In addition, since items belonging to different clusters are zoomed out / in on the contrary, it is easy to find a desired item.

  In the present embodiment, when the user selects a web page and zooms in on the clustering display unit 72, the web page belonging to the same cluster as the web page selected by the user is enlarged in conjunction with the selected web page. The web pages belonging to different clusters are displayed in a reduced scale in conjunction with each other, but the present invention is not limited to this. For example, when the user selects a web page and zooms in, the web page belonging to the same cluster as the web page selected by the user may be simply enlarged and displayed in conjunction with the selected web page. When the user selects a web page and zooms in, the web page belonging to the same cluster as the web page selected by the user may be displayed in a reduced scale in conjunction with the selected web page.

  In the present embodiment, the server 1 which is the content processing apparatus generates an index of the web page for the text of the web page and stores it in the index database, and the indexing unit 31 stores the index in the index database. A ranking unit 32 that generates a search result of a web page for a user query using an index, and accepts a web page group of search results by the ranking unit 32, but is not limited thereto. For example, the server 1 which is a content processing apparatus does not include the indexing unit 31 and the ranking unit 32, and the search part is performed by another device, or the search result is received using a general-purpose search engine. It may be.

  In this embodiment, a Web page search based on a keyword is performed for a general Web page on the network N which is the Internet, and a search result list of the Web page is output in a zooming description language (ZDL) format. However, the search is not limited to a general Web page on the network, and a search may be performed for a Web page having a fixed structure existing in a local database in the server.

DESCRIPTION OF SYMBOLS 1 Content processing apparatus 23 Display part 31 Indexing part 32 Ranking part 33 Clustering part 41 Keyword extraction part 42 Index generation part 51 Query input part 52 Query conversion part 53 1st similarity calculation part 54 Ranking generation part 61 2nd similarity Degree calculation unit 62 Cluster generation unit 71 Ranking display unit 72 Clustering display unit

Bill Kules and Ben Shneiderman, "Categorized graphical overviews for web search results: An exploratory study using US government agencies as a meaningful and stable structure," Proceedings of the Third Annual Workshop on HCI Research in MIS, Washington, DC, December 10-11 , 2004

Claims (11)

Among the contents clustered into a plurality of clusters, for the content belonging to one cluster, at least one of the detail level change and display media transition within the same screen is continuously realized in units of the one cluster, For the content belonging to a cluster, at least one of the detail change and display media transition within the same screen is linked to at least one of the detail change and display media transition of the content belonging to the one cluster. A zooming description generation unit that describes a zooming operation / behavior that is continuously realized in units of clusters;
An operation accepting unit that accepts a zoom-in / out operation on predetermined content from a user;
In accordance with the zoom-in / out operation accepted by the operation accepting unit, in accordance with the description generated by the zooming description generating unit, the clustered content is changed in detail in one cluster unit to which the operated content belongs and the display media transition A clustering display unit that performs at least one of the following, and performs at least one of the detail level change and the display media transition in association with the other cluster unit,
With
The clustering display unit hides content belonging to a cluster different from the content selected by the user when the user selects content and zooms in.
A content processing apparatus. When the user selects content and zooms in, the clustering display unit zooms out the content belonging to a different cluster from the content selected by the user and hides it.
The content processing apparatus according to claim 1. When the user selects content and zooms in, the clustering display unit zooms in on content that belongs to the same cluster as the content selected by the user in conjunction with the selected content.
The content processing apparatus according to claim 1. When the user selects content and zooms in, the clustering display unit zooms out content belonging to the same cluster as the content selected by the user in conjunction with the selected content.
The content processing apparatus according to claim 1. A search result receiving unit that receives a content group that is a result of being searched and ranked according to a search request;
A clustering unit for clustering the content group of the search results received by the search result receiving unit;
The content processing apparatus according to claim 1, further comprising: An indexing unit that generates a content index for content text and stores it in an index database;
Using an index stored in the index database, a ranking unit that generates a search result of content for a user query;
Further comprising
The search result receiving unit receives the content group of the search result by the ranking unit;
The content processing apparatus according to claim 5. The indexing unit includes a keyword extraction unit that extracts a representative keyword from content text, and an index generation unit that generates a content index based on the keyword extracted by the keyword extraction unit.
The content processing apparatus according to claim 6. The ranking unit converts a query input unit that allows a user to select one or a plurality of keywords from the extracted representative keywords, and a query that is a keyword selected by the user into the same format as an index. A query conversion unit, a first similarity calculation unit for calculating a similarity to a query of each content using an index stored in the index database, and a similarity calculated by the first similarity calculation unit A ranking generation unit that generates a search result of content for a user query based on
The content processing apparatus according to claim 6. The clustering unit uses an index stored in the index database to calculate a second similarity calculation unit that calculates a relative similarity between contents, and a relative similarity calculated by the second similarity calculation unit A cluster generation unit that generates a clustered search result for each content search result,
The content processing apparatus according to claim 6 . On the computer,
Among the contents clustered into a plurality of clusters, for the content belonging to one cluster, at least one of the detail level change and display media transition within the same screen is continuously realized in units of the one cluster, For the content belonging to a cluster, at least one of the detail change and display media transition within the same screen is linked to at least one of the detail change and display media transition of the content belonging to the one cluster. A zooming description generating means for describing a zooming operation / behavior that is continuously realized in units of clusters;
Operation accepting means for accepting a zoom-in / out operation for predetermined content from a user;
In accordance with the zoom-in / out operation accepted by the operation accepting means, according to the description generated by the zooming description generating means, the clustered content is changed in detail level and the display media transition in one cluster unit to which the operated content belongs. Clustering display means for performing at least one of the above, and performing and displaying at least one of detail level change and display media transition in association with this,
And execute
The clustering display unit hides content belonging to a cluster different from the content selected by the user when the user selects content and zooms in.
A content processing program. A content processing method in a content processing apparatus for generating and outputting content search results,
Among the contents clustered into a plurality of clusters, for the content belonging to one cluster, at least one of the detail level change and display media transition within the same screen is continuously realized in units of the one cluster, For the content belonging to a cluster, at least one of the detail change and display media transition within the same screen is linked to at least one of the detail change and display media transition of the content belonging to the one cluster. A zooming description generation process that describes zooming operations / behaviors that are continuously realized in cluster units;
An operation accepting step for accepting a zoom-in / out operation for predetermined content from a user;
In accordance with the zoom-in / out operation accepted in the operation accepting step, in accordance with the description generated by the zooming description generating step, the clustered content is changed in detail level and the display medium transition in one cluster unit to which the operated content belongs. A clustering display step of performing at least one of the above, and performing and displaying at least one of the detail level change and the display media transition in association with this,
Including
In the clustering display step, when the user selects content and zooms in, the content belonging to a cluster different from the content selected by the user is hidden.
A content processing method characterized by the above.