JP2737400B2 - Document retrieval device in document storage system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a search for a structured multimedia document in which a plurality of document parts composed of various media are structured so as to have a parent-child relationship with each other. The present invention relates to a document search device in a document storage system that stores multimedia documents.

[0002]

2. Description of the Related Art Structured multimedia documents of this kind are:
It is configured by giving a parent-child relationship to a plurality of document parts composed of various media, that is, characters, graphics, images, and the like. The parent document part has at least one child document part, and the child document part can be given a child document part by considering the child document part as a parent. Therefore, if a parent document part is used as a root, and a child document part and a grandchild document part with respect to the parent are used as branches, a structured multimedia document is configured by giving a plurality of document parts a tree structure relationship. It can be said that.

[0003] In a document storage system in which a large number of structured multimedia documents or ordinary unstructured documents are stored, the following two search methods are used to search for a desired document. there were.

[0004] On the other hand, a search method is to input an attribute given in advance for each document, that is, the title, creation date, owner name, etc. of the document as desired, and to search for a document corresponding to the input attribute. is there.

[0005] The other search method is to input desired contents, that is, words, graphic parts, image parts, and the like, and to search for a document containing the input contents.

[0006]

However, the search method based on the attribute of the document out of the above two conventional search methods does not involve the contents of the document at all, so that a detailed search cannot be performed. In addition, the method of searching based on the content of the other document has a disadvantage that it takes a long time to search since the entire document is checked to every corner. Furthermore, a search combining the above two methods, that is, selecting a relevant document based on the attribute of the document, and then examining whether or not this document contains desired contents can be performed. Even when a media document is a search target, the search was performed without considering the parent-child relationship of each document part, which is the structure of this document, and there was no more effective search method for structured multimedia documents. .

Therefore, the present invention focuses on the parent-child relationship of a structured multimedia document, and enables a detailed and efficient search, and a document search apparatus in a document storage system capable of reliably finding a desired document. The purpose is to provide.

[0008]

According to the present invention, there is provided a parent-child relationship retrieval unit for retrieving a parent-child relationship between a plurality of document parts, a content retrieval unit for retrieving the contents of the document parts in accordance with the medium of the document parts, Attribute search means for searching for a predetermined attribute of the document part; input means for designating at least one of the parent-child relationship, the content of the document part included in the parent-child relationship, and the attribute of the document part as a search condition And a search integration means for performing a search by integrating the parent-child relationship search means, the content search means, and the attribute search means. The search integration means, when a search condition is designated by the input means, Search for a structured multimedia document that meets the conditions.

[0009]

According to the present invention, when at least one of the parent-child relationship, the content of the document part included in the parent-child relationship, and the attribute of the document part is designated by the input means as the search condition, the search integration means performs the search. A structured multimedia document that satisfies the condition is searched for through a parent-child relationship search unit, a content search unit, and an attribute search unit. Therefore, the retrieved structured multimedia document satisfies at least one of the search condition, that is, the parent-child relationship, the content of the document component included in the parent-child relationship, and the attribute of the document component.

[0010]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an embodiment of a document search device in a document storage system according to the present invention. In FIG. 1, a sentence database 1 is a large-capacity memory including, for example, a magnetic disk, and stores a large number of structured multimedia documents. As described above, this structured multimedia document has a parent-child relationship tree structure such as a parent document part, a child document part, and a grandchild document part, and each document part has a plurality of types of document parts. The medium is composed of at least one medium of characters, figures, and images.

The structured multimedia document search device 2 is for searching for a document in the text database 1, and includes a character content search unit 3, a graphic content search unit 4, and an image content search unit for searching the contents of document parts. 5, an attribute search unit 6 for searching for an attribute given to the document part in advance,
The system includes a parent-child relationship search unit 7 for searching for a parent-child relationship of a document, and a search integration unit 8 for integrating these search units and executing a search.

The character content search unit 3 searches for character content composed of characters in a document part, and the graphic content search unit 4 performs graphic content (for example, a function indicating a graphic) composed of a graphic in a document part. To search for
Further, the image content search unit 5 searches for an image content composed of an image (for example, an image formed on a bitmap) in the document part.

The keyboard 9 and the mouse 10 are operated to input various data, commands, and the like. A display corresponding to the input by the operation is displayed on the display screen of the display device 11.

The input / output control device 12 controls input / output between the structured multimedia document search device 2 and the keyboard 9, mouse 10, and display device 11, and is formed on a display screen of the display device 11, for example. The structured multimedia document obtained is transmitted to the search integration unit 8 through the input / output control device 12, and is stored in the sentence database 1 from the search integration unit 8. Also, the search integration unit 8
The structured multimedia document read from the text database 1 is transmitted to the display device 11 through the input / output control device 12 and displayed on the display screen.

Now, in such a configuration, when a desired multimedia document is to be searched from a large number of structured multimedia documents in the document database 1, a search condition must be set. The procedure for setting the conditions will be described below.

First, by operating the keyboard 9 and the mouse 10, a parent-child relationship search formula for indicating the parent-child relationship of the multimedia document as shown in FIG. 2 is formed on the display screen of the display device 11.

Here, the parent-child relation retrieval formula is a multimedia document including a parent document part 101 and two child document parts 102 and 103 placed below the parent document part 101, or A multimedia document including a document part 101 of the parent document part 101 and one child document part 104 placed below the parent document part 101 is indicated.
That is, a multimedia document having a parent-child relationship as shown in FIG. 3 or a multimedia document having a parent-child relationship as shown in FIG. 4 is specified.

Subsequently, by operating the keyboard 9 and the mouse 10, the parent document part 10 shown in FIG.
The attribute values of the document components 102 and 103 are designated on the display screen of the display device 11. For example, the attribute value of the parent document part 101 is the part name of the document part 101, and as shown in FIG.
Is designated as a component name. At this time, the name of a content search formula (to be described later) indicating the content of the parent document part 101, that is, the content search formula name A is specified together with the attribute value. Similarly, an attribute value and a content search formula name are specified for each of the child document parts 102 to 104. For example, the attribute value of the child document part is "Main", "Chapter 1", "Chapter 2", and the like.

Note that the attribute value is used when creating a document part.
The document part is given to the document part so that the document part can be identified, and the document part is stored in the document database 11 together with its attribute value. As attribute values, not only the part names such as “Introduction” and “Main Discussion” but also the creation date,
It may be the owner's name.

Subsequently, by operating the keyboard 9 and the mouse 10, the content retrieval formula of the content retrieval formula name previously set for each of the parent document component 101 and each of the child document components 102 to 104 is displayed. 1
1 on the display screen. For example, the content search formula of the content search formula name A of the parent document part 101 shown in FIG.
It is as shown in the figure, and consists of a content system, display fonts and contents. Here, "character" is designated as the content system, "Gothic" is designated as the display font, and the word "invention" is composed of 50 characters or more.
"" Is specified. That is, the content search formula set for the parent document part 101 includes a character as a medium of the document part, the font of the character is Gothic, the document part is composed of 50 characters or more, and “ It contains the word "invention."

Another example different from the content search formula shown in FIG. 6 is as shown in FIGS. 7 and 8.

The search formula shown in FIG. 7 is given the name of the content search formula name B, "figure" is specified as the content system, "10 or less" is specified as the number of graphics, and "circle. Includes triangles. " That is,
This content search expression includes a figure as a medium of a document part, indicates that the number of figures is 10 or less, and includes a circle and a triangle.

The search formula shown in FIG. 8 is given the name of the content search formula C, "image" is specified as the content system, and "monochrome ratio 50% or less" is specified as the content. In other words, this content search expression includes an image as a medium of a document part, and indicates that the black occupancy is 50% or less for each of the black and white pixels constituting the image.

As described above, the parent-child relation composed of each document part is indicated by the parent-child relation expression, the attribute of each document part is indicated by each attribute value, and the content of each document part is expressed by each content relation expression. , The search condition of the multimedia document is set.

According to this search condition, the multimedia document has a parent-child relationship between the parent document part 101 and the two child document parts 102 and 103 as shown in FIG. The parent document part 101 as shown in the figure
And a parent-child relationship with one child document part 104 must be established. In addition, the parent document part 101 has a part name called “introduction” as an attribute value as shown in FIG. 5, and includes a character as a medium as shown in FIG. And the document part must be composed of 50 characters or more and include the word "invention".

The child document parts 102 to 103 have attribute values and content retrieval formulas set in the same manner as the parent document part 101.
It is assumed that a part name called "the main theme" is set as the attribute values of 02 and 103, respectively, and a content search formula related to the graphic shown in FIG. 7 is set. Similarly, the attribute value of the child document part 104 is set to “this book”
It is assumed that a part name is set and a content search formula related to the image shown in FIG. 8 is set.

When the search condition is set on the display screen of the display device 11, the search condition is notified to the search integration section 8 via the input / output control device 12.

When the search integration unit 8 is notified of the search condition, the search integration unit 8 first searches the parent-child relation expression shown in FIG.
Notify. When notified of the parent-child relation expression, the parent-child relation search unit 7 sequentially searches each multimedia document in the document database to find a multimedia document having a parent-child relation corresponding to the parent-child relation expression. As a result, a multimedia document having the parent-child relationship shown in FIG. 3 or FIG. 4 is found. Each time the parent-child relationship search unit 7 finds the relevant multimedia document, it instructs the search integration unit 8 with a pointer to the multimedia document together with its parent-child relationship.

When the pointer to the multimedia document is indicated by the parent-child relationship search unit 7 together with the parent-child relationship from the parent-child relationship search unit 7, the attribute value of each document component in the multimedia document according to the parent-child relationship, and the A pointer to the multimedia document is instructed to the attribute search unit 6. For example, if the multimedia document has the parent-child relationship shown in FIG. 3, the attribute value “Introduction” of the parent document part 101 and the attribute value “Main theme” of each of the child document parts 102 and 103 are the multimedia document. Along with the attribute search unit 6. For example, if the multimedia document has the parent-child relationship shown in FIG. 4, the attribute value “introduction” of the parent document part 101 is
The attribute value “main subject” of the child document part 104 is instructed to the attribute search unit 6 together with the multimedia document.

When the attribute search unit 6 is instructed with the attribute value of each document part according to the parent-child relationship and a pointer to the multimedia document, the attribute search unit 6 accesses the multimedia document in the document database 11 based on the pointer, The attribute value of each document part in the multimedia document is checked, and it is determined whether the attribute value of each document part obtained by the check matches each of the designated attribute values. Then, the attribute search unit 6 notifies the search integration unit 8 of this determination result.

If the search integration unit 8 obtains a determination result that the attribute value of each document part obtained by the search does not match the specified attribute value, the search-integrating unit 8 determines that the multimedia document to be searched is It is determined that the search condition is not satisfied, and the search for the multimedia document is stopped. Further, if the determination result that the attribute values of the two match each other is obtained, the search integration unit 8 converts the content search formula of each document part in the multimedia document into the character content search unit 3, the graphic content search unit 4, It is distributed to the image content search unit 5 and given. At this time, the content search formula of each document component is sequentially applied from the content search formula of the parent document component to the content search formula of the child document component, that is, from the upper order to the lower order.

Here, for example, the multimedia document is the third
Assuming that the document part 101 is composed of a parent document part 101 and two child document parts 102 and 103 as shown in the figure, the parent document part 101 is given a content search formula for characters shown in FIG. And each of the child document parts 102 and 103
Is given with the content search formula for the graphic shown in FIG. 7, the search integration unit 8 firstly executes the content search formula for the character shown in FIG. 6 together with the pointer to the document component 101 in the multimedia document. It instructs the character content search unit 3. Upon receiving the instruction, the character content search unit 3 determines whether or not the document part 101 of the multimedia document in the document database 11 satisfies the content search formula for characters shown in FIG. That is, the document part 101 includes a character, and the font of the character is Gothic,
It is determined whether it is composed of 50 characters or more and includes the word "invention". Then, the character content search unit 3 notifies the search integration unit 8 of the determination result.

When the search integration unit 8 receives the result of the determination that the document component 101 in the multimedia document does not satisfy the content search formula shown in FIG. 6, the search integration unit 8 stops searching for the multimedia document. If it is determined that the content search formula is satisfied, the search integration unit 8 applies the content search formula for the graphic shown in FIG. 7 to one of the child document parts 102 and 103 in the multimedia document. An instruction is given to the graphic content search unit 4 together with the pointer. The graphic content search unit 4 executes the content search formula and, for example, the child document part 102.
When the pointer to is pointed to, the document part 102 of the multimedia document in the document database 11
It is determined whether or not a content search formula for the graphic shown in the figure is satisfied. That is, it is determined whether or not the document part 102 includes a figure, the number of figures is 10 or less, and a circle or a triangle is included. The determination result is notified from the figure content search unit 4 to the search integration unit 8. You.

When the search integration unit 8 receives the result of the determination that the document component 102 in the multimedia document does not satisfy the content search formula shown in FIG. 7, the search integration unit 8 stops searching for the multimedia document. When it is determined that the content search formula is satisfied, the search integration unit 8 converts the content search formula related to the graphic shown in FIG.
An instruction is given to the graphic content search section 4 together with the pointer to the pointer 3. The graphic content search unit 4 is configured to execute the content search formula and the child document part 10.
When 3 is specified, the same determination as that of the child document part 102 is performed. That is, it is determined whether or not the document part 103 of the multimedia document in the document database 11 satisfies the content search expression relating to the graphic shown in FIG. Will be notified.

When the search integration unit 8 receives the result of the determination that the document component 103 in the multimedia document does not satisfy the content search formula shown in FIG. 7, the search integration unit 8 stops searching for the multimedia document. If it is determined that the content search formula is satisfied, the integration determination unit 8 determines that each of the document components 101 to 103 in the multimedia document has satisfied the content search formula.

It should be noted that even when the multimedia document to be searched is composed of the parent document part 101 and the child document part 104 as shown in FIG. 4, the multimedia document of FIG. The same determination as described above is repeated.
That is, each time a multimedia document having a parent-child relationship as shown in FIG. 4 is searched from the document database 11 by the parent-child relationship search unit 7, the attribute value of each document part in the multimedia document is output by the attribute search unit 6. Is determined whether the search condition is satisfied, and if it is determined that the search condition is satisfied, each content search unit determines whether the content of each document component satisfies the search condition.

In the process of searching for such a multimedia document, after a multimedia document having a parent-child relationship, which is one of the search conditions, is found, the attribute values of the document parts in the multimedia document satisfy the search condition. If not, or if the content of the document part does not satisfy the search condition, the search of the multimedia document is immediately stopped, and the next found document part in another multimedia document having the parent-child relationship is searched for. Will be searched for respectively.

If a multimedia document having a parent-child relationship, which is one of the search conditions, is found, and it is determined that the attribute values and contents of each document part in the multimedia document satisfy the search condition, the search is performed. After determining that the multimedia document satisfies all the search conditions, the integration unit 8 subsequently searches for another multimedia document.

In this way, the search of all the multimedia documents in the document database 11 is completed, and if there is at least one multimedia document that satisfies all the search conditions, the search integration unit 8 determines that the multimedia document satisfies all the search conditions. A list of documents is created, and the list is transmitted to the display device 11 through the input / output control device 12. As a result, a list of multimedia documents that satisfy all the search conditions is displayed on the display screen of the display device 11.

Here, by operating the keyboard 9 and the mouse 10 to select a desired multimedia document from the displayed list of multimedia documents and instruct the reading of the document, the input / output control is performed. Apparatus 1
2 to the search integration unit 8. In response,
The search integration unit 8 reads the selected multimedia document from the document database 11 and transmits the multimedia document to the display device 11 through the input / output control device 12. As a result, the selected multimedia document is displayed on the display screen of the display device 11.

As described above, in the present embodiment, when the parent-child relationship of the multimedia document, the attribute value of each document part constituting the parent-child relationship, and the content of each document part are designated as search conditions, a multimedia that satisfies all the search conditions is specified. Because the media documents are searched, you can do a fine-grained search,
Thus, it is possible to reliably find a desired multimedia document. In this search, a multimedia document having the designated parent-child relationship is found,
Next, it is determined whether or not the attribute value of each document part in the multimedia document satisfies a search condition, and thereafter, it is determined whether or not the content of each document part satisfies the search condition. If it is determined that the search condition is not satisfied, the search for the multimedia document is immediately stopped, and the search for the next other multimedia document is performed, so that the search process can be performed quickly. .

In the above embodiment, the attribute values and the contents are searched for all the document parts in the multimedia document having the designated parent-child relationship. , And the attribute value and the content may be searched for only some of the document parts.

Further, only one of the attribute value and the content of the document part may be used as a search condition, and the other may be excluded from the search condition. For example, if the contents of the document part are excluded from the search conditions, the name of the contents search expression shown in FIG. 5 is not indicated when indicating the attribute value of the document parts. Absent.
If the attribute value of the document part is excluded from the search condition, when indicating the attribute value of the document part, it is indicated that there is no limitation as the attribute value shown in FIG. Instruction, and thereafter, the content search expression of the name is indicated.

Further, not only a parent-child relationship composed of a parent document component and a child document component, but also a complicated parent-child relationship composed of a number of layers such as a parent document component, a child document component, and a grandchild document component are indicated. You may make it.

In the tree structure of the parent-child relationship, any document component can be regarded as a parent document component, and the lower order of the document component can be regarded as a child document component. To indicate that the parent document part does not need to be the document part corresponding to the root of the multimedia document, thereby searching for a parent-child relationship that regards any document part as a parent. Is also good.

Furthermore, although the medium is exemplified by characters, figures, and images, the present invention is not limited to these. If there is another medium, a content search unit corresponding to the other medium may be added. Needless to say, it can respond to the diversification of media.

[0048]

As described above, according to the document retrieval apparatus in the document storage system according to the present invention, at least one of the parent-child relationship, the content of the document component included in the parent-child relationship, and the attribute of the document component. Since a structured multimedia document that satisfies the search condition of can be searched, it is possible to perform a fine and efficient search, and thereby to reliably find a desired document.

[Brief description of the drawings]

FIG. 1 is a functional block diagram showing an embodiment of a document search device according to the present invention.

FIG. 2 is a diagram illustrating a parent-child relational expression.

FIG. 3 is a diagram exemplifying a parent-child relationship corresponding to the relational expression shown in FIG. 2;

FIG. 4 is a diagram illustrating a parent-child relationship corresponding to the relational expression shown in FIG. 2;

FIG. 5 is a view exemplifying a display mode when indicating an attribute value.

FIG. 6 is a diagram showing a content search formula relating to characters.

FIG. 7 is a diagram showing a content search formula for a graphic.

FIG. 8 is a diagram showing a content search formula relating to an image.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Document database 2 ... Structured multimedia document search device 3 ... Character content search part 4 ... Graph content search part 5 ... Image content search part 6 ... Attribute search part 7 ... Parent-child relation search part 8 ... Search integration part 9 ... Keyboard 10 mouse 11 display device 12 input / output control device

 ──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Noriyuki Uebayashi 3-57-6 Yoyogi Shibuya-ku, Tokyo Granforet Fuji Xerox Co., Ltd. (56) References JP-A-60-160475 (JP, A) JP-A-1-293433 (JP, A) Takahiro Aoki Kazuki Omon Yujiho Sasaki, "Study on Search System for Technical Papers Based on Document Structure", Kogakuin University Research Report No. 66, PP216-220, April 1989 March 28 Ryota Suzuki Noboru Sonehara Naoki Kobayashi et al., “Study of Total Media Architecture-One Development from DA”, Television Society of Japan Technical Report VOL. 12 NO. 9 P52, February 1988 Yoshihide Kikuchi Hitoshi Miyai, "Prototype Text Search System Using ISSP-Focusing on Searching Logical Structure Units-", Information Processing Society of Japan, 35th (late 1987) National Congress Lecture Papers PP1285-1286, 1987

Claims (1)

(57) [Claims] 1. A document storage system for storing a structured multimedia document in which a plurality of document parts composed of various media are structured so as to have a parent-child relationship with each other, and a parent-child relationship search for searching the parent-child relationship. Means, content search means for searching the contents of the document part according to the medium of the document part, attribute search means for searching for a predetermined attribute of the document part, and the parent-child relationship and the parent-child relationship Input means for instructing at least one of the content of the included document part and the attribute of the document part as a search condition, and causing the parent-child relation search means to search for the parent-child relationship specified by the input means, When a structured multimedia document that matches the relationship is found, the searched structured multimedia document is input by the input unit. A document comprising: search integration means for causing at least one of the content search means and the attribute search means to search for at least one search condition of the content of the indicated document part and the attribute of the document part. Document retrieval device in storage system.