JPH1153373A - Device and method for data retrieval and storage medium storing data retrieval program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To smoothly retrieve the temporary data which are not desired to be registered into a data base by preparing a 2nd storage means that stores a 2nd data base containing the attribute information on the temporarily registered data in addition to a 1st data base that is stored in a 1st storage means. SOLUTION: When the retrieval instruction given from an input device designates also the temporary data as a retrieval object, a CPU 101 selects the temporary data to be retrieved, e.g. an image file that is stored in an external storage 107 such as a CD-ROM, etc., and reads the image data out of the selected image file. Then the CPU 101 extracts and generates the image feature value and keywords, i.e., the attribute information from the image data stored in the image file of the storage 107 to register them in a memory in the same method as that which is applied when the attribute information is registered in a normal data base. In such cases, a new data base is prepared in a RAM 106 to store the temporary data and the attribute information and their related image data are registered in this data base.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data search method and apparatus and a storage medium storing a control program therefor, and relates to a data search technique.

[0002]

2. Description of the Related Art The flow of registration / search in a conventional database search program will be described with reference to FIG. First, the database search program performs initial settings for performing a database search in step S101.

In step S102, it is determined whether the instruction input by the user is data registration or database search. If it is determined in step S102 that the instruction is a registration instruction, the process proceeds to step S103, and data to be registered is read from a data file stored in a storage unit such as a hard disk or a CD-ROM.

Subsequently, in step S104, attribute information is extracted and generated from the registration data read in step S103.
In brief, for example, a word is extracted as a keyword from document information in registration data. In this word extraction, characters in a sentence are analyzed two by two characters, and the presence of a word of a level registered in the dictionary is determined. Also,
Sometimes we search not only words in a sentence but also information on headers and so on. In other cases, an image feature amount may be extracted other than as a keyword. The image feature is a feature unique to image data, and calculates the hue, brightness, contrast, and the position where those differences appear.
There is a technique for generating a value.

Next, in step S105, the information (keyword and image feature amount) extracted and generated in step S104 is registered in a database in a storage means such as a hard disk in association with a data file (with link information). Upon completion of the registration process, the process returns to step S102 and waits for an instruction from the user.

If the instruction from the user is a search in step S102, the process proceeds to step S106, and a search is performed from a database already registered in step S105. For example, when the keyword “sea” is specified by the user, the keyword “sea” is searched for from the keywords registered in the database, and a data file having a matching keyword is searched based on the link information. .

Next, in step S107, the result searched in step S106 is output to an output device such as a display.

[0008] As described above, the conventional search is performed.

[0009]

However, in the conventional search, there is a problem that the user does not want to register in the database, but must always register in the database to perform the search even if there is data to be treated as a search target. was there.

[0010] Even if the user wants to search immediately but does not have time to register the data in the database, the database must be accessed when the database is registered or searched, and the amount of the database is limited to the document data. There is a problem that it takes a considerable amount of unnecessary time because it is necessary to search for unnecessary data if the keyword is unnecessary.

Another problem is that after performing a temporary search, it is not possible to select whether to register the attribute information used in the search in the database.

[0012] In addition, since only one image can have attribute information, there is a problem that it takes a long time to perform a search process.

[0013] Further, since it has only attribute information such as keywords and image feature values, there is a problem that it is not possible to search for which image or color exists in which part of the image.

In addition, since there is only one attribute information search condition, there is a problem that it is not possible to narrow down the search by a plurality of conditions.

Further, since the entire image is displayed as a search result, there is a problem that the display processing takes a long time.

[0016]

As means for solving the above-mentioned problems, a first storage means for storing a first database including attribute information of registration data, and a first storage means for storing the first database stored in the first storage means. By having a second storage means for storing a second database including attribute information of temporary registered data separately from the database, a normal database and a database for temporary data are separated, so that high-speed data retrieval is performed. Becomes possible.

[0017] Further, by further providing a synthesizing means for registering the second database in the first database, it is possible to register a temporarily used database in a formal database.

[0018] Further, since the synthesizing means further comprises a selecting means for selecting whether or not to register the second database in the first database, it is possible to register the attribute information used in the search in a regular database. It is now possible to select whether or not the user wants.

[0019] In addition, the provision of the registration means for registering attribute information of data to be registered in the first or second database makes it possible to register attribute information in the database.

Further, the apparatus further comprises an input means for inputting attribute information to be searched and a search means for searching for the attribute information input by the input means, so that a high-speed search can be performed. .

The means described above is not limited to a device, but can also be realized by a storage medium storing a program that can be realized by controlling a method and an information processing device.

[0022]

BEST MODE FOR CARRYING OUT THE INVENTION

(First Embodiment) Hereinafter, the present invention will be described in detail with reference to the drawings.

FIG. 2 shows an example of the overall configuration of the data search device of the present invention.

In the configuration shown in FIG. 2, reference numeral 101 denotes a microprocessor (CPU) which performs operations for image information retrieval, logical judgment, and controls each component connected via a system bus 102.

The system bus 102 has a CPU
An address bus for transferring an address signal designating a component to be controlled by the CPU 101, a control bus for transferring a control signal of each component to be controlled by the CPU 101, and a communication between the components (devices). And a data bus for performing data transfer.

Reference numeral 103 denotes an input device, which is a pointing device such as a mouse or an electronic pen for inputting an instruction by moving a pointer or the like; a character symbol input key such as an alphabet key, a hiragana key, a katakana key, and punctuation; It assists the user in inputting, such as a keyboard having some of various function keys such as a search key and a cursor movement key for instructing cursor movement.

Reference numeral 104 denotes a scanner, which controls input from a document on a recording medium such as paper placed on a scanner base as bit image data.

Reference numeral 105 denotes a read-only memory (ROM), which is a fixed read-only memory and stores a system program.

Reference numeral 106 denotes a writable random access memory (RAM) composed of 16 bits per word, which is used for temporarily storing various data from each component.

Although not shown, the RAM 106 temporarily stores an input search word storage buffer WBUF, a search target identification buffer FBUF, and a search result output buffer OBUF. In addition to storing application programs, the CPU 1
01 is used as a working memory for arithmetic processing.

Reference numeral 107 denotes an external storage unit typified by a hard disk, a floppy disk, or the like. A reference table 107 associates an image ID described later with reference to FIG. 3 with an image file storage path.
A reference table in which image feature amounts are associated, a reference table in which image feature words and image features described later are associated in FIG. 5, a program executed in accordance with a flowchart described later in FIG. The word dictionary DIC storing information for word division and word type determination referred to in the step is stored. A data file of image data to be searched in the present embodiment is stored in a hard disk, a floppy disk, a CD-ROM, or the like, which is the external storage unit 107.

Reference numeral 108 denotes a display control unit (CRTC: CRT controller), which plays a role of displaying the contents stored in the display video memory on the display 109.

Reference numeral 109 denotes a display (CRT) for displaying the contents stored in the display video memory.

110 is a character generator (CG)
The character code input as the character code is developed as bit image data based on the font stored in the font storage unit in the ROM or CG.

Reference numeral 111 denotes a network interface (N
et / I / F), and is connected to a network via this network interface.

The data retrieval apparatus of the present invention comprising the above-described components operates according to various inputs from the input device 103 and various inputs supplied from the network interface 111. When an input from the device and an input from the network interface are supplied, an interrupt signal is first sent to the CPU.
To the external storage unit 107
Various control signals stored therein are read out, and various controls are performed according to the control signals.

In the data retrieval apparatus of the present invention comprising such components, when an input from the input device 103 and an input from the network interface 111 are supplied, first, an interrupt signal is sent to the microprocessor CPU 101, Microprocessor CPU
The 101 reads various control signals stored in the external storage unit 107, and performs various controls according to the control signals.

FIG. 12 shows an example in which the image shown in FIG. 11 is divided into a plurality of tiles. In this example, the target image is divided into a total of 20 tiles of 4 tiles in the vertical direction and 5 tiles in the horizontal direction. The numbers in the figure indicate tile numbers, and are 0, 1, 2,. . . Are sequentially numbered.

The following describes an example of an image format for dividing an image into tiles as described above.

[Explanation of FlashPix ^™ file format] FlashPix ^™ (FlashPix is based on Eastman K
In the odak (registered trademark) file format, the image attribute information and image data stored in the image header section are further structured and stored in a file. This structured image file is shown in FIGS.

Each property and data in the file is MS
-Access through storage and streams, equivalent to DOS directories and files. 13 and 14, the shaded portion is the storage and the unshaded portion is the stream. Image data and image attribute information are stored in the stream portion. The image data is hierarchized at different resolutions.
solution 0, 1, ... n. For each resolution image, the information required to read that image is
eader and the image data are stored in Sbuimage data.

A property set is defined by classifying attribute information according to the purpose of use and the contents.
Info. Property Set, Image Info. Property Set, Imag
e Content Property Set, Extension list property Se
There is t.

[Explanation of each property set] The first Su
The mmary Info. Property Set is not specific to FlashPix, but is an essential property set in Microsoft's structured storage. It stores the title, title, author, thumbnail image, etc. of the file.

The second Image Contents Property Set is an attribute that describes a method of storing image data (FIG. 15).
This attribute includes the number of layers of image data, the width and height of the image with the maximum resolution, the width, height, and color configuration for each resolution image, or the quantization table and Huffman table when using JPEG compression. Write the definition.

The third Image Info. Property Set stores various information that can be used when using an image, for example, information on how the image is captured and how it can be used. For example, there are the following. -Information on how to capture and / or generate digital data (File Source)-Information on copyright (Intellectual property)-Information on the content of the image (person, location, etc. in the image) (Content description)-Used for shooting Camera information (Camera inform
)) Information on camera settings (exposure, shutter speed, focal length, use of flash, etc.) at the time of shooting (Per Picture camera settings) ・ Information on digital camera specific resolution and mosaic filter (Digital camera characterization) ・ Film Manufacturer name, product name, type (negative / positive,
Information such as color / black and white (Film description) ・ Information on the type and size when the original is a book or print (Original document scan descriptio
n) ・ For scanned images, use the scanner or software
Information about the person who operated (Scan device)

The fourth Extension list property Set is an area used when adding information not included in the basic specifications of FlashPix.

The FlashPix Image View Object shown in FIG. 14 is an image file that stores viewing parameters and image data used when displaying an image. The viewing parameter is a set of processing coefficients stored for adapting the processing of image rotation, enlargement / reduction, movement, color conversion, and filtering when displaying an image.

Source / Result FlashPix Image Object is F
The source of lashPix image data, Source FlashPix Ima
ge Object is required, Result FlashPix Image Object is optional. Source FlashPix Image Object is the original image data, Result FlashPix Image Object
Stores an image obtained as a result of image processing using the viewing parameters.

The Source / Result desc. Property set is a property set for identifying image data, and stores an image ID, a property set for which change is prohibited, a last update date and time, and the like.

The Transform property set includes Affine conversion coefficients for rotation, enlargement / reduction, and movement, a color conversion matrix,
Stores a contrast adjustment value and a filtering coefficient.

Next, handling of image data will be described.

An image format including images of a plurality of resolutions divided into tiles will be described.

FIG. 5 shows an example of an image file composed of a plurality of images having different resolutions. In FIG. 16, the image of the maximum resolution has a column × row of X0 × Y0, the next largest image is X0 / 2 × Y0 / 2, and thereafter,
The size of each column and row is reduced by 1/2, and the process is repeated until both the columns and rows are equal to or less than 64 pixels.

As a result of the hierarchization as described above, as the attribute information of the image, “the number of layers in one image file” and the image of each layer, the header information and the image data described in the section of the conventional image format are used. Is required. Information about the number of layers in one image file, the width and height of the image with the maximum resolution, or the width, height, color configuration, compression method, etc. of the image with each resolution is described in Image Contents Pro.
The perty set is described in FIG.

Further, the image of each resolution layer is divided into 64 × 64 tiles as shown in FIG. When the image is divided into 64 × 64 tiles sequentially from the upper left of the image,
Depending on the image, blanks may occur in the right and bottom tiles. In this case, 64 × 64 pixels are constructed by repeatedly inserting the rightmost image or the lowermost image. In FlashPix ^™ , the image in each tile is stored in one of JPEG compression, single color, and non-compression. JPEG compression is an image compression method internationally standardized by ISO / IEC JTC1 / SC29, and the description of the method itself is omitted here. The tiled image data is stored in the Subimage data stream, and the total number of tiles, the size of each tile, the data start position, and the compression method are all stored in the Subimage header (FIG. 18).

The single color is a system in which the color of a tile is represented by one color without recording the value of each pixel only when all the tiles are composed of the same color. This method is particularly useful for images generated by computer graphics.

When each image is divided into tiles after the image is hierarchized in this way, a characteristic portion in the image described by the conventional keyword is separated into the respective tiles, and the tile including the respective keyword is separated. And tiles not included.

The image shown in FIG. 11 has a hierarchical image such as the FlashPix format, as shown in FIG. FlashP
The image file in the ix format has images of multiple resolutions. Here, the image with the lowest resolution is set to the resolution 0
(Layer number 0), and the resolutions are sequentially assigned to 1, 2, and 3. The resolution information, which is numerical information indicating the resolution, is hereinafter referred to as image designation information (layer number). This image file is divided for each resolution layer, the divided image data is divided into tiles of 64 × 64 pixels, and keywords included in each tile are described in FIG. That is, in the layer number 3 which is the image designation information, the tiles 0, 1, 2, 5, and 10 are empty, tile 3,
4, 8, 9 sky and sun, tiles 6 and 11 sky and men, tiles 7 and 12 sky and women, tiles 13 and 14 sky and yacht, tile 15 sea, tile 16 Are keywords of sea and men, tile 17 is sea and women, and tiles 18 and 19 are sea and yacht. Similarly, for layer number 0, the keywords of sky, sun, sea, men, women, and yachts are assigned to tile 0.

FIGS. 8, 9 and 10 are also examples of images divided into 20 tiles and stored as in FIG.

FIG. 3 shows a table in which image IDs and image file storage paths are associated with each other in the present invention. Reference numeral 301 denotes an image ID, which is an identification number uniquely assigned to one image file. Reference numeral 302 denotes a file path indicating the storage location of the image file corresponding to the image ID in the external memory DISK14, and corresponds to an MS-DOS directory and file. For example, in the image data of FIG.
Is given by 1563 and the search program of the present invention,
The file path is stored correspondingly. Image ID is 1563
8 (FIG. 8) is a file 080.FPX in a folder named FRUIT on drive X. 9, 10 and 11 are stored in the same manner.

FIG. 4 shows a table in which image IDs are associated with tile numbers as area designation information, keywords as attribute information, and image characteristics. 401 stores the image ID corresponding to 301 in FIG. 3, and 402 stores the tile number of each image ID. Four
Numeral 03 denotes words representing image features drawn on the tile corresponding to the tile number 402, and stores a keyword in a character code (for example, unicode). A plurality of keywords may be stored for one tile number. Reference numeral 404 stores the image feature of the tile corresponding to the tile number. As an image feature, in the present invention, as a representative color of the tile, a color component obtained by averaging the RGB pixel values constituting the tile and reducing the color to 27 colors is used. For the color reduction method, a known method such as a method of simply dividing each pixel value of RGB or a method using a reference table is used. In the figure, the image whose image ID is 1024 has 20 tiles, and the tile of tile number 0 is assigned the keyword “sky”, and a color component “11” is assigned as an image feature.

FIG. 5 is a table in which image feature words are associated with image features. In the present invention, a representative color of a tile is used as an image feature, and a word representing a color is used as an image feature word. In 501, the color name is used as a feature word,
code), and the image feature amount represented by the color name is stored in 502 as a color component. In the drawing, for example, the color name of a component having color components 2, 3, and 4 is "red".

In this embodiment, as described above, the representative color is divided into 27 as the image feature amount. This is R (Red)
Axis, G (Green) axis, and B (Blue) axis are each divided into three equal parts, and the RGB space is divided into 27 cells.
The representative colors are separated. Although the average value of the tile was obtained earlier, other methods are also conceivable. For example, the determination of the representative color of the tile is made by voting where the 64 × 64 pixels included in the tile apply to the 27 equally divided cells, and the cell with the most votes becomes the representative color. The cells divided into 27 equal parts are represented by numbers 0 to 26.

FIG. 6 is a flowchart for explaining the registration / search operation of the data search apparatus of the present invention. The operation will be described below with reference to FIG.

First, in step S601, the CPU 101
Starts a search program from the ROM 105 or the RAM 106 and performs initial setting.

Subsequently, in step S602, the CPU 101
Determines whether the instruction input from the input device 103 is an instruction to search or an instruction to register in the database according to the search program. Here, if it is determined that the input instruction is a search instruction, the process proceeds to step S603, and the registration instruction is a registration of a keyword or an image feature amount from a file as in the related art, or whether the registration instruction exists in the memory. It is determined whether the registration is from a temporary database.

If the CPU 101 determines in step S603 that the registration is from a file, the process proceeds to step S604 to read registration data from the file to be registered. The registration data is a document or image data existing in a file to be registered. For example,
The image data as shown in FIGS.

Next, in step S605, attribute information is extracted and generated from the file. The attribute information is information of registration data obtained by generating an image feature amount or extracting a keyword included in a document as described above. Further, the normal file is divided into an image header section and an image data section as shown in FIG. In the image header part, the image format identifier, the number of pixels of the image,
Contains attribute information such as size and resolution. In addition, one of the image formats, Flashpix (Flass)
In the hPix) format, new information can be added as non-image information, and thus it is possible that the user has added a keyword as non-image information.
Therefore, keywords added as non-image information are also extracted as attribute information. Taking FIG. 8 as an example, “apple”, “banana”, and “background” are extracted as keywords, and “23” is generated as an image feature amount.

It is also possible for a user to actually generate and register a keyword. Keyword extraction
The creation is performed, for example, by analyzing the text of the document data to create a word dictionary DI stored in the external storage unit 107.
Referring to C, extract and create the relevant word. The word dictionary DIC stores commonly used nouns. In the image data, creation of a keyword may be specified by a user, or a keyword may be assigned to an image in a certain area and stored. For example, if there is an image of "apple" in the image data, assign "apple" as a keyword to the tile where "apple" exists, and place it in the corresponding table in the database corresponding to the image ID, tile number, and layer number. "Apple" is generated as a keyword.

In this way, keywords and image feature amounts as attribute information are extracted and generated from image data.

In step S606, the attribute information extracted and generated in step S605 and the registration data from which the attribute information has been extracted are associated with each other and registered in a database such as a hard disk or the like in the external storage unit 107 as shown in FIGS. 3, 4, and 5. I do. As described above, by creating a correspondence table between an image ID and an image file path as shown in FIG. 3 and creating a correspondence table between an image ID and attribute information as shown in FIG. 4, attribute information and registered data can be linked.

When registration in the database is completed in step S606, the process returns to step S602, and waits for an instruction from the input device.

When the CPU 101 determines in step S602 that the instruction from the input device is a search instruction, the process proceeds to step S607, and determines whether the search instruction is a search instruction that also includes temporary data. The determination as to whether or not the temporary data is to be searched can be made by determining the user's instruction in the dialog box shown in FIG.
Here, the temporary data is not searched, that is,
If the search instruction indicates that the data not registered in the normal database is not to be searched, the process proceeds to step S608, and the search is performed from the external storage device 107 or the database on the RAM 106. Although this search process will be described later, for example, if attribute information is stored in a database as shown in FIGS. 3 and 4, if a keyword to be searched is designated as “apple”, the image ID 1563,
1564 and 1565 are searched. Further, when a search instruction having the image feature “15” is received, there is no image corresponding to 15, and the image ID 1024 having the closest image feature “16” is selected. The search is performed in this way.

In step S607, if the CPU 101 determines that the temporary data is also a search instruction indicating that search is to be performed, the process proceeds to step S609 to determine whether to add new temporary data to the search target. Judge. When searching for temporary data for the first time after the power of this device, it is necessary to add data. For the second time or later, since a temporary database has already been created in the RAM 106 or the external storage unit 107, it is determined whether to search from the temporary database or to register new temporary data. Here, it may be based on the search instruction instructed in step S602, or may be newly instructed by the user from the input device 103. An instruction as to whether or not to make the temporary data a search target is made by a user selecting a dialog box using a user interface as shown in FIG.

If it is instructed to add new temporary data to the search target, the process proceeds to step S610 to select temporary data to be searched. As shown in FIG. 23, for example, an image file existing in the external storage device 107 such as a CD-ROM is selected.

At step S611, image data is read from the image file selected at step S610.

Next, in step S612, the CD-ROM is recorded in the same manner as in the normal database registration (S605).
For example, an image feature amount and a keyword as attribute information are extracted and generated from image data in an image file of the external storage device 107.

Subsequently, in step S613, the image data and the extracted and generated attribute information are associated with each other and registered in the memory as in step S606. Here, instead of registering in the database, a new temporary data database is created on the RAM 106, and attribute information and related image data are registered.

In step S614, a search is performed from the temporary data database on the memory. For example, it is determined whether the keyword “empty” input from the user by the input device 103 exists in the temporary data database,
If it exists, the linked image ID or the image data itself is picked up. When the image data to be searched is specified, image data having the same or similar image feature amount as the image data to be searched is picked up in ascending order, and is set as a search result.

Subsequently, in step S615, a search process, which will be described later, is performed from an ordinary database on the external storage device 107 or the RAM 106 in the same manner as in step S614.

In step S616, the search results searched in steps S614 and S615 are combined. For example, the search results may be combined in the order of closest search results, or the search from the database and the search from the temporary data database on the memory may be separately divided and combined.

In step S617, the display control unit 108 controls the search results compiled in step S616, and
To display. After the display, the process returns to step S602.

If it is determined in step S609 that temporary data is not to be newly added, the registration process is skipped, the process proceeds to step S614, and the process is performed as described above.

In step S603, if the CPU 101 determines that the registration instruction is a registration from a temporary temporary database existing in the memory, the process proceeds to step S618, and the process proceeds to step S613, where the attribute information is registered. The general data database is registered in a normal external storage device 107 or a database on the RAM 106. Since the format of each database is the same, it is only necessary to combine the databases. Step S618
When the database registration (synthesis) process ends, the process returns to step S602, and enters an instruction waiting state.

FIG. 24 is a flowchart of a specific search processing operation of the data search device of the present invention. The search processing operation will be described in detail with reference to FIG.

First, in step S2401, a search word to be searched is input. This search word may be a keyword or an image feature amount. For example, "apple" may be specified as a keyword, or "red" as an image feature word.
May be entered. It is also possible to specify "red apple". Further, in the case of an image feature amount, when a certain image is designated and an image close to that image is searched, a search word can be obtained by extracting the image feature amount of the image to be searched.

Next, in step S2402, word classification is performed. This word type is used to determine the type of the input search word. If there is a word representing a color stored in the word dictionary DIC stored in the external storage unit 107, for example, “red” or “red”, it is an image feature word representing an image feature amount, and therefore, step S240
The process proceeds to step S4403. If the word is a word other than the color existing in the word dictionary DIC, the process proceeds to step S2403.

In step S2403, a keyword search is performed. Keyword search is performed in the external storage unit 107 or RAM
This is performed with reference to the data table of FIG. It is checked whether a keyword 1101 that matches the search word exists in the search table (data table), and if it exists, it is stored in the output buffer OBUF as a search result. If not, the process proceeds to step S2405 with the number of search tiles set to 0.

Step S2404 is processing when it is determined that the search word is a search target image feature word (“red”, “blue”, “black”, etc.), and the designated image feature amount is acquired. . The acquisition of the image feature amount refers to the data table of FIG. 5 stored in the external storage unit 107 and checks the color name 501 that matches the image feature word. If they match, the corresponding image feature quantity is extracted from 502 and stored in the search target image feature storage buffer FBUF temporarily stored in the RAM 106.
If no matching color name exists, an invalid value (for example, a negative number) is stored in the search target image feature storage buffer FBUF.

Subsequently, the search target specified by the search target image characteristic word is obtained by re-classifying the word. For example, the search target image feature word “red” is searched for a word that is modified, and a search target object “apple” is acquired. Retrieved search object (for example, "apple")
Is searched for a keyword that matches the search object in the same manner as in the step in step S2403, and the output buffer
The image ID and the tile number are stored in the OBUF as search results. Output buffer OBUF if keyword is found
The image ID and the tile number stored in the external storage unit 107 are extracted, and the image feature 404 corresponding to the image ID and the tile number is extracted with reference to the data table of FIG. It is checked whether or not the extracted image feature matches the set search target image feature storage buffer FBUF. In other words, when searching for “red apples”, search for “apples” by keyword search, and refer to the data table of FIG.
6) is obtained, and it is checked whether or not it matches the search target image feature (in the example, “red”).

If the image feature obtained by referring to the data table matches the search target image feature, the process directly proceeds to step S2405 with the output buffer OBUF. If the image feature does not match the specified image feature ("green apple"
) Deletes the unmatched tile number, image ID, and layer number from the output buffer OBUF, and also reduces the number of keywords.

In step S2405, image data is displayed based on the keyword search result and the image feature amount search result.

Referring to the output buffer OBUF output by any of the processes of steps S2403 and S2404, if the number of matching keywords is 0, a message indicating that there is no search result is displayed on the display 109. . If the number of keywords is not 0, a tile number and a layer number corresponding to the first image ID stored in the output buffer OBUF stored in the RAM 106 are displayed. An image ID and a tile number are designated, and image data corresponding to the tile number is sent from an image file on the external storage unit 107, and the display unit 109
To be displayed.

For example, when the keyword "yacht" is searched, the tile numbers 13, 14, 18, and 19 in FIG. 12 are searched, and only the four tiles are displayed on the display 109. The image of the “yacht” is one fifth of the area of the image of FIG. 12, and the time required for display is accordingly shortened. Further, when such an image is on a network having a low transmission speed, the time required for transferring the image data can be reduced as compared with the case where all images are transferred.

For example, if "red apple" is given as a search word, the keyword of the search object is "apple".
As a result, the image ID is 1563, the tile number is 7, 1
2, image feature 3, image ID 1564, tile number 13,
14, 18 and 19, the image feature is 15, the image ID is 1565, and the tile number is 6, 7, 11, 12 and the image feature is 15. Three images are obtained. Image ID with image feature 3 corresponding to "red"
Only the image data of 1563 is obtained as a search result, and only the image of FIG. 6 is displayed.

In the above description, only the tile of the image corresponding to the search is displayed. However, the present invention is not limited to this, and the entire image including the corresponding tile may be displayed.

FIG. 25 is a table in which the position designation word, which is one of the attribute information for designating the position of the search object in the image, is associated with the tile number, and reference numeral 2701 denotes the position designation word as a character code (for example, unicode). To store. In this embodiment, as the position designation words, "up", "upper left", "upper right", "middle",
Use 10 words: "center", "center left", "center right", "bottom", "bottom left", "bottom right". Reference numeral 2702 stores a tile number represented by a position designation word corresponding to the divided tile. For example, if "upper right" is specified as a position designation word, 3,
The tile numbers 4, 8, and 9 are specified as search targets. As described above, the position designation word (position designation information) indicates a position corresponding to the tile image.

Since the position designation word and the tile number are associated with each other, the following search can be performed.

In the search processing shown in FIG. 24, a case will be described in which a search term such as "white on top and blue on bottom" is processed. First, a method for inputting a search term will be described.

First, a position designation word is input. For example,
There are five options: upper, lower, right, left, and middle, and one of them is selected by a dialog box or the like.

Subsequently, an image feature amount is input. For example, selecting a location designation word will bring up a new dialog box where one of the colors can be selected. When "up" is selected in the input of the position designation word, "white", "blue", "red", "green", etc. are displayed as the image feature amounts,
Let the user select it again in a dialog box or the like.

When "up" and "white" are input by the user in this way, it is determined in a dialog box as shown in FIGS. 22 and 23 whether further conditions are to be added. When further conditions are to be added, the above processing may be repeated.

Next, the flow of the search process will be described.

At the time of retrieval, the tile number corresponding to the position designation word “upper” is extracted from the correspondence table of FIG. 25, and an image is added to the extracted tile number “0, 1, 2, 3, 4”. A search is made from the data table (search table) in FIG. 4 to determine whether an image feature amount “26” corresponding to the feature word “white” exists. 8, 9 and 10 are obtained as search results corresponding to the first search word.

Subsequently, the remaining conditions are searched from this image. That is, following the above-described processing, the tile number “15, 1” corresponding to the next position designation word “down”
6, 17, 18, and 19 ”are extracted from the correspondence table of FIG. 25, and any of the blue image feature amounts“ 9, 10 ”is included in the corresponding tile in the image searched by the temporary search. Is searched using FIG. Then, as shown in FIG.
The image of 563 is obtained as a search result, and the image of FIG.

When a search is performed together with the position designation word as a search result, the entire image is displayed as the search result.

Further, in the present embodiment, the search processing is performed on the image stored in the external storage unit 107. However, the image stored in the image server on the network is searched via the network interface 111, and the search result is used. The present invention can also be realized by downloading only a simple image. In this case, the process of displaying the search results in tile units is effective in that the load on the network can be reduced and the display can be performed at high speed.

As described above, the data search program (the processing in FIG. 6) of the data search apparatus of the present invention executes the search processing. This data search program is stored in ROM1
05. However, the present invention is not limited to this, and a storage medium such as a floppy disk or a CD-ROM or a network interface
The data retrieval program is input from the CPU, and the retrieval program is developed and stored in an external storage device 107 such as a RAM 106 or a hard disk. The present invention, of course, includes such a storage medium.

As described above, according to the present invention, it is possible to search for image data existing in an external storage device such as a CD-ROM without registering the image data in the database, and to store temporary data that is not desired to be registered in the database. You can now search smoothly. Furthermore, after the search is completed, the temporarily used temporary database can be registered in the normal database, so that usability is further improved. Also, if the process of registering in the normal database is not performed, attribute information such as keywords in the memory is discarded at the time of system termination, and this attribute information does not remain at the next start-up, so the data will be stored at the user's will. Can be performed.

[0110]

As described above, according to the present invention, the CD-ROM
It is possible to search for image data existing in an external storage device without registering it in a database, and it is possible to smoothly search for temporary data that the user does not want to register in the database.

Further, since the data of the temporarily used temporary database can be registered in the ordinary database after the search is completed, the usability is further improved.

Further, it is possible to select whether or not to register the data of the temporary database in the normal database according to the user's will. If the process of registering the data in the normal database is not performed, the memory is not stored when the system is terminated. The attribute information such as the above keyword is discarded, and this attribute information does not remain at the next start-up, so that it is possible to perform data processing according to the user's will.

Further, one image can be divided into a plurality of areas and attribute information can be given to each area, so that detailed search processing can be performed.

Further, by providing the position designation information as the attribute information, it is possible to obtain information as to which part of the image the target of the search result is.

Further, since a plurality of search conditions for attribute information can be selected, it is possible to narrow down and select.

Further, since only the image of the corresponding area can be displayed as a search result, the effect of reducing the display time can be obtained.

[Brief description of the drawings]

FIG. 1 is a flowchart illustrating the flow of a conventional search process.

FIG. 2 is a block diagram illustrating a configuration of a search device according to the present invention.

FIG. 3 is a table showing the correspondence between image IDs and image file paths.

FIG. 4 is a correspondence table diagram of image IDs and attribute information.

FIG. 5 is a correspondence diagram between color names and image features.

FIG. 6 is a flowchart illustrating the processing flow of a search program according to the present invention.

FIG. 7 is a memory map showing a format of an image file.

FIG. 8 is an example of image data.

FIG. 9 is an example of image data.

FIG. 10 shows an example of image data.

FIG. 11 is an example of image data.

FIG. 12 is an example in which image data is divided into tiles.

FIG. 13 shows the structure of a Flashpix image object.

FIG. 14 shows the structure of a flashpix image view object.

FIG. 15 shows attribute information used in FlashPix.

FIG. 16 is an image of an image file hierarchically having images of a plurality of resolutions.

FIG. 17 is an example in which an image is divided into a plurality of fixed areas.

FIG. 18 shows sub-image header information of FlashPix.

FIG. 19 is an example of an image that is hierarchized and divided into tiles.

FIG. 20 shows the structure of a table in which image IDs, tile numbers, keywords, and image feature amounts are associated with each other.

FIG. 21 is a dialog box for determining a user instruction in step S607.

FIG. 22 is a dialog box for determining a user instruction in step S609.

FIG. 23 is a user interface diagram for specifying a data position when selecting temporary data.

FIG. 24 is a flowchart illustrating the flow of processing of a search program according to the present invention.

FIG. 25 is a table structure in which position designation words and tile numbers are associated with each other.

[Explanation of symbols]

Reference Signs List 101 CPU 102 System bus 103 Input device 104 Scanner 105 ROM 106 RAM 107 External storage device (unit) 108 Display control unit 109 Display device 110 Character generator (CG) 111 Network interface (Net / I /
F)

Claims (33)

[Claims] A first storage unit for storing a first database including attribute information of registered data; and a first database stored in the first storage unit, wherein attribute information of temporary registered data is stored separately. A second storage unit for storing a second database including the second database. 2. The data search device according to claim 1, wherein the attribute information includes a keyword. 3. The data search device according to claim 1, wherein the attribute information includes an image feature amount. 4. The data search device according to claim 1, wherein the attribute information includes position designation information. 5. The data search device according to claim 1, further comprising a synthesizing unit that registers the second database in the first database. 6. The data search device according to claim 5, further comprising a selection unit for selecting whether or not to register the second database in the first database by the combining unit. 7. The data search device according to claim 1, further comprising a registration unit for registering attribute information of data to be registered in the first or second database. 8. The data search apparatus according to claim 1, further comprising: input means for inputting attribute information to be searched; and search means for searching from the first database or the second database. 9. The data search apparatus according to claim 8, further comprising display means for displaying a search result of said search means. 10. The data search apparatus according to claim 9, wherein said display means displays only an image of the searched area. 11. The method according to claim 7, wherein the search unit performs a search by narrowing down in stages based on a plurality of pieces of position designation information and image feature amounts included in the attribute information input by the input unit. Data retrieval device as described. 12. A first registration step of registering registration data attribute information in a first database stored in a first storage means, and temporary registration data attribute information is stored in a second storage means. A second registration step of registering in a second database. 13. The data search method according to claim 12, wherein the attribute information includes a keyword. 14. The data search method according to claim 12, wherein the attribute information includes an image feature amount. 15. The data search method according to claim 12, wherein the attribute information includes position designation information. 16. The data search method according to claim 12, further comprising a synthesizing step of registering said second database in said first database. 17. The data search method according to claim 16, further comprising a selecting step of selecting whether to register the second database in the first database in the synthesizing step. 18. The data search method according to claim 12, further comprising a registration step of registering attribute information of data to be registered in the first or second database. 19. The data search method according to claim 12, further comprising an input step of inputting attribute information to be searched, and a search step of performing a search from the first database or the second database. 20. The data search step according to claim 19, further comprising a display step of displaying a search result of the search step. 21. The data search method according to claim 20, wherein the display step displays only the image of the searched area. 22. The search method according to claim 19, wherein the search unit narrows down and searches stepwise based on a plurality of pieces of position designation information and image feature amounts included in the attribute information input by the input unit. Data retrieval device as described. 23. A first registration step of registering registration data attribute information in a first database stored in a first storage means, and temporary registration data attribute information is stored in a second storage means. A second registration step of registering in a second database, a storage medium storing a computer-readable program. 24. The storage medium according to claim 23, wherein the attribute information includes a keyword. 25. The storage medium according to claim 23, wherein the attribute information includes an image feature amount. 26. The storage medium according to claim 23, wherein the attribute information includes position designation information. 27. The storage medium according to claim 23, further comprising a combining step of registering said second database in said first database. 28. The computer-readable program according to claim 27, further comprising a selecting step of selecting whether to register the second database in the first database in the synthesizing step. Storage media. 29. The storage medium according to claim 23, further comprising a registration step of registering attribute information of data to be registered in said first or second database. 30. The computer-readable computer according to claim 23, further comprising: an input step of inputting attribute information to be searched; and a search step of performing a search from the first database or the second database. A storage medium on which a program is stored. 31. The storage medium according to claim 30, further comprising a display step of displaying a search result of said search step. 32. The storage medium according to claim 31, wherein the displaying step displays only the image of the searched area. 33. A search method according to claim 30, wherein said search means narrows down and searches stepwise based on a plurality of pieces of position designation information and image feature amounts included in the attribute information input by said input means. A storage medium storing the computer-readable program according to the above.