JP3325326B2 - Electronic filing equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electronic storage device for storing image information and the like in an external storage device such as an optical disk by adding a keyword and a text (title) indicating an outline, and searching for the image information by specifying a keyword. It relates to a filing device.

[0002]

2. Description of the Related Art In an electronic filing apparatus, when searching for target image data (image information), a keyword for search is inputted, and a keyword search or a search for selecting image data to which the keyword is added is performed. For example, there is a method such as a free text search for inputting a character string and selecting image data to which a title including the character string is added.

In an electronic filing apparatus, a keyword or a title added to each image data to be managed is recorded on the same optical disk as a recording destination of the image data.
Recorded on another hard disk. When performing the above search using the electronic filing device, the keywords and titles added to all the managed image data are read from the optical disk or the hard disk, and are compared with the input search keywords and search character strings to perform the search processing. It is carried out. Usually, a keyword number is assigned to a keyword added to image data, and the keyword number is recorded for each image data.

For example, if the total number of managed image data is 1000, and each image data is added with an average of 5 keywords, and the keyword number is expressed in 2 bytes, 10 bytes per image data is required. Become. Therefore, data of 10,000 bytes (10 KB) in total is read and compared with the search keyword.

If the character string length of the title added to the image data is 100 bytes, the total length is 1 byte.
The data of 00000 bytes (100 KB) is read and compared with the search character string.

[0006] The number of image data to be managed increases,
When the number of keywords added to image data increases or the length of the title character string added increases, the amount of data to be compared with the search keyword or search character string increases, and the search process takes time. was there.

The image information search device disclosed in Japanese Patent Application Laid-Open No. 4-86947 re-examines the classification information added to an already stored image.
It is not intended to improve search speed.

In order to solve this problem, Japanese Patent Laid-Open Publication No. Hei 4-90054 discloses a method in which keywords are divided into several groups in numerical order to improve search speed by using group keywords having a smaller number of bits than search information. When grouping, only the keyword number is used, so the primary search result using the group keyword can be used only to narrow down the number of images to be secondary searched.

[0009]

As described above, conventionally, the number of image data to be managed is increased, the number of keywords added to the image data is increased, or the length of the title character string added is reduced. If the length becomes long, the amount of data to be compared with the search keyword or the search character string becomes enormous, and there is a problem that the search process takes time. When keywords are grouped for the purpose of improving search speed, a primary search result using a group keyword using only a keyword number can be used only to narrow down the number of images to be secondary searched. There was a problem. Accordingly, it is an object of the present invention to provide an electronic filing apparatus capable of speeding up search processing and performing efficient image information search.

[0010]

An electronic filing apparatus according to the present invention comprises: first storage means for storing search information including a plurality of keywords for searching for each image corresponding to each image; Similarity calculating means for calculating the similarity between keywords of the search information stored in the storage means,
Group dividing means for grouping keywords having high similarity calculated by the similarity calculating means into the same group, generating means for generating a group keyword for each group grouped by the group dividing means,
A second storage unit that stores a group keyword generated by using the generation unit from the plurality of keywords of the search information stored in the first storage unit in correspondence with each image; And a group keyword generated from the keyword input by the input means using the generation means, and
A first extraction unit for extracting an image including a keyword in units of groups by sequentially comparing the group keywords stored in the storage unit of (a), and an image extracted by the first extraction unit Correspondingly the first
A second extraction unit that sequentially compares the keywords stored in the storage unit with the keyword input by the input unit to extract an image that matches the keyword; Images,
Or a selecting means for selecting a desired image from the images extracted by the first extracting means.

[0011]

According to the present invention, search information including a plurality of keywords for searching for each image corresponding to each image is stored in the first storage means, and the similarity between the stored search information keywords is calculated. Grouping the calculated keywords having a high degree of similarity into the same group, generating a group keyword for each of the grouped groups by the generation unit, and generating a plurality of search keywords stored in the first storage unit. A group keyword generated by using the generating unit from the keyword is stored in the second storage unit corresponding to each image, a keyword for search is input, and a keyword is generated from the input keyword by using the generating unit. By sequentially comparing each group keyword stored in the second storage means with the group keyword to be set, a key for each group is obtained. Extracting an image including the keyword by the first extracting means, and sequentially comparing the keyword stored in the first storage means with the keyword input by the input means in correspondence with the extracted image. , An image having the same keyword is extracted by the second extracting means, and a desired image is selected from the images extracted by the second extracting means or the images extracted by the first extracting means. It is something to do.

[0012]

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

FIG. 1 shows an example of the configuration of an electronic filing apparatus according to the present invention. That is, the electronic filing device includes the control module 10 and the memory module 1
2, image processing module 14, communication control module 1
6, a scanner device 18, an optical disk 19 and an optical disk device 20, a keyboard 23, a CRT display device 24, a printer device 25, a magnetic disk 26 and a magnetic disk device 27, a mouse 29, a system bus 30, and an image bus 32. .

The control module 10 includes a CPU 34 for performing various controls for storing, retrieving and editing images, and the like.
And optical disk device 20, magnetic disk device 27 and C
It comprises an interface circuit 36 for connecting the PU 34. The keyboard 23 and the mouse 29 are connected to the CPU 34.

The memory module 12 includes a main memory 38 for storing various control programs for image storage, search, editing, etc. and management information, a character recognition unit 39 for recognizing characters of the image, and a document for several pages of A4 size document. A page memory 40 as an image memory having a storage capacity corresponding to an image,
It comprises a display memory 42 as a display interface, a display control unit 44 and the like. A buffer memory area 40a is provided in a part of the page memory 40. The page memory 40 is, for example, the optical disk 19
Is a memory for temporarily storing an image stored in the optical disk 19 and an image read from the optical disk 19.

The display memory 42 is for temporarily storing an image displayed by the CRT display device 24. The display memory 42 enlarges, reduces, rotates, inserts or inserts an image actually displayed in a display window of the CRT display device 24, for example, an image from the page memory 40 or an image from the buffer memory 40a. An image subjected to black-and-white inversion is stored. The display controller 44 controls the display processing of the CRT display device 24 and the like.

The image processing module 14 includes an enlargement / reduction circuit 46 for performing image enlargement / reduction processing, a vertical / horizontal conversion circuit 48 for performing image rotation processing, and an encoding processing for compressing an image (reducing redundancy). A compression / decompression circuit (CODEC) 50 for performing a decoding process for performing and decompressing (reducing the reduced redundancy), a scanner interface 52 for the scanner device 18, a printer interface 54 for the printer device 25, and scaling. An internal bus 56 connects the circuit 46 and the vertical / horizontal conversion circuit 48 to the compression / expansion circuit 50, the scanner interface 52, and the printer interface 54. The compression / expansion circuit 50 uses the MH (Modified Huffman) method or the MR (Modified Read) method to perform band compression,
Alternatively, band extension is performed.

The communication control module 16 is constituted by a communication interface 58 such as a BCP (Bus Communication Processor) connected to a local area network (Local Area Network). Further, the communication control module 16 has a universal communication process (UCP) connected to an external device such as an FCP (facsimile connection mechanism) or a personal computer via an interface.
r) may be provided.

The system bus 30 is a bus for control signals of various devices, and connects the control module 10 with the memory module 12, the image processing module 14, and the communication control module 16. The image bus 32 is an image bus, and connects the memory module 12, the image processing module 14, and the communication control module 16.

The scanner device 18 is, for example, a two-dimensional scanning device that obtains an electric signal corresponding to an image on the original by scanning the original (document) two-dimensionally with a laser beam.

The optical disk device 20 includes a scanner device 18
Are sequentially stored on the optical disk 19. Further, an image corresponding to the search code specified by the keyboard 23 or the like is searched from the optical disc 19.

The keyboard 23 is used to input a unique search code corresponding to an image stored in the optical disk 19 and various operation commands such as storage, search, and edit processing. The mouse 29 moves the cursor (not shown) displayed on the display window of the CRT display device 24 arbitrarily in up, down, left, and right directions, and gives an instruction at a desired position. The display contents (various operation modes, titles, designation of areas for image editing, icons, etc.) are selected or indicated.

The CRT display (cathode ray tube display) 24 displays an image read by the scanner 18 and an image retrieved from the optical disk 19. Icons and the like are displayed at the upper, lower, and right ends of the display window of the CRT display device 24.

The printer 25 prints out (hard copies) an image read by the scanner 18, an image retrieved from the optical disk 19, or an image displayed on the CRT display 24.

The magnetic disk device 27 stores various control programs on a magnetic disk 26 mounted on the magnetic disk device 27, and also stores a search code input from the keyboard 23 and an image corresponding to the search code. It stores search data (search information) including a storage address on the optical disk 19, an image size, a search frequency, and the like.

In this embodiment, the image data (image information) recorded on the optical disk 19 includes an image management table 1 shown in FIG. 2, an image management table 2 shown in FIG. 3, and a keyword correspondence table 3 shown in FIG. , And a keyword correspondence table 4 shown in FIG. These tables are stored in the magnetic disk 2 for each optical disk 19 to be managed.
6 is recorded.

The image management table 1 stores the name of the optical disk 19 on which image data is recorded and the total number of recorded image data. The name of the optical disk 19 shown in FIG. 2 is “1990 patent data”, and the total number of images is 100.

The image management table 2 shown in FIG. 3 includes, for each image data (image information), an image No. (image number), an image size, and a head position on the optical disk 19 which indicates where the image is recorded on the optical disk 19. An address, a title added to the image data, ten keyword numbers, and a group keyword are stored.

The title of the image management table 2 is SJI
Expressed in S code, a 100-byte area is reserved, and in the case of 2-byte kanji code, up to 50 digits, ANK
In the case of a one-byte code, a maximum of 100 digits can be stored.

The keyword No. of the image management table 2 is
It is represented by 2 bytes (16 bits), and if "0" is set, it means that no keyword is set. 65535 at maximum (2 to the 16th power -1)
It is possible to set different types of keywords.

The group keyword in the image management table 2 is the keyword No. of the keyword added to the image.
From the group keyword calculation process described later. The group keyword is composed of 2 bytes (16 bits) and indicates which group keyword is added to this image data.

For example, as shown in FIG. The image data of No. 1 has an image size of A4, stores image data starting from address 0 on the optical disk, has a SJIS character string of “electronic filing device” added to the title, and has a keyword of Keyword No. . Those indicated by 1 and 2 are added, and the group keyword is “0”.
x0003 "(hexadecimal). The keyword correspondence table 3 shown in FIG. 4 stores the total number of types of keywords added to image data.

The keyword correspondence table 4 shown in FIG.
The keyword No., the keyword name corresponding to the keyword No., and the group No. to which the keyword belongs are stored for the total number of types of the keyword No. 4 and 5, the total number of types of keywords is 50.
The keyword name of “1” is “optical disk” and the group number is “0”.

FIG. 6 shows the structure of a group keyword. In this embodiment, the keywords are classified into 16 groups, the group keywords have a length of 16 bits (2 bytes), and the group numbers are changed from “0” to “1”.
The group No. is made to correspond to the bit position up to “5”. It should be noted that the present invention is not limited to 16 groups, and the point is that a group keyword may be configured. Next, a process for newly registering image data will be described with reference to a flowchart shown in FIG.

First, a document of a document to be registered (document) is set on the scanner device 18, and a title and some keywords are input from the keyboard 23 to instruct registration. The CPU 34 performs the following processing according to the program loaded in the main memory 38.

The image data obtained by reading the original by the scanner device 18 is written in an empty area of the optical disk 19 of the optical disk device 20. Next, “1” is added to the total number of images in the image management table 1, one line is added to the image management table 2, a new image No. is assigned, the image size written on the optical disc 19 and the head address written Is set. The title input by the operator is stored in the image management table 2
To the keyword N from the keyword name input by the operator using the keyword correspondence tables 3 and 4.
The value o is set in the image management table 2. Next, if the keyword name input by the operator is new, the CPU 34 performs the following processing.

"1" is added to the total number of keywords in the keyword correspondence table 3, one line is added to the keyword correspondence table 4, a new keyword No. is assigned, and a keyword name is set. Further, the similarity between keywords is calculated, and the keyword is divided into groups to determine which group each keyword belongs to (group division processing). Further, a group keyword is generated from the keyword No. set in the image management table 2 (group keyword generation processing described later), and set in the image management table 2. Next, C
The group division processing when a new keyword is added by the PU 34 will be described with reference to the flowchart in FIG.

First, using the image management table 2 shown in FIG. 3, the number of co-occurrence frequencies for all combinations of two keywords is determined, and the maximum co-occurrence frequency is set to 100 to normalize the frequencies. The co-occurrence frequency is the number of images to which two keywords are both added.

FIG. 9 is a normalized co-occurrence frequency table 5 in which the co-occurrence frequency numbers for all combinations of two keywords are normalized. In FIG. The normalized co-occurrence frequency number of 1 and 3 is 70.

Next, a portion exceeding a certain value (for example, 50) is selected from the values of the first row (one having a high similarity to the keyword 1) of the normalized co-occurrence frequency table 5, and these are designated as the keyword 1 and The keyword correspondence table 4 is updated as the same group. In the example shown in FIG. 9, keyword 1 and keyword 3 are set as the same group. The above processing is performed for each row, and the keyword correspondence table 4 is updated. At this time, if the number of groups exceeds 15, all remaining keywords are set to group 15.

The present invention is not limited to the method of calculating the similarity between keywords, but may use a synonym dictionary or a method of checking whether or not the same word is included, and calculating. Next, the group keyword generation processing by the CPU 34 will be described with reference to the flowchart in FIG.

First, all bits of the group keyword are turned off, a group number in the keyword correspondence table corresponding to each input keyword number is obtained, and bits on the group keyword corresponding to the obtained group number are turned on. .

For example, keywords No. 1 and No. 2 are added to the image data of image No. 1 in the image management table 2 shown in FIG. 2, and the keyword correspondence table 4 shown in FIG.
Since the group numbers are “0” and “1”, respectively, the group keyword 0x0003 (hexadecimal) determined by the group keyword generation processing is set.

Also, keywords Nos. 3 and 4 are added to the image data of the image No. 3 and the group Nos. Are “0” and “1” from the keyword correspondence table 4, respectively. , The same group keyword 0x as the image No1
0003 is set. Next, processing at the time of retrieval in the present embodiment will be described with reference to the flowchart in FIG.

The operator inputs some keywords for search and instructs keyword search. CPU34
Performs the following processing according to the program loaded in the main memory 38.

First, a keyword No. is determined from several keyword names for search input by the operator using the keyword correspondence table 4, and a group keyword for search is determined from these keyword Nos. This processing is similar to the above-described group keyword generation processing.

The logical product of the search group keyword obtained here and the group keyword for each image data in the image management table 2 is obtained, and all the image data having the same value as the search group keyword are obtained. Image N
o is temporarily stored. (First search)

Next, the keyword No. in the image management table 2 is read from the temporarily stored image No., and the image data including all the ones that match the above-mentioned search keyword No. is selected. (Secondary search) After that, according to the instruction of the operator, the primary or secondary search is performed.
The image searched by the next search processing is displayed on the CRT display device 24.

For example, if the operator inputs "optical disk" and "scanner" as search keywords, the images of images No. 1 and 3 are searched in the first search, and the images are further narrowed down in the second search. Only the image No. 1 is searched. In this case, "optical disk" and "optical disk device", "scanner" and "scanner device"
Are keyword names belonging to the same group. Next, the CRT at the time of search processing in the present embodiment
The screen display of the display device 24 will be described.

FIG. 12 shows an initial screen at the time of search processing displayed on the CRT display device 24. The left side of the initial screen is composed of a search condition setting area 61 and display areas 62 and 63 corresponding to the number of search results.
A keyword set from the keyboard 23 is displayed in the search condition setting area 61, and a first
The number of hits in the next search result,
The number of hits in the next search result is displayed. In the example shown in FIG. 12, five keywords can be set, and the area 62a in which the number of the results of the keyword search between the items and
63a.

The right side of the initial screen is an area 64 where a document image is displayed. The lower part of the initial screen is a function key corresponding to the function key of the keyboard 23. The F1 key 65, the F2 key 66, the F4 key 6
7, F9 key 68. For example, when the F1 key 65 is pressed, the primary search process is executed from the keyword set as the search condition, and the result is displayed in the number-of-matches display area 62. When the F2 key 66 is pressed, the first
A second search process is performed on the next search result, and the result is displayed in the number-of-matches display area 63. When the F4 key 67 is pressed, the screen is switched to the search result screen shown in FIG. 14 or FIG. If the F9 key 68 is pressed, the search processing ends.

FIG. 13 shows a display screen of search conditions after the search keyword has been set by the operator and the primary and secondary search processes have been executed. “Optical disk” is set as the first keyword, and “10” items are displayed as the result of the primary search using the group keyword No. to which the keyword of “optical disk” obtained using the keyword correspondence table 4 belongs. Second
Similarly, for the keyword “scanner”, “8” items are displayed as the primary search result.

When the keyword of "optical disk" is set, a document to which the keyword of "optical disk device" belonging to the same group is added is also extracted as the corresponding document in the first search. Similarly, for the “scanner”, a document to which the keyword “scanner device” is added is also extracted. The number of documents to which the group keywords corresponding to “optical disk” and “scanner” are both added is displayed as “4” in the area 62 a of the inter-item and result. Set the number of these keywords for each keyword and press the F1 key 6
After input of 5, it is processed and displayed.

Next, when the F2 key 66 is pressed, the number of results is displayed in the secondary number of hits display area 63. In this example, the number of documents to which the keyword of “optical disk” is added is “5”, and the number of documents to which the keyword of “scanner” is added is “3”. Then, the number of documents to which both the keywords “optical disk” and “scanner” are added is “2” in the area 63a of the inter-item and result.
Is displayed.

FIG. 14 shows a display screen of the primary search result. In the left area 71, the number, title, and keyword of each document image, which are the primary search results, are displayed. An area 72 on the right side is a document image display area, and a document image corresponding to the hatched “No. 1” of the area 71 is displayed.

On the display screen of the primary search result, F1
Key 73 is a key for switching between the primary and secondary search result screens, F2 key 74 and F3 key 75 are keys for selecting a document image to be displayed, F4 key 76 is a key for returning to the search condition screen, and F5 key. 77 is a key for printing the displayed document image, and F9 key 78 is a key for terminating the search process.

The display screen of the primary search result has the four document image information extracted in the primary search, that is, the keyword of “optical disk” or “optical disk device”, and “scanner” or “optical disk device”. A document image with the keyword "scanner device" is displayed.

FIG. 15 shows a display screen of the secondary search result. Here, two pieces of document image information of only those having the keywords “optical disk” and “scanner” are displayed from the document images extracted by the primary search.

As described above, the first keyword based on the group keyword is used.
By displaying the next search result and the second search result using the keyword, the search result under the loose condition can be compared with the search result under the hard condition with only the designated keyword, and the document image intended by the operator can be compared. It is possible to search more efficiently.

For example, if the target document image is not included in the secondary search result, the target document image can be obtained by referring to the primary search result. It is also possible to set more accurate search conditions using a keyword added to the document image extracted as a result.

As described above, the similarity between the keywords is obtained, the keywords having a high similarity are classified into the same group, the group corresponding to the individual keyword added when the image is registered is obtained, and the group corresponding to the group is obtained. A value in which the position bit is turned on and the other bits are turned off is created, and this value is stored as a group keyword of the image as a pair with the registered image.

When performing a search by specifying a keyword, a search group keyword in which the bit corresponding to the group is turned on is obtained from each keyword specified for the search.
The group keyword of the image created at the time of storing the image is compared with the group keyword for search, and a candidate image to be searched is extracted in units of a plurality of grouped keywords (first search process).

Further, a conventional search (secondary search process) is performed on the extracted candidate image by using the keyword itself as a unit, and a desired image is extracted from the candidate images. It was done. That is, when the group keyword is expressed by 2 bytes (16 bits) as described in the above embodiment, the keywords are classified into 16 groups.

For example, assuming that the total number of images is 1000 and each image data has an average of 5 keywords, conventionally, 1000 images × 5 × 2 bytes of 1000 images
0 bytes (about 10 KB) of data had to be read and compared.

However, when the present invention is applied, first, 2000 bytes (about 2 KB) of 1000 × 2 bytes
, And performs a primary search process using a group keyword, and performs a secondary keyword search only on the image data selected in this process.

Here, how much image data is selected in the primary search processing differs depending on how keywords of recorded image data are attached and in which group the keywords are classified. However, in the above example, when 80% or less of the entire image data is selected in the primary search, the total amount of comparison data in the primary and secondary is smaller than in the conventional keyword search.

For example, if 10% of the image data is selected in the primary search, 2000 bytes (primary search) +
100 items x 5 items x 2 bytes (secondary search) total 300
This means that 0-byte data is compared, and the amount of comparison data is reduced to about 30% as compared with the related art.

As described above, by performing a rough search process using a group keyword, the number of image data for performing a normal keyword search can be greatly reduced.
Keyword search can be speeded up.

Further, since keywords having a high degree of similarity are classified as the same group, the image data obtained in the primary search processing is one to which keywords having a high relation with the keyword input by the operator are added. By properly using the primary and secondary search results, desired image data can be efficiently extracted.

As described above, according to the above embodiment,
By performing a rough search process using a group keyword, the number of image data for performing a normal keyword search can be significantly reduced, and the keyword search can be speeded up.

Further, since keywords having a high degree of similarity are classified as the same group, the image data obtained in the primary search processing is one to which keywords having a high relation with the keyword input by the operator are added. Desired image data can be efficiently extracted by properly using the primary and secondary search results.

[0072]

As described in detail above, according to the present invention,
It is possible to provide an electronic filing apparatus capable of performing high-speed search processing and performing efficient image information search.

[Brief description of the drawings]

FIG. 1 is a block diagram schematically showing a configuration of an electronic filing apparatus according to one embodiment of the present invention.

FIG. 2 is a diagram for explaining an image management table.

FIG. 3 is a diagram for explaining an image management table.

FIG. 4 is a diagram for explaining a keyword correspondence table.

FIG. 5 is a diagram for explaining a keyword correspondence table.

FIG. 6 is a diagram for explaining the structure of a group keyword.

FIG. 7 is a flowchart illustrating an operation of registering image data.

FIG. 8 is a flowchart for explaining the operation of a group division process.

FIG. 9 is a diagram for explaining a normalized co-occurrence frequency table.

FIG. 10 is a flowchart for explaining the operation of a group keyword generation process.

FIG. 11 is a flowchart for explaining a processing operation at the time of search.

FIG. 12 is a view for explaining an initial screen at the time of search processing.

FIG. 13 is a diagram for explaining a search condition display screen after a search process has been executed.

FIG. 14 is a view for explaining a display screen of a primary search result.

FIG. 15 is a view for explaining a display screen of a secondary search result.

[Explanation of symbols]

 1, 2, image management table 3, 4, keyword correspondence table 5, normalized co-appearance frequency table 18, scanner device 19, optical disk 23, keyboard 24, CRT display device 26, magnetic disk 34, CPU 38, main memory

────────────────────────────────────────────────── ─── Continued on the front page (58) Fields surveyed (Int. Cl. ⁷ , DB name) G06F 17/30 210 G06F 17/30 170 G06F 17/30 320 JICST file (JOIS)

Claims (1)

(57) [Claims] A first storage unit that stores search information including a plurality of keywords for searching for each image corresponding to each image;
Storage means, similarity calculation means for calculating the similarity between the keywords of the search information stored in the first storage means, and the keywords having a high similarity calculated by the similarity calculation means in the same group. Group dividing means for grouping; generating means for generating a group keyword for each group grouped by the group dividing means; and a plurality of keywords of search information stored in the first storage means, A second storage unit that stores a group keyword generated using the generation unit corresponding to each image; an input unit that inputs a keyword for search; and the generation unit based on the keyword input by the input unit. Is used to sequentially compare each group keyword stored in the second storage means. A first extraction unit for extracting an image including a keyword in units of groups, and a keyword stored in the first storage unit corresponding to the image extracted by the first extraction unit. A second method of sequentially comparing the keywords input by the input means to extract images matching the keywords;
Extraction means, and selection means for selecting a desired image from the images extracted by the second extraction means or the images extracted by the first extraction means. Electronic filing equipment