JPS6043776A - Picture memory retrieving device - Google Patents

Abstract

PURPOSE:To obtain a picture memory retrieving device which is available to various types of documents by selecting each mode in response to the performance of characters, etc. to be processed. CONSTITUTION:It is decided whether a flag F1 is set at ''1'' or not. If the flag F1 is set at ''1'', the key word input is accepted from an input part 7 as long as the 1st mode is designated from the part 7. An operator stores successively the picture information while feeding a key word for each sheet of originals. While it is decided whether a flag F2 is set at ''1'' or not, in case the flag F1 is set at ''0''. Then the 2nd mode state is decided if the F2 is set at ''1''. In the 2nd mode the operator designates a key word area every time a sheet of original is read out. Then both storage of the picture information and registration of the key word are carried out successively while recognizing the key word through the key word area. Hereafter the 3rd, 4th- mode states can be designated by flags F3, F4- respectively.

Description

[Detailed Description of the Invention] [Technical Field] The present invention relates to an image storage and retrieval device that stores image information such as a document on an information storage medium such as an optical disk or a magnetic disk, and searches for image information on the information storage medium. Regarding.

[Prior art]

In such an image storage and retrieval device, when newly storing image information such as a document, it is necessary to register a search keyword for the document or the like. However, in the past, since the operator had to input search keywords one by one using a keyboard or the like, it was not possible to automate the storage process of image information such as a single document, which was inefficient. In addition, the burden on the operator was heavy, and keyword errors were likely to occur due to incorrect operations.

On the other hand, Japanese Patent Application Laid-Open No. 56-22225 discloses a method for reading page numbers of documents, etc., storing image information of documents, etc. on an optical disk in order of page numbers, and associating storage addresses of image information with the page numbers. An apparatus for storing information in a memory is disclosed. This device can search for image information such as a corresponding document by specifying a page number as a keyword. However, page numbers often overlap, and there is no direct relationship between the content of a document or the like and the page number, and there are also documents that do not have page numbers. Therefore, this device is not suitable for any purpose other than storing image information such as a series of documents with consecutive page numbers organized in order of page numbers.
It does not have versatility as an image storage and retrieval device.

〔the purpose〕

SUMMARY OF THE INVENTION An object of the present invention is to provide a highly versatile image storage and retrieval device that can eliminate the above-mentioned inconveniences related to keywords and efficiently store image information such as documents.

〔Example〕

FIG. 1 is a block diagram showing one embodiment of the present invention. In the figure, 1 is a reading unit that reads image information on documents, etc., and 2
3 is a page buffer memory for temporarily storing image information for one page, and 3 is an output unit having a CRT display device. 4 is a character recognition unit that performs character recognition to obtain a search keyword, and 5 is an address memory for storing image information storage addresses and search keywords in association with each other. Reference numeral 6 denotes a control unit for controlling the operations of each part of the apparatus, 7 an input unit such as a keyboard, and 8 a large-capacity image storage unit using an information storage medium such as a magnetic disk or an optical disk.

FIG. 2 is a schematic cross-sectional view of the reading section 1. FIG. The reading unit 1 includes a pair of rollers 14° 15, 16 for conveying originals such as documents.
, a tray 18, 19 from which the original is ejected after reading, a switching claw 17 for distributing the original to the tray 18, 19, a contact glass 20, a light source 21 for illuminating the original, and a contact lens 20. A lens 22 for forming an image on the image sensor 23, a sensor 25 for detecting the leading edge of the original, a sensor 24 for detecting the trailing edge of the original, and a sensor 24 for detecting the trailing edge of the original. It has a mechanism for taking out the sheets one by one and feeding them into the pair of rollers 14, a circuit for converting the output signal of the image sensor 23 into binary image information, and the like.

This device can specify five modes as storage processing modes for image information. Mode designation is performed by operating a mode key on the input unit 7.

Next, the image information storage device will be explained.

When new image information becomes unprocessable, the control unit 6 issues a reading command to the reading unit l.

The reading unit 1 rotates the roller pairs 14 to 16 to feed one document from the stacker to the roller pair 14 and convey it. When the leading edge of the document crosses the sensor 25 and a detection signal is output from the sensor, the control section 6 starts writing the image information output from the reading section l into the page buffer memory 2, and also transfers the image information to the output section 3. Display on a CRT display device. Thereafter, when the sensor 24 detects the trailing edge of the document and outputs a detection signal, the reading section 1 stops the rotation of the roller pairs 14 to 16 and causes the switching claw 17 to stop the document.

When the control section 6 receives the detection signal from the sensor 15, it stops writing to the page buffer memory 2 and checks the state of the mode key of the input section 7. That is, as shown in the flowchart of FIG. 3, the first to fifth mode keys are examined in order until one in the "on" state is found (■ to ■). The control unit 6 sets a flag F1 corresponding to each mode.
~F5, set the flag corresponding to the mode key that first detected ``on'' to P1'' (■ ~
[phase]).

Note that this determination may be made while reading the image information.

After reading the image information and completing the determination, the control section 6 starts processing according to flowchart 1 shown in FIG.

First, it is determined whether the flag F1 is //Ipg (0).

If YES, the first mode is designated in the input unit 7, and therefore the keyword input from the input unit 7 is accepted (■). When confirming completion of keyword input (0)
, At the same time as issuing a normal ejection command to the reading section 1, the transfer of image information from the page buffer memory 2 to the image storage section 8 is started (0). When the image information for one page has been stored in the image storage section 8, search data that is a combination of the storage address of the image information and the keyword already inputted from the input section 7 is stored in the address memory 5. ([phase]).

This completes the first mode of processing, and the control section 6 issues a command to read the next document to the reading section 1.

In other words, the first mode is a mode in which the operator inputs keywords for each document and sequentially stores image information. In addition, if the above normal discharge command is issued, the reading section 1
The roller pairs 14 to 16 are rotated while the switching claw 17 is held at the angle shown in FIG. 2, and the document that has been stopped after reading is discharged onto the tray 18.

If the determination result in step 0 is No, the control unit 6 determines whether the flag F2 is set to "1" for the cell 1 ([phase]
). If YES, the second mode is specified, and the keyword area specification from the input section 7 is accepted (0).
. This will be explained with reference to FIG.

FIG. 5 shows the screen of a CRT display device.

10 is an image of the read document. The operator operates the pointer 29 on the screen from the input section 7, and selects the document image 1.
A pointer is positioned at the corner of the area of characters (including symbols and numbers) desired to be used as a search keyword for 0, and the controller 6 is instructed to detect the pointer address at that time. For example, if you want to use the character string t [i'57-10A, Il shown in Figure 6 as a keyword, use the address (a, , b, ) (a2. b, ) of the corner of each character area.
-(a2.b, ) is detected by the control unit 6. Note 1
If the character recognition unit 4 or control unit 6 is provided with a character extraction function, only the addresses of two corners of the entire character string area, that is, (a,,b,) and (a2.b,) in FIG. It is sufficient to specify.

When the area of each character used as a keyword is specified in this way, the control unit 6 controls 1. □tzR-
';Ay 77-1,291.・111 Input to character recognition unit 4 and perform character recognition (0).

When all characters of the keyword can be recognized ([phase]), the control section 6 displays the recognition result on the screen of the C'RT display device of the output section 3 ([phase]). In the example in Figure 5, the character strings 1i'57-10A" and "1i'57-10A" used as keywords are used as keywords.
The recognition results of 5-9-IJ are displayed as 11 and 12, respectively.

The operator checks the character recognition results displayed on the screen, and if there are no errors, turns on the OK key on the input unit 7.

If there is an error, the operator turns on the NG key on the input section 7.

The control unit 6 should check the status of the OK key and NG key■
, ■) If the OK key is detected to be on, issue a normal ejection command to the reading unit 1, and then store the image information in the image storage unit 8.
(■), and search data, which is a combination of the keyword recognized by the character recognition unit 4 and the storage address of the image information, is stored in the address memory 5 (■). NG key”
If an error is detected, an abnormal discharge command is issued to the reading unit 1, and the process jumps to step [phase].If an abnormal discharge command is issued, the reading unit 1 rotates the switching claw 17 upward. Rotate roller pairs 14 to 16 from
- Discharge to Ray 19.

In other words, the second mode is a mode in which the operator specifies a keyword area each time a document is read, and sequentially stores image information and registers the keyword while recognizing the keyword from that area.

Note that if the keyword cannot be recognized, an abnormal discharge command is issued to the reading section 1, and the process is terminated ([phase]).

This search data will be explained. Each character of the keyword is expressed, for example, in JISn code (1 byte), and the storage address of image information is expressed, for example, in 2 bytes. Assuming that the maximum number of characters of a keyword is n, the search data is sequentially stored in the address memory 5 with (n4-2) bytes as one block. If you also want to use kanji as keywords, for example J
Use TS Kanji code (2 bytes).

Although not mentioned in the above explanation, before registering a keyword, the control unit 6 searches the address memory 5 to see if the same keyword has been registered. If the same keyword does not exist, a new block is created and search data including the power keyword is stored there. If the same keyword is found, expand that block and
A new image information storage address is added to that block. In this case, a specific code that is neither used as a keyword character code nor an address code is inserted at the boundary between the keyword character code and the address, so that the keyword and address can be easily distinguished during a search.

This will be explained with reference to FIG. 4 again. If the determination result in step [stock] is No, the control unit 6 determines whether the flag F3 is "1" (■). If YES, the third mode is specified, so the third mode processing is performed. We (■〜＠、■、■
). This mode differs from the second mode only in that when a keyword is recognized, image information is stored and search data is registered directly without the operator's judgment.

If the determination result in step [phase] is NO, it is determined whether flag F4 is 1114g ([phase]). If YES, the fourth mode has been designated, so the process of that mode is executed. This mode differs from the second mode in that the keyword area designation from the input unit 7 is not accepted again, but the keyword character recognition is performed for the previously specified area. Steps [phase] to ■ are the same as steps [phase] to @ of the second mode.

In other words, the fourth mode is a mode in which image information is recorded and search data is registered while performing keyword character recognition for previously specified keyword areas without the operator specifying a keyword area for each document. It is.

If the determination result in step [phase] is NO, it is determined whether the flag F5 is rr 1 n (■). This judgment result is YES
In the case of A, the fifth mode is specified. In this mode, the same processing as in the fourth mode is performed (■ to @). However, after step ■, the process proceeds to step ■ without relying on the operator's judgment.

That is, this mode is a mode in which the storage of image information of a series of originals and the registration of search data are automatically performed without operator intervention.

If the determination result in step (2) is NO, an error interrupt occurs because the mode specification is incorrect, and the control unit 6 displays an error message to the operator and interrupts the process.

Here, a supplementary explanation will be given regarding character recognition of keywords. In steps [phase], ■, [phase], ■, if there is an unrecognized character in the specified keyword area, an area where the character area is moved up or down by a distance y,
The character recognition unit 4 receives image information of an area that has been moved by a distance X to the left or right, or an area that has been moved up, down, left and right at the same time.
input in sequence and have character recognition performed. For example, if the recognition of the characters 1i'5'' shown in Figure 6 fails, the following ■~■
Character recognition is attempted for the area with the corner address until recognition is successful.

■(ar+bl X)+(al+b2 x), (a,,
b, x) + (a, Llbz x) ■ (a , +y,
b, ), (a , +y, 1)2), (a2+y, b
l), (a 2+y, b,)■(al Lbl)+
(81y+b2), (a, Y, bl)+(a2 Ll
)2)■(a, ,b, +x) I(a, lbL+
X) I (a, lb, +x) l (a, b2+X)■
(a, +y, b, -x) I (a, y l
b z-x) r (a, +y, b, -x) I
(a, +y+ b, -x) ■ (a, -y, b, -x
), (a, -y, b, -x), (a, -y, b
, -x), (a, -y, b2-x)■(a, -
y, b, +x), (a, -y, b, Llx), (
a, -y, b, +x), (a, -y, b2+x)■
(a, y, b, +x), (a, +y, b2
+x), (a, -1-y, b, +x), (a
, +y, b2-1-x) If all the characters of the keyword can be recognized in this way, the determination results in steps [phase], [phase], [phase], and [stock] become YES. If even one character cannot be recognized, the determination result in the same step is No.

By using each of the above modes, the device of this embodiment can
It is possible to efficiently store image information such as characters. That is, by specifying the first mote, the operator can process each document, etc., while determining a keyword at his discretion, as in the past. By specifying the fifth mode, a large amount of standard documents can be automatically processed at high speed without operator intervention. If the fourth mode is specified, the operator will have to intervene to check the keywords, but the burden on the operator is lighter than when inputting keywords, and the operator operation time is shorter, so it is possible to handle large amounts of fixed documents etc. can be processed efficiently. Although the processing speed of the fourth mode is lower than that of the fifth mode, more reliable keywords can be registered. In addition, by specifying the second or third mode, you can easily process documents while inputting keywords, including kanji, etc., more quickly and easily than in the second mode, with only simple operator operations for specifying keyword areas. .

Next, searching for stored image information will be explained.

When the operator specifies a search mode using the input section 7, the control section 6 switches to the search mode and waits for a keyword to be input. When the operator inputs a keyword from the input unit 7, the control unit 6 displays the keyword on the CRT display device, converts it into a code used internally (such as the above-mentioned JISII code), and stores the keyword in the address memory 5.
Search for. When a block with the same keyword is detected, the corresponding image information is read out from the image storage section 8 using the recorded address contained therein and displayed on the CRT display device. Thereafter, when the operator inputs an output instruction, the control section 6 causes the printer of the output section 3 to print out the image information.

〔effect〕

As explained in detail above, the present invention is applicable to a mode in which an operator inputs a keyword depending on the performance of characters to be processed, a mode in which the operator specifies a keyword area for each page and keywords are recognized from that area, and a mode in which the operator specifies a keyword area for each page and recognizes the keyword from that area. Since it is possible to select a mode that recognizes multiple characters and other keywords from the same keyword area, it is possible to store and process image information of various documents more efficiently than before, reducing the burden on the operator. .

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is a schematic sectional view of the reading unit, Fig. 3 is a flow diagram of mode I setting, and Fig. 4 is a flow diagram of image information storage processing. , FIG. 5 is a diagram showing an example of screen display, and FIG. 6 is an explanatory diagram of keyword area specification. 1...Reading section, 2...Bage buffer memory, 3.
...Output section, 4...Character recognition section, 5...Address''
1st・6°°“control part・7°゛°”force part・8.1
1... Image storage unit. □ 7・−゛Representative Patent Attorney Makoto Suzuki（・□゛

Claims (1)

[Claims] In an image storage and retrieval device for storing and retrieving image information such as a document, a first means for specifying a mode and an arbitrary area of a document, etc., and inputting a search keyword; A second means for obtaining a search keyword by performing character recognition on the area specified by the first means, when the mode of the first means is specified, uses the keyword input by the first means,
and third means for storing the keywords obtained by the second means in association with respective storage addresses of image information to be newly stored when another mode is specified by the first means. However, there are two types of the latter mode, and if one of the modes is specified by the first means, the area is specified anew by the first means for each page of image information, and then the second means specifies the area. An image storage and retrieval device characterized in that the second means performs character recognition on the same area that has already been designated for a series of image information when the other mode is designated.