JP3279772B2 - OCR recognition device and facsimile device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a memory having a small capacity.
An OCR recognition device capable of recognizing an OCR using the device, and
The present invention relates to a facsimile machine.

[0002]

2. Description of the Related Art In recent years, in an image communication apparatus such as a facsimile, an OCR recognition function and a PC modem are mounted on a terminal device, and when an OCR original is read by a scanner, the contents thereof are recognized by an OCR recognition section. To send OCR data to a destination such as a center through a PC modem
The one provided with the reader for R has been used. In a conventional example of such an OCR reader, in order to perform OCR recognition, a page memory as shown in FIG. 9 is provided as an OCR recognition memory for temporarily storing read data in the OCR reader, and the OCR original is read. Is read for one page, and then the OCR recognition operation is started. Since image data for one page of the OCR document is stored in the page memory, the OCR mark information 2 is developed in a predetermined address area of the page memory. Therefore, the OCR mark information can be recognized and read by the operation of the OCR recognition unit (or the execution of the OCR recognition algorithm).

[0003]

However, in such a conventional OCR reader, a page memory is used as an OCR recognition memory, so that a memory having a large storage capacity is required. For example, if the OCR document is A
Assuming a size of four, approximately 512 for OCR recognition
A KB (kilobyte) page memory is required, and there is a problem that the memory provided on the side of the terminal device, which is small and has a function integration as described above, is too large in capacity and increases the cost.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and has as its object to provide an OCR recognition apparatus and a facsimile apparatus capable of recognizing an OCR using a small-capacity memory.

Another object of the present invention is to provide an OCR recognition device and a facsimile device capable of performing OCR recognition by effectively utilizing an image data storage memory for OCR recognition .

[0006]

In order to achieve the above object, the present invention provides a reading means for reading an OCR document,
The image data obtained by the reading means is
OCR recognition memo with multiple storage areas
And one OCR recognition memory in the OCR recognition memory.
-If a block is stored in different blocks,
From the block that stores the tip of the OCR recognition mark
Character recognition means for sequentially performing OCR recognition;
Character data which is the recognition result of the stage is used for data communication.
And a data modem for performing modulation and demodulation.

Second, a reader who reads an OCR document
And image data obtained by the reading means.
OCR recognition memory and OC sent from the center
Format information storage method for storing information about the R original
And the OCR based on the information about the OCR document.
Only the lines on the CR document where the characters for character recognition exist
Stored in the R recognition memory, and stored in the OCR recognition memory.
A character recognition device that performs OCR recognition on stored image data.
And the character that is the recognition result of this character recognition means.
A data modem for modulating and demodulating data for data communication.
It is a thing.

[0008]

According to the present invention, first, image data read by the scanner is temporarily stored in the OCR recognition memory for several lines in block units by the above configuration.
It recognizes for each block whether or not there is OCR mark information in the image data stored in the recognition memory. If there is no mark information, the data of the recognized block is released and new image data is simultaneously stored in the memory area. Store.
If there is mark information, OCR character content recognition is performed across block boundaries, and then data release of the memory area and storage of new image data are performed in parallel.

Secondly, information for OCR recognition sent from the center to the OCR recognition device of the terminal is stored in a recognition information storage memory, and the OCR recognition device performs OCR recognition in accordance with the OCR recognition information. In addition to performing the recognition, the information for OCR recognition includes the OCR recognition area information in the OCR document, and when reading the OCR document on the OCR recognition device side, only the information for OCR recognition according to the OCR recognition area information is used. Is read to perform OCR recognition.

[0010]

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

FIG. 1 shows a communication system including an OCR-FAX device having a facsimile function and an OCR recognition function, and an OCR center device connected via a line.
The OCR-FAX device 1 has a facsimile communication function for transmitting and receiving image data and a data communication function for transmitting and receiving character data and various command data. A display 3 is provided on the operation surface to display an OCR recognition result and the like. The function key 5 performs switching of operations such as display, printing, and transmission. The OCR center apparatus 2 connected by a line has a host computer having a facsimile communication function and a data communication function. The host computer has a facsimile modem for counting the slips sent from the OCR-FAX apparatus 1 and confirming the sent slips. It returns and downloads the form formats to each OCR-FAX device 1. The OCR-FAX device 1 is installed in a sales office or business place on the orderer side, and the OCR center device 2 is installed in the order receiving center.

FIG. 2 is a block diagram showing the configuration of the OCR-FAX apparatus. The CPU unit 11 controls the whole,
It is connected to each part. The operation unit 12 is a device in which the operator gives an instruction using a keyboard or the like, and performs correction of an OCR recognition result, transmission instruction, display on a screen, printing instruction, and the like. The time generating circuit 13 supplies time information necessary for displaying the transmission time and the form creation time. The communication control unit 14 performs transmission and reception by selecting a facsimile modem or a data modem (PC modem) according to the data. Code processing unit 1
Numeral 5 compresses the transmission data of the facsimile data and decompresses the reception data at the time of reception. The PC communication processing unit 16 performs a process of transmitting and receiving code data. The scanner control unit 17 controls a scanner 18 that reads a facsimile document or an OCR sheet. The scanner 18 includes a lens 20 that focuses light from the OCR original 19, a CCD image sensor 21 that reads an image of the OCR original 19, a reading control unit 22 that controls a reading operation of the CCD image sensor 21, and a preset threshold. And an image correction unit 23 that corrects the read image based on the hold level. Further, the switching unit 24 includes the code processing unit 15,
The input and output of the PC communication processing unit 16 and the scanner control unit 17 are switched.

The memory control unit 25 controls input / output data and memory. The IC card 26 stores software required for OCR recognition, and the memory control unit 2
5 functions as an OCR-FAX device. The memory 27 stores data read by the scanner 18 and stores recognition data subjected to OCR recognition processing.
Further, as shown in FIG. 3A, the memory 27 has a two-port SRAM having a capacity (for example, 32 KB, see the description below) much smaller than a capacity required to store one page of the document. Is used to divide the inside of the SRAM into two blocks, a recognition block 1 and a recognition block 2, and a block boundary L is provided at a boundary between the blocks 1 and 2. Each of the recognition blocks 1 and 2 is set to a capacity capable of storing one line of the document,
In normal writing and reading operations, data corresponding to one row is stored in the recognition blocks 1 and 2, and in the next operation, the previously written data is read from the recognition block 1 or 2 in parallel with the reading. The original is read, and the read image data is stored in the same area of the recognition block 1 or 2. For example, one of the recognition block 1 and the recognition block 2 is operated by the memory control unit 25. Data writing and reading are performed simultaneously on one or both.

The character recognition processing unit 28 loads the software of the IC card 26, reads the data of the OCR sheet read by the scanner 18 stored in the memory 27,
After performing the recognition process, it is stored in the memory 27. The format data storage unit 29 stores OCR sheet format information, display format information, format information, and the like downloaded from the OCR center device 2.
For example, as shown in FIG. 6, the format information is composed of general information 38 and unique information 39 expressed in a table format. The general information 38 is information that is always required as information for OCR recognition, and includes information such as a format type of the OCR information, a mark width for OCR recognition, and a character interval. The unique information 39 is characteristic information of the OCR document, and is information created in consideration of performing a special process in the OCR recognition operation. The unique information 39 includes information such as the total number of recognition marks, the distance dimension from the leading edge of the document to the first recognition mark, and the distance dimension from the first mark to the next recognition mark. The form paper creation unit 30 creates data for printing an OCR sheet as shown in FIG. 7 based on the OCR sheet format information.

The character generator 31 generates a character pattern for displaying or printing code data. The display control unit 32 develops the OCR-recognized data into a bitmap on the display memory 33 according to the display format information, and displays the bitmap on the display 34. The recording control unit 35 develops the OCR-recognized data into a bitmap on the recording memory 36 according to the print format information, and prints the data in the recording unit 37. Further, based on the creation data of the form paper creation unit 30, O
The CR sheet is printed by the recording unit 37.

FIG. 3 is a diagram for explaining the configuration of the OCR recognition memory and various ways of writing the OCR information into this memory. FIGS. 4 and 5 show the operation of the OCR recognition memory control device according to the embodiment. It is a flowchart explaining.

The operation of the facsimile apparatus of the present embodiment configured as described above will be described below. FIG. 4 is a flowchart for explaining a recognition processing operation procedure in a case where a line of OCR mark information is written in the facsimile apparatus according to this embodiment so as to be correctly accommodated in the memory area of each of the recognition blocks 1 and 2. When the facsimile apparatus starts operation, in a processing step (hereinafter simply referred to as a step) ST1, a CCD image sensor 2 is started.
1 starts reading the OCR original 19. CC
The image data read by the D image sensor 21 is A / D converted by the image correction unit 23 and stored in the data storage memory 27. The memory address is output from the recognition unit 28 to the memory control unit 25 of the data storage memory 27. As an example, assume that the memory address is Start0000 (HEX). The OCR manuscript 19 first has two blocks End7
Read up to FFF (HEX). In this embodiment, as an example, one recognition block is 16 KB (128 Kb).
it). The basis is that the number of bits in one line is 2048 bits and 1 mm (mm)
Eight lines are required for sub-scanning, and the recognition width is 8 mm
(Millimeters), it is about 16 KB. In two blocks, it is 32 KB (256 Kbit).

Next, in step ST2, mark information search processing is performed. Here, in order to recognize the mark information, first, a black box of the recognition mark is searched, and it is checked whether or not there is a black box in the first block. If there is a black box, a character content recognition operation is performed from the mark in step ST3 based on the identification parameter (recognition area size, gap, number of recognized characters, etc.), and in the next step ST4, the memory area for the recognition block 1 is allocated. After the release, in step ST5, the next area of the OCR original 19 is set to 1 by the CCD image sensor 321.
Read blocks. On the other hand, in step ST2, 1
When it is determined that there is no black box in the block, the memory area for the recognition block 1 is released in step ST6, and then the process proceeds to step ST5. Step S
When reading of one block at T5 is completed, a recognition process of mark information is performed again at step ST7. Here, as in step ST2, a black box of the recognition mark is searched for, and whether or not there is a black box in the second block is checked. If there is a black box, a character content recognition operation is executed from the mark based on the identification parameter information in step ST8.
After the memory area for the recognition block 2 is released in step 9, the next area of the OCR original 19 is read by the CCD image sensor 21 in step ST10 for two blocks. On the other hand, when it is determined in step ST7 that there is no black box in the first block, the memory area for the recognition block 2 is released in step ST11, and the process proceeds to step ST10. After reading the OCR original 19 in step ST10, the process returns to step ST2.

When the reading and mounting member of the OCR original 19 is performed by such processing, the OCR reading is performed when the line of the mark information as shown in FIG. 3 (b) is correctly stored in the memory area of each of the recognition blocks 1 and 2. It is. However, in the actual OCR reading, it is rare that such a line of the mark information is correctly stored in the memory area of each of the recognition blocks 1 and 2. Generally, the mark information comes to the boundary of the block of the memory 27 when the mark information reaches the boundary of the block. It happens that one recognition block 1 and 2 is straddled. FIG. 5 is a flowchart illustrating a recognition processing operation procedure at a memory boundary in such a case.

In this case, when the facsimile apparatus starts the reading operation, in processing step ST21, the CC
The reading of the OCR original 19 by the D image sensor 21 is started. First, two blocks of the OCR original 19 are read. The image data read by the CCD image sensor 21 is A / D converted by the image correction unit 23 and stored in the data storage memory 21. Recognition unit 2
8 to the memory control unit 25 of the data storage memory 27. Next, in step ST22, a search process for mark information is performed. Here, in order to recognize the mark information, first, a black box of the recognition mark is searched and it is checked whether or not this box is over the boundary of the recognition block. The processing operation according to the flowchart of FIG.
On the other hand, if it is determined in step ST22 that the mark information covers the boundary portion, it is checked in step ST23 whether or not the mark information exists in both the recognition block 1 and the recognition block 2. While error processing is performed, if both are present, the address of the last part with the mark information, that is, the mark end address value is stored in step ST24.

Here, there are two cases where the mark information is detected at the boundary between the recognition blocks 1 and 2. That is, (1) a case where a black box as mark information is detected across the block boundary L from the recognition block 1 to the recognition block 1 (FIG. 3C). And (2)
When a black box is detected across recognition block 2 to recognition block 1, that is, at the lower boundary of recognition block 2. In this case, the rest is at the upper boundary of the recognition block 1 (FIG. 3D). By the way,
The mark information in which the mark end address value is stored in step ST24 exists from the recognition block 1 to the recognition block 2 at the block boundary L shown in (1). In such a case, since the mark information exists in both the recognition blocks 1 and 2, the data stored in the recognition blocks 1 and 2 cannot be released until the character content recognition processing is completed.

Therefore, when the storage of the mark end address value is completed as described above, the process proceeds to step ST25, where the character content of the portion where the mark information is present is recognized. This character content recognition processing is performed in the next step ST26.
It is repeatedly checked whether or not the recognition process has been completed. If not, the monitoring operation is continued until the process is completed.
At 27, the memory area up to the area corresponding to the mark end of the recognition block 1 and the recognition block 2 in the memory 27 is released. This memory release area extends from the address number 0000 to 60000 (24 KB) according to the memory map shown in FIG. This memory 2
In parallel with the partial release process of step 7, reading of the OCR original 19 by the CCD image sensor 21 in step ST28 is started. The image data read by this reading operation is stored in the block end of the recognition block 2, that is, in the area from the head of the memory 27 to the recognition block 1. As described above, since the memory 27 is a two-port memory, the release of data and the storage of another data in place of it are executed in parallel.

When the reading of the OCR original 19 and the storing of the image data are completed, a mark recognizing process is started in a step ST29 from the mark end of the recognition block 2 to the block boundary L which is the block end of the recognition block 1. Mark recognition is performed for the area up to the portion, and it is constantly checked whether this processing is completed. Accordingly, in this mark recognition processing, mark information is detected from the remaining 8 KB data of the recognition block 2 and the 16 KB data newly read by the CCD image sensor 21. When the mark recognition processing for the area is performed, in step ST30, as the second half operation of the mark recognition processing, it is determined whether or not there is mark information in the recognition block 2, that is, whether or not there is mark information that extends from the recognition block 2 to the recognition block 1. Check if. That is, in this mark recognition process, the presence of the mark information described in the above (2) and shown in FIG. Such a search for presence / absence of mark information is performed depending on whether or not mark information is detected at the lower boundary in the recognition block 2. And the step ST
If it is determined in step 30 that there is no mark information in the recognition block 2, there is no mark information that extends from the recognition block 2 to the recognition block 1, so that the recognition block 2 in the memory 27 is released in step ST31, and then step ST32 It is checked whether or not there is any mark information in the recognition block 1. If there is no mark information in the recognition block 1, the recognition block 1 is released in step ST33, and the process returns to step ST21. On the other hand, if there is any mark information in the recognition block 1 in step ST32,
R character content recognition processing and step ST
At 35, it is checked whether or not this character content recognition processing has been completed. If completed, the recognition block 1 is released at step ST33, and the process returns to step ST21.

Further, when it is determined in step ST30 that there is mark information in the recognition block 2, there is mark information that extends from the recognition block 2 to the recognition block 1. Therefore, in step ST36, the recognition block 2 changes to the recognition block. In step ST37, an OCR character content recognition process is performed.
In step ST38, it is checked whether or not this character content recognition processing has been completed, and in step ST38, the memory 27 is released from the recognition block 2 to the recognition block 1, and the process returns to step ST21.

Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 6 is a diagram showing the data structure of the information for OCR recognition, FIG. 7 is a diagram showing the actual structure of the OCR document indicated by the information for OCR recognition in FIG. 6, and FIG. 8 is related to the second embodiment. 5 is a flowchart illustrating an OCR recognition processing operation procedure in the facsimile machine. When the facsimile apparatus starts operating, in a processing step (hereinafter simply referred to as a step) ST1, first, in this embodiment, the center apparatus 2 sends general information 38, which is information for OCR recognition, to the terminal facsimile apparatus 1 and unique information. 39, the facsimile machine 1 stores the information for recognition in the format data storage unit 29. When this is completed, a reading control operation is performed in step ST2, and document feeding is started. In this document feed control operation, it is checked in step ST3 whether or not the leading edge of the OCR document 19 has been detected.
A document front-to-front mark distance value given as 9 is set. In step ST5, reading control is performed while document feeding is continued. Image data read during this time is not stored in the image data storage memory 4 and is discarded. On the control side, reading of the OCR document 19 is performed in step ST6. It is checked whether or not the position has reached the set value of the document front-end mark distance. Here, when the reading position of the OCR original 19 reaches the set value of the original leading end-leading mark distance, it is recognized that the original reading start position has been reached for the first time, and the reading control recognition width (8 mm in this embodiment) is determined in step ST7. Is set, reading control is performed while document feeding is continued, and image data read during this time is stored in the data storage memory 27. To perform this processing operation, the character recognition processing unit 2
Numeral 8 designates the number of original movement lines to the scanner control unit 17, and data of the specified line is stored in the data storage memory 27.

On the other hand, the character recognition processing section 28
At 8, it is checked whether the reading position of the OCR original 19 has reached the recognition width set value. Here, the OCR manuscript 19
When the reading position has reached the recognition width set value, step ST9
In the character recognition accepting process, the character recognition processing unit 28
Performs a process of recognizing the stored data at this processing stage and a process of counting the mark width. When the recognition process is completed,
The distance value to the next mark in the unique information 39 from the format data storage unit 29 is taken in as the distance value to the next recognition mark. In this embodiment, the recognition area width is set to be slightly smaller than the mark width so as to absorb an error in reading distance movement. As an exceptional operation, the character recognition processing unit 28 moves the specified distance and performs a recognition processing operation of the data stored in the data storage memory 27. However, the character recognition processing unit 28 stores data exceeding the mark width due to a mechanical error. It may be stored in the memory 27. If a distance value to the next recognition mark is set and reading is performed in this state, errors are accumulated, the specified distance is moved for recognition data storage, and the data is stored in the data storage memory 27 as recognition data.
May not be the data stored in the mark width area correctly. Therefore, in step ST10, it is checked whether or not there is data after the mark width. If it is determined that there is no data, in the next step ST11, it is checked whether or not the total number of marks has been reached (in this embodiment, “5” is set as the unique information 39). If it has reached, it is assumed that a series of OCR recognition processing has been completed and the processing ends. If the total number of marks has not been reached, step S
In T12, the reading control is performed while setting the distance value to the next mark and continuing the document feeding, but the image data read during this time is discarded without being stored in the data storage memory 27, while the control side executes step ST13.
In step ST12, the reading position of the OCR original 19 is
Check whether the value set in step has been reached. And
If it is determined that the set value has been reached, the process returns to step ST7.

If it is determined in step ST10 that there is data after the mark width, as described above, the data is moved by a specified distance for storing recognition data due to accumulation of errors and stored in the data storage memory 27 as recognition data. In step ST14, the character recognition processing unit 28 determines in the character recognition processing unit 28 that the image data exceeding the mark width due to a mechanical error is stored in the image data storage memory 4. The number is counted, and a subtraction process is performed in step ST15, and the scanner control unit 17 is given a new distance value to the next recognition mark based on a value obtained by adding a correction value to the distance value to the next mark. Set the number of original movement lines. Thus, in this embodiment, the error accumulation of the moving distance is corrected whenever it becomes necessary every time. Then, in the next step ST16, it is checked whether or not the total number of marks has been reached. If the total number has been reached, a series of OCR recognition processing is completed and the processing is terminated. If the total number of marks has not been reached, the reading control is performed while setting the distance value to the next mark and continuing the document feeding in step ST17. The image data read during this time is stored in the data storage memory 27. On the other hand, the control side proceeds to step ST13 to check whether the reading position of the OCR original 19 has reached the value set in step ST17. Then, when it is determined that the set value has been reached, step ST
It returns to the process of 7.

When the OCR original 19 is read by such a series of processing, the OCR original 19 is read in a state where the recognition position is known in advance, so that the memory capacity of the data storage memory 27 and the like can be reduced. , The cost of the OCR recognition device can be reduced.
In this embodiment, the mark reference method is used as a method of detecting the error of the moving distance. However, the error may be detected from the height information of the character for character recognition.

[0029]

As described above, according to the present invention, first, the OCR recognition memory of the OCR recognition apparatus is a small-capacity memory having a plurality of blocks and capable of storing read image data in block units. A recognition unit for retrieving the presence of OCR mark information from the data in the OCR recognition memory and recognizing the character content corresponding to the mark information; releasing the data in the OCR recognition memory; Since a memory control unit that executes storage in parallel is provided, and mark information is searched and recognized for each block,
The advantage is that accurate character recognition can be performed with a small memory capacity. Further, when the mark information is stored across different blocks, recognition is performed across blocks, so that OCR recognition can be performed without storing the mark information for each block. .

Secondly, according to the present invention, an image data read by a scanner is subjected to an OC by a character recognition processing unit.
An OCR recognition device for performing R recognition is provided with a memory for storing information relating to a recognition area for storing information for OCR recognition sent from the center, and the recognition processing unit performs the OCR recognition stored in the format data storage memory. Since the OCR recognition is performed by causing the scanner to read only the OCR recognition information in accordance with the area information for use, the effect that accurate character recognition can be performed by effectively using a small-capacity memory is obtained.

[Brief description of the drawings]

FIG. 1 shows an OCR facsimile apparatus and an OC according to the present invention.
The figure which shows the communication system which consists of an R center apparatus

FIG. 2 is a block diagram showing a configuration of an OCR facsimile apparatus according to the present invention.

FIG. 3A is a map diagram showing an internal configuration of a memory at the time of the OCR recognition memory control processing according to the embodiment; FIG. 3B is a mark on the memory at the time of the OCR recognition memory control processing according to the embodiment; A map diagram showing an example of a storage state of information. (C) A map diagram showing another example of a storage state of mark information in a memory at the time of the OCR recognition memory control process according to the embodiment. The map figure which shows another example of the storage state of the mark information in the memory at the time of such OCR recognition memory control processing.

FIG. 4 is a flowchart for explaining a recognition processing operation corresponding to a mark information storage state as shown in FIG. 3A in the embodiment.

FIG. 5 is a flowchart for explaining a recognition processing operation corresponding to a mark information storage state as shown in FIGS. 3B and 3C in the embodiment.

FIG. 6 is a diagram showing a data configuration of information for recognition used in the OCR recognition apparatus according to the embodiment.

FIG. 7 is a view showing the actual configuration of an OCR document represented by the information for OCR recognition shown in FIG. 2;

FIG. 8 is a flowchart for explaining the operation of the OCR recognition apparatus according to the embodiment.

FIG. 9 is a map diagram showing an internal configuration of a memory used in a conventional OCR reading device.

[Explanation of symbols]

 17 Scanner control unit 19 OCR document 18 Scanner 27 Data storage memory 28 Character recognition processing unit 29 Format data storage memory

Continuation of front page (56) References JP-A-57-137974 (JP, A) JP-A-56-6560 (JP, A) JP-A-4-289985 (JP, A) JP-A-59-32074 (JP, A) JP-A-59-32065 (JP, A) JP-A-55-119779 (JP, A) JP-A-58-103266 (JP, A) JP-A-5-37680 (JP, A) JP-A-4-218172 (JP, A) JP-A-4-156172 (JP, A) JP-A-55-112683 (JP, A) JP-A-55-110377 (JP, A) JP-A-2-224083 (JP, A) A) (58) Field surveyed (Int. Cl. ⁷ , DB name) H04N 1/00-1/00 108 H04N 1/32-1/36 H04N 1/42-1/44 G06K 9/20

Claims (5)

(57) [Claims] A reading means for reading an OCR document;
The image data obtained by the sampling means is
Memory for OCR having a plurality of areas for storing data in units of
And one OCR recognition memory in the OCR recognition memory.
If the block is divided into different blocks and stored,
From the block that stores the tip of the CR recognition mark
Character recognition means for performing next OCR recognition, and this character recognition means
Character data, which is the recognition result of
An OCR recognition device including a data modem for performing demodulation. 2. A reading means for reading an OCR document, and the reading means
The image data obtained by the acquisition means is plotted for several lines.
Memory for OCR having a plurality of areas for storing data in units of
And straddle the block boundary of the OCR recognition memory.
When the OCR recognition mark of the image data is stored
The character recognition hand that performs OCR recognition while straddling
And the character that is the recognition result of this character recognition means.
A data modem for modulating and demodulating data for data communication.
OCR recognition device. 3. A reading means for reading an OCR document, and the reading means.
The image data obtained by the acquisition means is plotted for several lines.
Memory for OCR having a plurality of areas for storing data in units of
And straddle the block boundary of the OCR recognition memory.
When the OCR recognition mark of the image data is stored
In this case, after the OCR recognition processing of the image data, the block boundary
Image data beyond the field to the end of the OCR recognition mark
Memory control means for performing processing for releasing the data,
The resulting character data is modulated and demodulated for data communication.
An OCR recognition device comprising a data modem for performing the operation. 4. A reading means for reading an OCR document, and the reading means
OCR for storing image data obtained by the
Knowledge memory and OCR manuscript sent from the center
A format information storage memory for storing information
The text of the OCR manuscript based on the information about the OCR manuscript
Only the line where the character recognition character exists exists in the OCR recognition memo.
Image data stored in the OCR recognition memory.
Character recognition means for performing OCR recognition on
Character data, which is the recognition result of the character recognition means, is
OCR certification with a data modem for modulation and demodulation for communication
Sense device. 5. The OCR recognition device according to claim 1, wherein
And modulate and demodulate the image data for facsimile communication.
Facsimile modem and the OCR recognition device
Switching between data modem and facsimile modem
Communication control means for transmitting the recognition result to the line.
Facsimile machine.