JP3217442B2 - Optical character reader - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical character reader for dynamically controlling a dictionary subset.

[0002]

2. Description of the Related Art In recent years, OCR (optical character recognition device) has been developed.
With the advancement of character recognition technology, small size and high performance have been achieved, and they have been supplied at relatively low cost. This OCR
Has become widespread in general, and has recently been connected as a peripheral device of a relatively high-performance personal computer, and has been frequently used for all applications.

The OCR scans a form with a scanner, converts the form image into image data, cuts out a character pattern from the image data, and performs character recognition by comparing and comparing this character pattern with a built-in recognition dictionary.

Incidentally, the above-mentioned OCR is conventionally used in FIG.
When recognizing the forms (a) to (f) written or printed in various formats as shown in the amount column of the read form shown in FIG. That is, in the example of FIG. 4, "0-9", "*",
Recognition is made using “￥”.

[0005]

According to the above-described conventional example, even when a "numeric character" is recognized, "*" or "@"
Since it is necessary to perform matching with the dictionary of FIG. 4, it may hinder speeding up, and the recognition rate may be unnaturally erroneously recognized as “{123} *” or the like shown in FIG. Not good.

The present invention has been made in view of the above circumstances, and recognizes characters in a predetermined order. When recognizing characters in the middle, a dictionary subset is recognized based on the recognition result of the characters recognized so far. It is an object of the present invention to provide an optical character reading device in which a recognition time is shortened and a recognition rate is improved by dynamically changing the character.

[0007]

Optical character reader of this invention, in order to solve the above-mentioned object, at least numbers first from the second symbol and the number indicating the non-first symbol from the monetary units that a unit amount A plurality of characters are arranged in a direction from one end to a second end, and when the first and second symbols are arranged, the characters are arranged only on the first end side of the numeral. An optical character reading device for reading the characters of a form on which is recorded, an image scanner that scans the form and reads the image of the form to obtain image data, and an image scanner that reads the image data read by the image scanner. An image memory for storing, a character extracting unit for extracting a character pattern from the image data stored in the image memory, and the first memory for collating with the character pattern. A dictionary memory storing a dictionary corresponding to a symbol, the second symbol, and the numeral; and a direction from the first end character to the second end of the plurality of characters recorded in the form. Character recognition means for recognizing a character by comparing a character pattern cut out from the cutout means with a character pattern sequentially cut out from the cutout means and a character recognition means when the character recognition means recognizes a character. This is for designating a dictionary stored in the dictionary memory. As an initial setting, it is designated to match with the dictionary corresponding to the first symbol, the second symbol, and the numeral, and thereafter, the character is sequentially designated. When recognizing the next character based on the character recognition result of the recognizing means, when the character recognition result is the first symbol, it is checked against the dictionary corresponding to the second symbol and the numeral. specify Characterized by comprising a write subsets control means. Further, the optical character reading apparatus of the present invention is characterized in that at least a numeral is selected from a first end and a second end from a first symbol indicating a unit of money and a second symbol and a numeral indicating a unit other than the unit of money. When the first and second symbols are arranged, a plurality of characters arranged only on the first end side of the numeral are recorded on a form in which characters are recorded. An optical character reader for reading, comprising: an image scanner that scans the form to read an image of the form to obtain image data; an image memory that stores image data read by the image scanner; Character extracting means for extracting a character pattern from image data stored in a memory; and a first symbol and a second symbol for collating with the character pattern A dictionary memory in which a dictionary corresponding to a serial number is stored, and characters that are adjacent to each other in the direction from the first end to the second end of the plurality of characters recorded in the form are sequentially cut out. Character recognition means for recognizing a character by collating a character pattern cut out from the means with a dictionary stored in the dictionary memory; and a dictionary stored in the dictionary memory for collating when the character recognition means recognizes a character. Is specified as an initial setting to match with the dictionary corresponding to the first symbol, the second symbol, and the numeral, and thereafter, based on the character recognition result of the character recognition unit. When recognizing the next character, if the character recognition result is the numeral, specify that the character is to be compared with the dictionary corresponding to the numeral, and the character recognition result is the first symbol and the character immediately before the first symbol. of If character recognition result of the second symbol Specifies a match dictionary corresponding to the numbers,
If the character recognition result is the first symbol and the immediately preceding character recognition result is not the second symbol, a dictionary that specifies to collate with the dictionary corresponding to the second symbol and the number And a subset control unit.

[0008]

According to the optical character reading apparatus of the present invention, at least a numeral is selected from a first symbol indicating a unit of money, a second symbol indicating a unit other than the unit of money, and a numeral. In the case where the first and second symbols are arranged, the document is scanned on a form on which a plurality of characters arranged only on the first end side than the numbers are recorded. The image data of the form image is obtained from the image data, a character pattern is cut out from the image data, and a plurality of characters recorded on the form are sequentially read from a character at a first end to a character adjacent to a second end. Characters are recognized by comparing the cut-out character pattern with the dictionary stored in the dictionary memory. At this time, the dictionary subset control means :
(1) Initially, it is specified to match with the dictionary corresponding to the first symbol, the second symbol, and the numeral. Thereafter, when recognizing the next character based on each character recognition result, character recognition is performed. If the result is the first symbol, check with the dictionary corresponding to the second symbol and the number; (2) Initially set the dictionary corresponding to the first symbol, the second symbol and the number When the character recognition result is a number when the next character is sequentially recognized based on each character recognition result, it is specified to perform the matching with the dictionary corresponding to the number. If the result is the first symbol and the preceding character recognition result is the second symbol, it specifies to match with the dictionary corresponding to the number, and the character recognition result is the first symbol and the first If the previous character recognition result is not the second symbol, the second
Specifies to match with the dictionary corresponding to the symbol and the number. In this way, by dynamically controlling the dictionary for collating with the character to be recognized next based on the recognition result of the character that has already been recognized, the recognition time can be reduced and the recognition rate can be improved. become.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of the present invention.

In FIG. 1, reference numeral 1 denotes a scanner, which is a component that scans a form and converts it into image data. Reference numeral 2 denotes an image memory, which stores image data obtained via a scanner. Reference numeral 3 denotes a character extracting unit, which is a component for extracting a character pattern from the image data stored in the image memory 2. Reference numeral 4 denotes a character recognizing unit, which is a component for recognizing characters by collating a cut-out character pattern with a recognition dictionary described later. Reference numeral 5 denotes a dictionary memory for storing a recognition dictionary. Reference numeral 6 denotes a recognition order control unit, which controls the recognition order of characters.

Reference numeral 7 denotes a dictionary subset control unit which is a point of the present invention, and controls a dictionary subset to be collated with a character to be recognized next, based on a recognition result of a character which has already been recognized. FIG. 2 is a flowchart showing the operation of the embodiment of the present invention.

Hereinafter, the operation of the embodiment of the present invention shown in FIG. 1 will be described in detail with reference to FIG. Here, FIG.
An example of recognizing the amount column shown in FIG. The formats handled in this case are six types shown in the following (A) to (F). (A) Numeric characters only (B) A single character “￥” before a numeral (C) A character “*” before a numeral (D) A plurality of characters “*” before a numeral (E) A character “￥” before a numeral "" Is one character, and "*" is one or more characters before it. (F) "*" is one or more characters before the number, and "￥" is one character before it. (A) is a number after a number, ( B ) is a number that does not appear after ￥, ( C )
Must be a number after “*”.

Therefore, the recognition order control unit 6 controls the character recognition unit 4 so that recognition is performed in order from the left. The dictionary subset control unit 7 first sets the initial dictionary subset to “0”
9 ”,“ * ”,“ ￥ ”, and then (A),
The dictionary subset is changed according to the flowchart shown in FIG. 2 so as to satisfy (B) and (C). That is, it is checked whether or not the recognition result is a number.
The dictionary subset is set to “0 to 9”, and the process proceeds to the step of repeatedly recognizing the next character. If not, then it is checked whether the recognition result is ￥. If “￥” here, ￥ is removed from the dictionary subset, and it is checked that the recognition result of the previous character is *. Here, if it is recognized as *, the dictionary subset is set to 0 to 9; if not, the process proceeds to the step of repeatedly recognizing the next character. Thus, unnecessary collation with the dictionary can be eliminated, and unnatural misreading can be eliminated.

FIG. 3 shows a specific example. In the case of recognizing the address column as shown in FIG. 3, it is possible to limit the dictionary subset when recognizing the next city / district column based on the recognition result of the prefecture column. For example, when the recognition result of the prefecture column is “Tokyo”, the city / county column may be recognized with a dictionary subset of characters included in the city / county name in Tokyo.

[0015]

As described above, according to the present invention, the recognition time is dynamically controlled by dynamically controlling the dictionary subset to be collated with the character to be recognized next, based on the recognition result of the character already recognized. This contributes to shortening and improving the recognition rate.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a flowchart showing the operation of the embodiment of the present invention.

FIG. 3 is a view showing an example of a character recognition form for explaining the operation of the embodiment of the present invention.

FIG. 4 is a diagram showing an example of a conventional character recognition form.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... scanner, 2 ... image memory, 3 ... character extraction part, 4 ... character recognition part, 5 ... dictionary memory, 6 ... recognition order control part, 7 ... dictionary subset control part.

Claims (2)

(57) [Claims] 1. A first symbol indicating a unit of money and a unit of money
At least a number from the second symbol and the number that indicate other than the place
The characters are arranged in a direction from the first end to the second end.
If the first and second symbols are arranged, the number
A plurality of sentences arranged only on the first end side of the character
Optical characters to read the characters of the form where the characters were recorded
A reading device that scans the form and reads an image of the form.
An image scanner that obtains image data and an image scanner that reads the image data
An image memory for pay, text from the image data stored in this image memory
Character extracting means for extracting a character pattern; and a first symbol for comparing the character pattern with the first symbol.
A dictionary storing a dictionary corresponding to the second symbol and the numeral
And a first memory of the plurality of characters recorded in the form.
In order from the character at the end to the character adjacent to the direction of the second end.
Next character pattern cut out from the cutting means and the previous
By comparing with the dictionary stored in the dictionary memory
Character recognition means for character recognition, and the dictionary to be collated when the character recognition means recognizes characters
This is used to specify a dictionary stored in memory.
The first symbol, the second symbol, and the numeral as setting
To match the dictionary corresponding to
The next character is recognized based on the character recognition result of the character
When recognizing, the character recognition result is the first symbol
Matches with the dictionary corresponding to the second symbol and the number
To and a dictionary subset control means for specifying that
An optical character reading device, comprising: 2. A first symbol indicating a unit of money and a unit of money.
At least a number from the second symbol and the number that indicate other than the place
The characters are arranged in a direction from the first end to the second end.
If the first and second symbols are arranged, the number
A plurality of sentences arranged only on the first end side of the character
Optical characters to read the characters of the form where the characters were recorded
A reading device that scans the form and reads an image of the form.
An image scanner that obtains image data and an image scanner that reads the image data
An image memory for pay, text from the image data stored in this image memory
Character extracting means for extracting a character pattern; and a first symbol for comparing the character pattern with the first symbol.
A dictionary storing a dictionary corresponding to the second symbol and the numeral
And a first memory of the plurality of characters recorded in the form.
In order from the character at the end to the character adjacent to the direction of the second end.
Next character pattern cut out from the cutting means and the previous
By comparing with the dictionary stored in the dictionary memory
Character recognition means for character recognition, and the dictionary to be collated when the character recognition means recognizes characters
This is used to specify a dictionary stored in memory.
The first symbol, the second symbol, and the numeral as setting
To match the dictionary corresponding to
The next character is recognized based on the character recognition result of the character
When the character recognition result is the number,
Specifies to match the dictionary corresponding to the digit,
The recognition result is the first symbol and the character recognition result immediately before the first symbol.
If the result is the second symbol, a dictionary corresponding to the number
Is specified, and the character recognition result is
And the character recognition result immediately before the character is the second symbol.
If not, the letter corresponding to the second symbol and the number
An optical character reading device, comprising: dictionary subset control means for designating collation with a book .