JPH07105321A - Word recognition device, address recognition device and word recognition method - Google Patents

Abstract

PURPOSE:To provide the word recognition device by which a word is quickly recognized. CONSTITUTION:An address recognition device 9 is made up of a word recognition device 17, an address recognition circuit 19 and an address dictionary 21. The word recognition device 17 is provided with a word recognition circuit 39 to reference a word dictionary 41 thereby recognizing a word and listing up word candidates. The processing to list up the word object is made by comparing a character candidate outputted from a character recognition circuit 35 with a word in the word dictionary 41 and the word recognition circuit 39 decides a character threshold level for each character in the word dictionary 41 and uses it as one of criteria and a dissident character number is counted for each character and it is compared with a maximum dissident character number decided by the word dictionary.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a word recognition device, an address recognition device and a word recognition method.

[0002]

2. Description of the Related Art A conventional word recognition device recognizes a character and, as a result of the character recognition, lists a plurality of character candidates together with a character evaluation value (a value determined by the character candidate likeness), and refers to the word dictionary from these. To recognize words.

In the above-mentioned word dictionary, a plurality of words and the maximum number of non-matching characters and a threshold value corresponding to each of the plurality of words are defined and registered. The maximum number of non-matching characters is the limit value of the number of characters counted when there is no candidate that should be used as a character that constitutes a word among the character candidates listed by the result of the above-mentioned character recognition. Words that are not candidates for words. In addition, the threshold serves as a reference for determining a word when recognizing a word, and a word in which the sum of the character evaluation values of the characters forming the word does not reach the threshold is not a word candidate.

In such a word recognition device, since the check based on the maximum number of non-matching characters in the word dictionary is performed after recognizing all the characters constituting each word, the comparison process must be performed at least as many times as the number of word characters. For example, if it is predetermined in the word dictionary that the maximum number of non-matching characters is 0, the comparison process for the second and subsequent characters may be unnecessary if there is no character candidate that matches the first character of the character string. Instead, the comparison process was performed.

Further, since the threshold value of the word is compared with the total character evaluation value forming the word, even if the character evaluation value of each character candidate varies, it is evaluated uniformly. become. For this reason, the reliability of the word candidates is low.

[0006]

As described above, in the conventional word recognition device, since the check based on the maximum number of mismatched characters in the word dictionary is performed after recognizing all the characters constituting each word, at least the number of word characters Only the comparison process must be performed. For example, if it is predetermined in the word dictionary that the maximum number of non-matching characters is 0, the comparison process for the second and subsequent characters may be unnecessary if there is no character candidate that matches the first character of the character string. Instead, the comparison process was performed.

Further, since the threshold value of the word is compared with the sum of the character evaluation values forming the word, even if the character evaluation value of each character candidate varies, it is evaluated uniformly. become. For this reason, the reliability of the word candidates is low. Therefore, an object of the present invention is to provide a word recognition device that eliminates the above-mentioned drawbacks and can perform word recognition processing quickly.

[0008]

In order to achieve the above object, the first invention comprises a reading means for reading a word written on a paper sheet. The character recognition unit recognizes the word of the paper sheet read by the reading unit for each character, and outputs a plurality of character candidates to which an evaluation value is given for each character. The storage means stores a plurality of words to be written on the paper sheet, an allowable limit value of the number of non-matching characters corresponding to each word, and a threshold value set for each character of the word. The comparison means compares whether or not each character of the word stored in the storage means matches the character candidate output by the character recognition means. The counting means, when the comparison by the comparison means, when the evaluation value of the character candidate output from the character recognition means is smaller than the threshold value of the corresponding character stored in the storage means, a non-matching character. Count as if there is. The determining means is a result of counting by the counting means,
The word recognition device is characterized in that when the count value exceeds the allowable limit value of the number of non-matching characters stored in the storage means, the word is not a corresponding word.

In order to achieve the above object, the second invention comprises a reading means for reading a word written on a paper sheet. The character recognition unit recognizes the word of the paper sheet read by the reading unit for each character, and outputs a plurality of character candidates to which an evaluation value is given for each character. The storage means stores a plurality of words to be written on the paper sheet and a threshold value determined for each character of these words. The comparison means compares each character of the word stored in the storage means with the character candidate output from the character recognition means. The determining means is characterized by determining that the character candidate is an unmatched character when the evaluation value of the character candidate is smaller than the threshold value of the corresponding character stored in the storage means during the comparison by the comparing means. It is a word recognition device.

In order to achieve the above object, the third invention comprises a reading means for reading a word written on a paper sheet.
The character recognition unit recognizes the word of the paper sheet read by the reading unit for each character, and outputs a plurality of character candidates to which an evaluation value is given for each character. The storage means stores a plurality of words to be written on the paper sheet and an allowable limit value of the number of mismatched characters corresponding to each word. The determining means sequentially compares the plurality of corresponding character candidates from the first character of the word stored in the storage means,
It is determined whether or not there is a character candidate that matches the word among the character candidates. As a result of the discrimination by the discriminating means, the word recognizing means recognizes as a character forming a word when there is a matching character candidate. When there is no matching character candidate as a result of the judgment by the judging means, the counting means counts the number of characters as an unmatched character. The discrimination means is a word recognition device characterized in that, as a result of the counting by the counting means, the word cannot be recognized when the counted value exceeds the allowable limit value of the number of mismatched characters stored in the storage means.

In order to achieve the above object, the fourth invention comprises a reading means for reading a word written on a paper sheet.
The character recognition unit recognizes the word of the paper sheet read by the reading unit for each character, and outputs a plurality of character candidates to which an evaluation value is given for each character. The storage means stores a plurality of words to be written on the paper sheet, an allowable limit value of the number of non-matching characters corresponding to each word, and a threshold value set for each character of the word. The comparison means compares whether or not each character of the word stored in the storage means matches the character candidate output by the character recognition means. The address recognition unit recognizes an address from the word when the word and the character candidate match as a result of the comparison by the comparison unit. The counting means, when the comparison by the comparison means, when the evaluation value of the character candidate output from the character recognition means is smaller than the threshold value of the corresponding character stored in the storage means, a non-matching character. Count as if there is. The discriminating means is an address recognition device characterized by discriminating that the word is not a corresponding word when the counted value exceeds the allowable limit value of the number of mismatched characters stored in the storage means as a result of the counting by the counting means.

In order to achieve the above object, in the fifth invention, in a word recognition method for recognizing a word from a word stored in a word dictionary and a candidate for each character constituting a word obtained by character recognition, In a first step, a word, a threshold value determined for each character of this word, and an allowable limit value of the number of non-matching characters of the word are stored in advance in a word dictionary. In the second step, the words in the medium description are read. In the third step, the word of the medium read in the second step is recognized for each character, and a character candidate to which an evaluation value is given for each character is output. In a fourth step, the threshold value for each character of the word in the word dictionary stored in the first step is compared with the evaluation value for each character of the character candidate output in the third step. . In a fifth step, it is determined whether the character of the word whose evaluation value is determined to be equal to or more than the threshold value in the fourth step and the character candidate match. In the sixth step, when the evaluation value is determined to be smaller than the threshold value in the fifth step, the characters of the word are counted as non-matching characters. In a seventh step, it is determined whether or not the number of non-matching characters counted in the sixth step is larger than the allowable limit value of the number of non-matching characters stored in the first step. In the eighth step, as a result of the determination in the seventh step, when the number of non-matching characters is larger than the allowable limit value of the number of non-matching characters, it is determined that this word is not a corresponding word. Is the way.

[0013]

In the word recognition apparatus according to the first aspect of the present invention configured as described above, the character threshold is determined for each character in the word dictionary, and this is used as a criterion for determination, and the number of non-matching characters is counted for each character. Word recognition is performed by comparing with the maximum number of non-matching characters that is predetermined in the word dictionary.

In the word recognition apparatus according to the second aspect of the present invention thus constructed, the character threshold value is determined for each character in the word dictionary, and this is used as the criterion. In the word recognition device according to the third aspect of the invention configured as described above, a comparison is made with the maximum number of non-matching characters set in the word dictionary each time a character of the word is recognized.

In the address recognition apparatus according to the fourth aspect of the present invention configured as above, a character threshold value is set for each character in the word dictionary, and this is used as a criterion for determination, and the number of mismatched characters is further counted for each character. The word recognition is performed by comparing with the maximum number of non-matching characters set in advance in the word dictionary, and the address recognition is performed based on these judgments. In the word recognition method according to the fifth aspect of the invention thus configured, a predetermined word is not recognized as a corresponding word based on the character threshold value for each character and the count of the number of mismatched characters.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment according to the present invention will be described in detail below with reference to the drawings. First, with reference to FIG. 2, the configuration of a mail sorter capable of mounting the word recognition device of this embodiment will be described. FIG. 2 is a front view of the mail sorting machine according to the present embodiment.

The postal sorter 1 is provided with a supply unit 3 from which postal items P, such as postcards and sealed letters, are supplied collectively and in a standing position. The postal matter P supplied from the supply unit 3 is conveyed to the take-out unit 5, and the take-out unit 5 takes out the postal matter P one by one and conveys it to the take-in conveyance path 7.
The take-in / conveyance path 7 is composed of a belt and a plurality of rollers, and nips and conveys the postal matter P.

An address recognition device 9 is installed on the upstream side of the take-in / transport path 7. The address recognition device 9 is a word recognition device provided with an address discrimination function, and reads the address information described in the mail P and outputs the classification designation data.
The address recognition device 9 will be described in detail later.

The postal matter P passing through the address recognition device 9 is then conveyed to the sorting section 11. The sorting unit 11 is composed of, for example, a plurality of sorting boxes 13 ... Divided by address or postal code. Mail pieces P are sorted and accumulated in the sorting boxes 13 ... In accordance with the sorting designation data from the address recognition device 9 described above.

The delivery of the postal matter P to this sorting box 13 ...
This is performed by the section conveyance paths 15a to 15g. Next, the configuration of the address recognition device 9 of the present embodiment will be described with reference to FIG. FIG. 3 is a block diagram of the address recognition device according to the present embodiment.

The address recognition device 9 comprises a word recognition section 17, an address recognition circuit 19 and an address dictionary 21. The word recognition unit 17 includes a photoelectric conversion unit 23. The photoelectric conversion unit 23 is for irradiating the postal matter P with light by a light source (not shown), receiving the reflected light, and converting the reflected light into an electric signal.

The binarization circuit 25 binarizes the read signal of the image of the entire surface of the postal matter P output from the photoelectric conversion unit 23 with a predetermined threshold value. The binarized signal of the image of the entire surface of the postal matter P binarized by the binarization circuit 25 is output to the address area detection unit 27.

The address area detection unit 27 specifies the address area in the image of the mail P by coordinates by the projection method or the like,
It is something to detect. After the detection, a signal indicating that the predetermined range is the address area is output.

The selection circuit 29 inputs the output signal from the address area detection unit 27 and selectively outputs only the binarized signal of the address area among the binarized signals input from the binarization circuit 25. .

The character detection / cutout circuit 31 detects, for example, a line from the binarized signal in the address area input from the selection circuit 29.
The characters are cut out, characters are detected character by character according to the estimated pitch, and the characters are cut out and a signal for each character is output.

The normalization circuit 33 includes a character detection / cutout circuit 31.
Each character is normalized to a certain size from the signal input from. The signal normalized here is output to the character recognition circuit 35.

The character recognition circuit 35 is connected to a character dictionary 37, and a large number of characters are registered in this character dictionary 37. The character recognition circuit 35 refers to the character dictionary 37 to perform character recognition. This character recognition is performed by matching with a reference pattern of a character registered in the character dictionary 37, for example, and character candidates are listed. Since the same character candidates as the character candidates described with reference to FIG. 2 can be mentioned by the character recognition of this embodiment, the description of the character candidates will be omitted. The character recognition circuit 39 outputs the information of the character candidates to the word recognition circuit 39.

The word recognition circuit 39 is connected to a word dictionary 41, and a large number of words are registered in this word dictionary 41. In the word recognition circuit 39, this word dictionary 41
The word recognition is performed with reference to to list word candidates. The details of the word dictionary 41 and the flow of word recognition will be described later. The word recognition circuit 39 outputs the information of this word candidate to the address recognition circuit 19.

Although the address recognition circuit 19 is connected to the address dictionary 21, the existing address is registered in the address dictionary 21. The address recognition circuit 19 refers to the address dictionary 21 to recognize an address from the word candidates and outputs the classification designation data. This sorting designation data is used for sorting and stacking the sorting boxes 13 in the mail sorting machine 1 as described above.

Next, referring to FIG. 4, the character candidates picked up by the character recognition circuit 35 of this embodiment will be described. Figure 4
[FIG. 6] is a drawing showing character candidates of the present embodiment. Figure 4
As a result of character recognition of 5 characters, the character candidates listed are shown.
The character of the rank candidate is “K” and the score is 100, the character of the second rank is “K” and the score is 90, and the character and the score are determined for each rank. The Nth candidate is the character “SA”. Score 2
It is 0. Similarly, a character and a score are determined for each of the character numbers 2, 3, 4, and 5.

The correlation between this character and the score is not constant. That is, although the character number is 1 and the second candidate is the character “ke” and the score is 90, the character number is 2
The number 2 candidate is also "Ke" but the score is 9
As shown in No. 5, even if the same character is listed as a candidate, the score differs depending on the character to be recognized.

Next, the word dictionary 41 of this embodiment will be described with reference to FIG. FIG. 5 is an example of a word dictionary used in word recognition according to the present embodiment. As shown in FIG. 5, in the word dictionary 41, a character string, a character threshold value, a maximum non-matching character number, and a threshold value are determined corresponding to each word number and registered as information. That is, if the word number is A, "Kawasaki" is registered as the character string, the character threshold value is 50 corresponding to each character, and the maximum non-matching character number is 0, and the threshold value is 250. It is registered. Similarly, the word number B is a character string, a character threshold value, a maximum number of non-matching characters, and a threshold value of "Kawasakiku" and 5 respectively.
0, 0, 250, and the word number C are "Cywaik",
0, 1, and 250 are registered.

The character threshold value is determined corresponding to each character string and is compared with the score of the character candidate. That is, it is determined that a character candidate having a score equal to or higher than this character threshold matches the word dictionary 41, but a character dictionary having a score lower than this character threshold has a word dictionary 41. No comparison is made with.

This character threshold value is uniformly set to 50 for all characters here, but of course it may be a different value for each character depending on the degree of erroneous recognition at the time of word recognition. .

Next, the operation of word recognition by the word recognition circuit 39 of this embodiment will be described with reference to FIGS. The present embodiment will be described by taking the case where five characters are recognized as one word.

FIG. 1 is a flowchart showing the word recognition method of this embodiment. First, as a premise, this method is
The processing is performed for each word number, and the processing is not performed in parallel.

First, as initialization, the total character evaluation value (total score) is set to 0, and the number of mismatched characters is set to 0 (step 1). Then, among the character strings registered in the word dictionary 41, i
Retrieve the th character (step 2). That is, when recognizing a word, since the characters are sequentially extracted from the first registered character, it will be described below that i is 1. However, in the case of the word number A shown in FIG.

Then, the jth character candidate is taken out of the character candidates (step 3). That is, step 1
The j-th character candidate of the character corresponding to the character listed in (1) is extracted. Since it is necessary to select a certain character to form a word, it will be explained below that j is 1 by sequentially recognizing from the first candidate, but since the first character was mentioned in step 1. , The first candidate corresponding to this character is listed. In the character candidates shown in FIG. 4, "ka" is taken out as the first candidate.

Thereafter, the character threshold value extracted from the word dictionary 41 in step 2 is compared with the score of the character candidate extracted in step 3 (step 4). That is, it is determined whether the score of the character candidate is equal to or higher than the character threshold value from the word dictionary 41.

When it is recognized in step 4 that the value is equal to or more than the threshold value, it is determined whether these characters match (step 5). If it is determined that they match, the score of the jth character candidate is added as the character evaluation value as the character evaluation value of the i-th character (step 6). In this example, the first character candidate “K” has a score of 100.
And the character threshold value is 50 for the first character “KA” of the word number A in the word dictionary 41, the step 4,
Stepping 5 and 6, 100 is added as the character evaluation value.
This total character evaluation value is stored in the word recognition circuit 39 for each word number.

After this step 6, the process returns to step 2 and the character string of the succeeding word dictionary 41 is taken out. That is,
In the case of the word number A shown in FIG. 5, "wa" is taken out. If the characters do not match as a result of step 5, the subsequent character candidates are extracted in step 3. That is, the j + 1th character candidate is taken out, but in the drawing of the character candidate shown in FIG. 4, “ke” is taken out and compared with the threshold value in step 4, and whether or not the character matches in step 5 is determined. Is judged.

While repeating steps 3, 4 and 5, it is determined in step 4 that the character threshold value of the i-th character listed in step 3 is smaller than the score of the j-th character candidate. If the i-th character is a non-matching character, 1 is added to the number of non-matching characters (step 7). The addition of the non-matching characters is also performed when the score of the character candidate is equal to or higher than the character threshold in steps 3, 4 and 5, but there is no matching character as a result of comparison up to the Nth candidate. The total number of unmatched characters is stored in the word recognition circuit 39 for each word number.

Subsequently, it is judged whether or not the total number of the maximum non-matching characters added in step 7 is less than or equal to the predetermined non-matching character number in the word dictionary 41 (step 8). If it is determined in step 8 that the number of characters is less than or equal to the predetermined maximum number of non-matching characters, the process returns to step 2 and the subsequent character string in the word dictionary 41 is extracted.

If it is determined in step 8 that the number of characters is larger than the predetermined maximum number of non-matching characters, it is determined that no word candidate is obtained (step 9). Thereafter, the word number thus judged is not judged as a character.

In this way, when the processing is completed up to the end of the character string in the word dictionary 41 in step 2, the total number of non-matching characters of the characters listed in this character candidate is compared with the maximum number of non-matching characters predetermined in the word dictionary. Is performed (step 10).

If the result of step 10 is that the total is less than or equal to the maximum number of non-matching characters, then the total character evaluation value is compared with a threshold value predetermined in the word dictionary (step 1
1). As a result of step 11, if the total character evaluation value is equal to or more than the threshold value, it is assumed that a word candidate is obtained (step 1
2), the process ends.

If the result of step 10 is that the total is larger than the maximum number of non-matching characters, and if the result of step 11 is that the total character evaluation value is smaller than the threshold value, then no word candidate is obtained in step 9, The process ends.

Thereafter, the same processing is subsequently performed for the word number B. According to such a method, for example, in consideration of the processing of the word number C, the first "sa" and the second "a" in the character string are in the character candidates, but the scores are 20 respectively. And 30, which are smaller than the threshold value 50, are counted as the number of non-matching characters in step 7, respectively. That is, the number of unmatched characters is 2, and the word number C
Since the maximum number of non-matching characters is larger than 1, it is determined that C cannot be obtained as a word candidate at the stage of selecting the second character, and it is not necessary to access subsequent character strings. That is, for the word number C, only the number of stored non-matching characters needs to be determined.

In this way, the address recognition circuit 19 selects a word suitable as an address from the listed candidates for words. Further, the word can be uniquely specified depending on the threshold value or the value of the maximum number of non-matching characters.
The process at 9 becomes faster.

In this way, the character threshold value is set for each character, and this is set as one of the judgment criteria. Further, the number of non-matching characters is counted for each character, and the maximum number of non-matching characters defined in the word dictionary is compared. Therefore, word recognition is performed quickly.
The present invention is not limited to the above embodiment, but can be variously modified without changing the gist of the present invention.

[0051]

As described above, according to the present invention, the character threshold value is set for each character and this is set as one of the judgment criteria.
Furthermore, since the number of non-matching characters is counted for each character and compared with the maximum number of non-matching characters defined in the word dictionary, word recognition is performed quickly.

[Brief description of drawings]

FIG. 1 is a flowchart showing a word recognition method according to this embodiment.

FIG. 2 is a front view of a mail sorting machine according to the present embodiment.

FIG. 3 is a block configuration diagram of an address recognition device according to the present embodiment.

FIG. 4 is a diagram showing character candidates according to the present embodiment.

FIG. 5 is an example of a word dictionary used in word recognition according to the present embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Mail sorter 9 Address recognition device 17 Word recognition unit 35 Character recognition circuit 37 Character dictionary 39 Word recognition circuit 41 Word dictionary

Claims (5)

[Claims] 1. A reading unit for reading a word written on a paper sheet, and a plurality of character candidates for recognizing the word of the paper sheet read by the reading unit for each character and giving an evaluation value for each character. A memory for storing a plurality of words to be described on the paper sheet, an allowable limit value of the number of mismatched characters corresponding to each of these words, and a threshold value determined for each character of the word. Means, comparing means for comparing each character of the word stored in the storing means with the character candidate output by the character recognizing means, and the comparing means for comparing the characters with the character When the evaluation value of the character candidate output from the recognition means is smaller than the threshold value of the corresponding character stored in the storage means, the counting means counts the character as a non-matching character, and the result of the counting by the counting means. , Word recognition device characterized in that the count value is composed of, and discriminating means for discriminating a non-corresponding word when exceeds the allowable limit value of the stored mismatched characters in the storage means. 2. A reading means for reading a word written on a paper sheet, and a plurality of character candidates for recognizing each word of the paper sheet read by the reading means and giving an evaluation value for each character. A character recognition means for outputting, a storage means for storing a plurality of words to be written on the paper sheet and a threshold value determined for each character of these words, and a word stored in this storage means Comparing means for comparing whether or not each character of the character candidate matches the character candidate output from the character recognizing means, and the evaluation value of the character candidate stored in the storing means when the comparison is performed by the comparing means. A word recognition device comprising: a discriminating means for discriminating a non-matching character when it is smaller than a threshold value of the character to be recognized. 3. A reading unit for reading a word written on a paper sheet, and a plurality of character candidates for recognizing the word of the paper sheet read by the reading unit for each character and giving an evaluation value for each character. A character recognition means for outputting, a storage means for storing a plurality of words to be described on the paper sheet and an allowable limit value of the number of non-matching characters corresponding to each of the words, and a storage means for storing the words stored in the storage means. A determination unit that determines whether or not there is a character candidate that matches the word among the character candidates by sequentially comparing the plurality of corresponding character candidates from the first character, and the result of the determination by the determination unit, If there is a matching character candidate, the word recognition means for recognizing as a character forming a word, and if there is no matching character candidate as a result of the judgment by the judging means, the number of characters is regarded as a non-matching character Counting means for counting, and as a result of counting by the counting means, a judging means for recognizing that the word cannot be recognized when the counted value exceeds the allowable limit value of the number of non-matching characters stored in the storage means. And a word recognition device. 4. A reading unit for reading a word written on a paper sheet, and a plurality of character candidates for recognizing the word of the paper sheet read by the reading unit for each character and giving an evaluation value for each character. A memory for storing a plurality of words to be described on the paper sheet, an allowable limit value of the number of mismatched characters corresponding to each of these words, and a threshold value determined for each character of the word. Means, comparing means for comparing each character of the word stored in the storing means with the character candidate output by the character recognizing means, and the result of the comparison by the comparing means, the word and the character When the candidates match, the evaluation value of the character candidate output from the character recognition means is stored in the storage means when the address recognition means for recognizing the address from this word and the comparison means are compared. A pair When the count value is smaller than the threshold value of the characters to be counted, the counting means counts the characters as non-matching characters, and as a result of the counting by the counting means, when the count value exceeds the allowable limit value of the number of non-matching characters stored in the storage means. An address recognition device comprising: a determination unit that determines that the word is not a corresponding word. 5. A word recognition method for recognizing a word from a word stored in a word dictionary and a candidate for each character constituting the word obtained by character recognition, wherein the word is previously stored in the word dictionary and each character of this word is recognized. A first step of storing a predetermined threshold value and a permissible limit value of the number of non-matching characters of the word; a second step of reading the word written in the medium; and a step of reading the word of the medium read by the second step. A third step of recognizing each character and outputting a character candidate to which an evaluation value is given for each character; a threshold value for each character of the words in the word dictionary stored in the first step; A fourth step of comparing with the evaluation value for each character of the character candidates output in the third step; and of the word for which the evaluation value is judged to be equal to or greater than the threshold value by the fourth step. Letters and A fifth step of determining whether or not the character candidates match, and when the evaluation value is determined to be smaller than the threshold value by the fifth step, the character of the word is regarded as a non-matching character. A sixth step of counting, and a seventh step of judging whether or not the number of non-matching characters counted by the sixth step is larger than an allowable limit value of the number of non-matching characters stored in the first step. And, as a result of the judgment in the seventh step, if the number of non-matching characters is larger than the allowable limit value of the number of non-matching characters, the eighth step of judging that this word is not a corresponding word, How to recognize words.