JPS63316288A - Character recognizing device - Google Patents

Character recognizing device

Info

Publication number
JPS63316288A
JPS63316288A JP62151373A JP15137387A JPS63316288A JP S63316288 A JPS63316288 A JP S63316288A JP 62151373 A JP62151373 A JP 62151373A JP 15137387 A JP15137387 A JP 15137387A JP S63316288 A JPS63316288 A JP S63316288A
Authority
JP
Japan
Prior art keywords
character
characters
recognition
result
memory
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
JP62151373A
Other languages
Japanese (ja)
Inventor
Yuji Shinozaki
祐司 篠崎
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
NEC Corp
Original Assignee
NEC Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by NEC Corp filed Critical NEC Corp
Priority to JP62151373A priority Critical patent/JPS63316288A/en
Publication of JPS63316288A publication Critical patent/JPS63316288A/en
Pending legal-status Critical Current

Links

Abstract

PURPOSE:To improve a reading efficiency and to lower an erroneous reading ratio by segmenting a character by one character, comparing and synthesizing a recognized result for every one character with a result obtained by segmenting more than two characters together and recognizing as the character of more than two characters. CONSTITUTION:A control part 1 transmits a control signal (a) to a character segmenting circuit 2 to segment the one character, transmit a control signal (c) and store in the recognition result of the one character in a memory 11. Subsequently, the control signals (a), (c) are generated to recognize a second character and store in the recognition result in the memory 11. At this time, since the input pattern of a memory 4 is a pattern in which the two characters are connected, the control signals (f), (b), (c) are generated to segment and recognize the two characters. The control part 1 generates a control signal (e) and a general decision part 12 segments every one character at the time of setting a switch 13 to an A mode and segments the characters together and defines the recognized result to be a decision result when a decision is impossible. At the time of setting the switch 13 to a B mode, when the results for segmenting the one character and the two characters coincide, it is defined to be the decision result.

Description

【発明の詳細な説明】 〔産業上の利用分野〕 本発明は文字認識装置に関するものである。[Detailed description of the invention] [Industrial application field] The present invention relates to a character recognition device.

〔従来の技術〕[Conventional technology]

従来の文字認識装置は、2文字以上くっついて記載され
た文字を認識する場合などでも、文字を1文字ずつに区
切って切り出し、1文字ずつの文字として認識処理を行
なっていた。
Conventional character recognition devices, even when recognizing two or more characters written together, separate the characters into individual characters and perform recognition processing on each character.

〔発明が解決しようとする問題点〕[Problem that the invention seeks to solve]

上述した従来の文字認識装置は、2文字以上くっついた
文字を認識する場合などに文字を1文字ずつ区切って切
り出すため、不読(読解不可)や誤読が多く発生すると
いう欠点があった。
The above-mentioned conventional character recognition device has the disadvantage that it often causes illegibility (inability to read) and misreading because it separates characters one by one when recognizing two or more characters stuck together.

例えば、第3図に示すように、文字枠の中に数字のゼロ
を2文字くっつけて記載された文字を認識するとき、1
文字ずつ区切って切り出しそれぞれの切出し文字パター
ンについて認識を行なうために、文字同士のつながり具
合の情報が欠落し、数字のゼロ“0”をロク“6”に誤
読したり、判別できずに不読となる場合などがあった。
For example, as shown in Figure 3, when recognizing a character written with two zeros attached in a character frame, 1
Because each character is separated and recognized for each cut-out character pattern, information about the connection between characters is lost, resulting in the number zero "0" being misread as "6" or being unreadable. There were cases where this was the case.

また、上記欠点を減少させるために、第4図にに示すよ
うに、文字切出しの条件を変え、文字の切出し・認識を
複数回(同図では3回)行ない、総合的に文字の判定を
行なう手法もあったが(同図の場合読解できたものを採
用している)、文字の切出し・認識に多くの時間を要し
たり、大規模なハードウェアが必要となったりしていた
In addition, in order to reduce the above drawbacks, as shown in Figure 4, the conditions for character extraction were changed, character extraction and recognition were performed multiple times (three times in the figure), and the characters were comprehensively judged. There was a method to do this (in the case of the figure, the one that could be read was used), but it took a lot of time to extract and recognize the characters, and it required large-scale hardware.

〔問題点を解決するための手段〕[Means for solving problems]

このような問題点を解決するために本発明による文字認
識装置は、文字を1文字ずつ切り出し1文字ずつ認識す
る単数文字認識手段と、2文字以上の文字を1文字ずつ
に区切らずに2文字以上いっしょに切り出し2文字以上
の文字として認識する複数文字認識手段と、文字を1文
字ずつ切り出しI文字ずつ認識した結果と2文字以上い
っしょに切り出し2文字以上の文字として認識した結果
とから認識結果を比較合成する比較合成手段とを設ける
ようにしたものである。
In order to solve such problems, the character recognition device according to the present invention has a single character recognition means that cuts out characters one by one and recognizes them one by one, and a single character recognition means that separates two or more characters into two characters without dividing them into individual characters. The recognition result is based on the multi-character recognition means that cuts out the above characters together and recognizes them as two or more characters, the results of cutting out characters one by one and recognizing them one by one, and the results of cutting out two or more characters together and recognizing them as two or more characters. A comparison and synthesis means for comparing and synthesizing the information is provided.

〔作用〕[Effect]

本発明による文字認識装置においては、2文字以上くっ
ついた文字を認識する場合に、読取り率が向上し、誤読
率が低下する。
In the character recognition device according to the present invention, when recognizing two or more characters stuck together, the reading rate improves and the misreading rate decreases.

〔実施例〕 第1図は、本発明に係わる文字認識装置の一実施例の要
部を示す概略系統図である。第1図において、1は装置
全体を制御する制御部であり、制御部lは制御信号a、
b、c、d、eを発生することができる。2.3は文字
切出回路で2.メモリ4.5にそれぞれハス6.7を介
し接続されている。文字切出回路2は、制御部1からの
制御0信号aにより、メモリ4に格納された人力文字イ
メージから文字を1文字分切り出し、切出しパターンを
メモリ5に格納する。この切り出した文字が2文字以上
つながっていた場合は、上記のように切出しパターンの
格納を行なうと共に、識別信号fを発生する。文字切出
回路3は制御信号すによりメモリ4に入力された入力文
字イメージから文字を2文字分切り出し、切出しパター
ンをメモリ5に格納する。
[Embodiment] FIG. 1 is a schematic system diagram showing essential parts of an embodiment of a character recognition device according to the present invention. In FIG. 1, 1 is a control unit that controls the entire device, and the control unit l is a control signal a,
b, c, d, and e can be generated. 2.3 is a character cutting circuit. Each memory 4.5 is connected via a bus 6.7. The character cutting circuit 2 cuts out one character from the human character image stored in the memory 4 in response to the control 0 signal a from the control unit 1, and stores the cutting pattern in the memory 5. If two or more of the cut-out characters are connected, the cut-out pattern is stored as described above, and an identification signal f is generated. The character cutting circuit 3 cuts out two characters from the input character image input to the memory 4 in response to a control signal, and stores the cutting pattern in the memory 5.

8.9はメモリ5に格納された切出しパターンをバス7
を介して入力し認識する回路で、それぞれ、制御部1か
らの制御信号c、  dによって起動し、認識結果をバ
ス10を介しメモリ11に格納する。認識回路8は1文
字ずつ切り出された文字を認識するための認識辞書81
を内蔵しており、1文字ずつ切り出された文字を認識す
るために使用する。認識回路9は2文字分切り出された
文字を認識するための認識辞書91を内蔵しており、2
文字くっついた文字を認識するために使用する。
8.9 transfers the cutting pattern stored in the memory 5 to the bus 7.
These circuits are activated by control signals c and d from the control unit 1, respectively, and store the recognition results in the memory 11 via the bus 10. The recognition circuit 8 includes a recognition dictionary 81 for recognizing characters cut out one by one.
It has a built-in and is used to recognize characters cut out one by one. The recognition circuit 9 has a built-in recognition dictionary 91 for recognizing characters cut out for two characters.
Used to recognize characters that are stuck together.

前記文字切出回路2と認識回路8は単数文字認識手段を
構成し、前記文字切出回路3と認識回路9は複数文字認
識手段を構成する。
The character extraction circuit 2 and the recognition circuit 8 constitute a single character recognition means, and the character extraction circuit 3 and the recognition circuit 9 constitute a plural character recognition means.

また、バス10には、メモリ11に格納された。Further, the information was stored in the memory 11 on the bus 10.

認識結果を総合的に判定する比較合成手段としての総合
判定部12が接続されており、制御部lからの制御信号
eにより起動される。総合判定部8にはスイッチ13が
接続されており、モードA。
A comprehensive determination section 12 is connected as a comparison and synthesis means for comprehensively determining recognition results, and is activated by a control signal e from a control section l. A switch 13 is connected to the comprehensive determination section 8, and mode A is selected.

モードBの2つのモードの切替えができる。モードAは
通常のモード、モードBは高精度のモードを意味する。
You can switch between two modes: Mode B. Mode A means normal mode, and mode B means high precision mode.

次に、第1図に示す装置の動作について説明する。第2
図は、本装置で認識する文字パターンの例を示す文字パ
ターン図である。同図において、数字“5”と“6”2
文字がくっついた文字21が記載されている。なお、ス
イッチ13はモードAに設定されているものとする。予
めメモリ4に第2図の入カバターン21が格納されてい
るものとする。まず制御部lは文字切出回路2に対し制
御信号aを発生し、1文字分の文字の切出しを行ない、
さらに制御信号Cを発生し、1文字分の認識結果をメモ
リ11に格納する。続いて同様に制御信号a、cを発生
し、2文字目の認識を行ない、認識結果をメモリ11に
格納する。このとき、メモリ40入カバターンは2文字
がくっついているパターンであるため、1文字目の切出
しで文字切出回路2から識別信号fが発生されるので、
このときはさらに制御部lは文字切出回路3に対する制
御信号す、制御信号Cを順次発生し、2文字分の文字の
切出し、2文字の認識を行なう。第2図において、22
.23は1文字の切出しパターン、24は2文字の切出
しパターンである。切出しパターン22.23について
は認識辞書81を使用する認識回路8で認識し、切出し
パターン24については認識辞書91を使用する認識回
路9で認識する。認識結果は、切出しパターン22につ
いては5″、切出しパターン23については“?”すな
わち不読、切出しパターン24については“56“であ
る。
Next, the operation of the apparatus shown in FIG. 1 will be explained. Second
The figure is a character pattern diagram showing an example of a character pattern recognized by this device. In the same figure, the numbers “5” and “6”2
Character 21, which is a combination of letters, is written. It is assumed that the switch 13 is set to mode A. It is assumed that the input cover pattern 21 shown in FIG. 2 is stored in the memory 4 in advance. First, the control unit l generates a control signal a to the character cutting circuit 2 to cut out one character,
Furthermore, a control signal C is generated, and the recognition result for one character is stored in the memory 11. Subsequently, control signals a and c are generated in the same manner, the second character is recognized, and the recognition result is stored in the memory 11. At this time, since the cover pattern with 40 memories is a pattern in which two characters are stuck together, the identification signal f is generated from the character cutting circuit 2 when the first character is cut out.
At this time, the control unit 1 also sequentially generates control signals S and C to the character cutting circuit 3 to cut out two characters and recognize the two characters. In Figure 2, 22
.. 23 is a one-character cutting pattern, and 24 is a two-character cutting pattern. The cutout patterns 22 and 23 are recognized by the recognition circuit 8 that uses the recognition dictionary 81, and the cutout pattern 24 is recognized by the recognition circuit 9 that uses the recognition dictionary 91. The recognition results are 5'' for cutout pattern 22 and ``?'' for cutout pattern 23. ”, that is, unreadable, and for the cutout pattern 24, it is “56”.

次に、制御部1は制御信号eを発生し、総合判定を行な
う。総合判定部12は、スイッチ13がモードAに設定
されている時には、1文字ずつ切り出して認識できない
場合でも、2文字いっしょに切り出して認識できたとき
はその結果を判定結果とする。また、スイッチ13がモ
ードBに設定されている時は、1文字ずつ切り出して認
識した結果と2文字いっしょに切り出して認識した結果
とが一致した場合のみ、その結果を判定結果とし、その
他の場合は不読とする。
Next, the control section 1 generates a control signal e and performs comprehensive determination. When the switch 13 is set to mode A, the comprehensive judgment section 12 takes the result as the judgment result even if it cannot recognize each character by cutting them out one by one, but when it cuts out two characters together and recognizes them. Also, when the switch 13 is set to mode B, only when the result of recognition by cutting out one character at a time matches the result of recognition by cutting out two characters together, that result is used as the judgment result, and in other cases are not readable.

モードA、モードBのそれぞれの場合の判定の例を表1
1表2に示す。表1において、番号2゜3などの場合は
、1文字ずつ切り出し時に不読が発生しても、2文字い
っしょに切り出したときに認識できれば、その認識の結
果を判定結果とする例である。
Table 1 shows examples of determination in each case of mode A and mode B.
1 Shown in Table 2. In Table 1, in the case of numbers 2.degree.3, etc., even if unreadable characters occur when characters are extracted one by one, if they can be recognized when two characters are extracted together, the recognition result is used as the determination result.

また、表2において、番号1.2などの場合は、1文字
ずつ切り出し時に誤読となっても、判定結果が不読に修
正される例である。
Further, in Table 2, in the case of number 1.2, etc., even if a misreading occurs when cutting out one character at a time, the determination result is corrected to be unreadable.

第2図の入力例は表1の番号2に相当するため、総合判
定結果として“56”が得られる。
Since the input example in FIG. 2 corresponds to number 2 in Table 1, "56" is obtained as the overall judgment result.

以上のように本装置は、予めメモリ4に格納されている
第2図に示すような2文字くっついた入カバターンにつ
いて文字認識を行ない、認識結果をメモリ11に格納す
ることができる。なお、本装置において、予めメモリ4
に格納され1文字ずつ且つ(つつかずに記載された文字
パターンを認識し、その認識結果をメモリ11に格納す
る場合は、制御部1から制御信号a、cを順次発生すれ
ばよいことは以上までの説明で明らかである。また本装
置は、文字切出回路と文字認識回路を拡張することによ
り、3文字以上くっついた文字を認識するために応用す
ることも可能であることはいうまでもない。
As described above, this apparatus can perform character recognition on the cover pattern of two characters stuck together, as shown in FIG. Note that in this device, the memory 4
In order to recognize a character pattern written one character at a time (without touching) and store the recognition result in the memory 11, it is sufficient to sequentially generate control signals a and c from the control unit 1. This is clear from the above explanation.It goes without saying that this device can also be applied to recognize characters that are three or more characters together by expanding the character extraction circuit and character recognition circuit. do not have.

さらに本装置においては、スイッチ7をモードAに設定
しておくと、1文字ずつ切り出して認識できない場合で
も、2文字いっしょに切り出して認識できたときは、そ
の結果を判定結果とするため、結果として読取り率を向
上させることができる。また、スイッチ7をモードBに
設定しておくと、1文字ずつ切り出して認識した結果と
2文字いっしょに切り出して認識した結果とが一致した
場合のみ、その結果を判定結果とし、その他の場合は不
読とするため、結果として誤読率を低下させることがで
きる。このとき文字の切出し・認識は各文字について高
々2回1行なうだけなので、文字切出しの条件を変え文
字の切出し・認識を複数回行なう手法と比較すると、処
理時間を低減できることがわかる。
Furthermore, in this device, if switch 7 is set to mode A, even if one character is extracted one by one and cannot be recognized, if two characters are extracted and recognized together, the result will be used as the judgment result. As a result, the read rate can be improved. In addition, if switch 7 is set to mode B, only when the result of recognition by cutting out one character at a time matches the result of recognition by cutting out two characters together, that result will be used as the judgment result, and in other cases, Since the information is not readable, the misreading rate can be reduced as a result. At this time, character extraction/recognition is performed only twice for each character at most, so when compared with a method in which character extraction/recognition is performed multiple times by changing the character extraction conditions, it can be seen that the processing time can be reduced.

〔発明の効果〕〔Effect of the invention〕

以上説明したように本発明は、文字を1文字ずつ切り出
し1文字ずつ認識し、2文字以上の文字を1文字ずつに
区切らずに2文字以上いっしょに切り出し2文字以上の
文字として認識し、文字を1文字ずつ切り出し1文字ず
つ認識した結果と2文字以上いっしょに切り出し2文字
以上の文字として認識した結゛果とから認識結果を比較
合成することにより、2文字以上くっついた文字を認識
する場合に種々の判断を行なうことができるので、読取
り率を向上させ、誤読率を低下させることができる効果
がある。
As explained above, the present invention cuts out characters one by one and recognizes them one by one, and cuts out two or more characters together and recognizes them as two or more characters without separating them into individual characters. When recognizing two or more characters joined together by comparing and synthesizing the recognition results from the results of cutting out one character at a time and recognizing them one by one, and the results of cutting out two or more characters together and recognizing them as two or more characters. Since various judgments can be made, the reading rate can be improved and the misreading rate can be reduced.

−Cに2文字いっしょに切り出して認識を行なう場合は
認識結果のカテゴリーの数は増加するが、読取り文字の
種類が限られカテゴリーの数も大して増加しない場合た
とえば数字のみ記載される場合などに本発明は特に有効
である。
- If two characters are extracted and recognized at the same time in C, the number of categories in the recognition result will increase, but if the types of characters that can be read are limited and the number of categories does not increase significantly, for example, when only numbers are written, the number of categories in the recognition result will increase. The invention is particularly effective.

【図面の簡単な説明】[Brief explanation of the drawing]

第1図は本発明に係わる文字認識装置の一実施例を示す
系統図、第2図はその動作を説明するための文字パター
ン図、第3図、第4図は従来の文字認識装置で2文字く
っついた文字を認識する方法を説明するための説明図で
ある。 1・・・制御部、2,3・・・文字切出回路、4,5゜
11・・・メモリ、6,7.10・・・バス、8,9・
・・認識回路、12・・・総合判定部、13・・・スイ
ッチ、81.91・・・認識辞書。
Fig. 1 is a system diagram showing an embodiment of the character recognition device according to the present invention, Fig. 2 is a character pattern diagram for explaining its operation, and Figs. FIG. 2 is an explanatory diagram for explaining a method for recognizing characters that are stuck together. DESCRIPTION OF SYMBOLS 1...Control part, 2,3...Character cutting circuit, 4,5゜11...Memory, 6,7.10...Bus, 8,9.
... Recognition circuit, 12... Comprehensive judgment unit, 13... Switch, 81.91... Recognition dictionary.

Claims (1)

【特許請求の範囲】[Claims] 文字を1文字ずつ切り出し1文字ずつ認識する単数文字
認識手段と、2文字以上の文字を1文字ずつに区切らず
に2文字以上いっしょに切り出し2文字以上の文字とし
て認識する複数文字認識手段と、文字を1文字ずつ切り
出し1文字ずつ認識した結果と2文字以上いっしょに切
り出し2文字以上の文字として認識した結果とから認識
結果を比較合成する比較合成手段とを備えたことを特徴
とする文字認識装置。
A single character recognition means that cuts out characters one by one and recognizes them one by one, and a plurality of character recognition means that cuts out two or more characters together and recognizes them as two or more characters without separating them into individual characters. Character recognition characterized by comprising a comparison and synthesis means for comparing and synthesizing the recognition results from the results of cutting out characters one by one and recognizing them one by one and the results of cutting out two or more characters together and recognizing them as two or more characters. Device.
JP62151373A 1987-06-19 1987-06-19 Character recognizing device Pending JPS63316288A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP62151373A JPS63316288A (en) 1987-06-19 1987-06-19 Character recognizing device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP62151373A JPS63316288A (en) 1987-06-19 1987-06-19 Character recognizing device

Publications (1)

Publication Number Publication Date
JPS63316288A true JPS63316288A (en) 1988-12-23

Family

ID=15517141

Family Applications (1)

Application Number Title Priority Date Filing Date
JP62151373A Pending JPS63316288A (en) 1987-06-19 1987-06-19 Character recognizing device

Country Status (1)

Country Link
JP (1) JPS63316288A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6738519B1 (en) 1999-06-11 2004-05-18 Nec Corporation Character recognition apparatus
JP2006059124A (en) * 2004-08-20 2006-03-02 Hitachi Omron Terminal Solutions Corp System and apparatus for recognizing character string in landscape
JP2007058803A (en) * 2005-08-26 2007-03-08 Canon Inc Online hand-written character recognition device, and online hand-written character recognition method

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6738519B1 (en) 1999-06-11 2004-05-18 Nec Corporation Character recognition apparatus
JP2006059124A (en) * 2004-08-20 2006-03-02 Hitachi Omron Terminal Solutions Corp System and apparatus for recognizing character string in landscape
JP4713107B2 (en) * 2004-08-20 2011-06-29 日立オムロンターミナルソリューションズ株式会社 Character string recognition method and device in landscape
JP2007058803A (en) * 2005-08-26 2007-03-08 Canon Inc Online hand-written character recognition device, and online hand-written character recognition method

Similar Documents

Publication Publication Date Title
JPS63316288A (en) Character recognizing device
JPH0520794B2 (en)
KR19990049667A (en) Korean Character Recognition Method
JPH0528324A (en) English character recognition device
JPS5842904B2 (en) Handwritten kana/kanji character recognition device
JPH0527915B2 (en)
JP2746345B2 (en) Post-processing method for character recognition
JPH0338787A (en) Character recognition processor
JPH0944604A (en) Character recognizing processing method
JPS59158482A (en) Character recognizing device
JPH0475556B2 (en)
JPS58125184A (en) Optical character reader
JP2931485B2 (en) Character extraction device and method
JPH0218689A (en) Device and method for recognizing character
JPS6073796A (en) Optical character reader
JPH0475186A (en) Character reader
JPS5820075B2 (en) pattern recognition device
JPS6252691A (en) Character recognizing device
JPS6160184A (en) Optical character reader
JPH01237778A (en) Character recognizing method
JPS61194584A (en) Word recognition system
JPS60110089A (en) Character recognizer
JPS62145486A (en) Character recognition system
JPH01166187A (en) Method for recognizing character
JPH07220028A (en) Character recognition device