JPH08272906A - Character recognition device - Google Patents
Character recognition deviceInfo
- Publication number
- JPH08272906A JPH08272906A JP7075010A JP7501095A JPH08272906A JP H08272906 A JPH08272906 A JP H08272906A JP 7075010 A JP7075010 A JP 7075010A JP 7501095 A JP7501095 A JP 7501095A JP H08272906 A JPH08272906 A JP H08272906A
- Authority
- JP
- Japan
- Prior art keywords
- recognition
- character
- character recognition
- processor
- processing
- 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
Links
Landscapes
- Character Discrimination (AREA)
Abstract
(57)【要約】
【目的】高性能の文字認識手段を必要とすることなく文
字認識精度を向上する文字認識装置を提供することにあ
る。
【構成】帳票のイメ−ジデータを画像メモリ１６に読取
り、認識プロセッサＡ１８で順次文字認識を行う。認識
プロセッサＡ１８で不読となった文字は、認識プロセッ
サＡ１８と異なる文字認識手法で文字認識を行う認識プ
ロセッサＢで文字認識を行う。異なる認識手法で不読に
なった同一文字パターンに対して認識処理を施すことに
より文字認識精度を向上する。
(57) [Abstract] [Purpose] To provide a character recognition device that improves character recognition accuracy without requiring a high-performance character recognition means. [Structure] Image data of a form is read into the image memory 16 and the recognition processor A18 sequentially recognizes characters. Characters that have become unreadable by the recognition processor A18 are recognized by a recognition processor B that performs character recognition using a character recognition method different from that of the recognition processor A18. Character recognition accuracy is improved by performing recognition processing on the same character pattern that becomes unreadable by different recognition methods.
Description
【０００１】[0001]
【産業上の利用分野】本発明は、帳票のイメ−ジデータ
を読取り、文字認識を行う文字認識装置に関する。BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition device for reading image data of a form and performing character recognition.
【０００２】[0002]
【従来の技術】従来、文字認識装置においては、単一の
手法で文字認識を行うことが通例であった。また複数の
認識手法を持ち、複数の認識手段を並列に動作させる文
字認識装置が、特開平６−１６２２６２で提案されてい
る。2. Description of the Related Art Conventionally, in a character recognition device, it has been customary to perform character recognition by a single method. A character recognition device having a plurality of recognition methods and operating a plurality of recognition means in parallel has been proposed in Japanese Patent Laid-Open No. Hei 6-162262.
【０００３】[0003]
【発明が解決しようとする課題】上記従来技術は、単一
の文字認識手法による文字認識処理ため、その手法にそ
ぐわない字形の文字にたいしては不読になりやすい欠点
があった。また複数の認識手法を持ち、複数の認識手段
を並列に動作させることにより、処理速度を低下するこ
となく文字認識率を向上させる文字認識装置も提案され
ているが、この提案では同一性能を持つ複数の認識手段
を持つ必要があり装置が高価になる欠点があった。The above-mentioned prior art has a drawback in that since the character recognition processing is performed by a single character recognition method, a character having a character shape that does not match the method is likely to be unreadable. A character recognition device that has a plurality of recognition methods and operates a plurality of recognition means in parallel to improve the character recognition rate without lowering the processing speed has also been proposed, but this proposal has the same performance. It is necessary to have a plurality of recognition means, and there is a drawback that the device becomes expensive.
【０００４】また、文字認識処理と文字きりだし処理等
の独立した異なる処理を複数のプロセッサで処理する案
も提案されているが、この提案では、一般的に文字認識
処理のプロセスを担当するプロセッサの負荷が大きくな
り、プロセッサ間で稼働率のアンバランスが生じる問題
があった。Further, a proposal has been proposed in which a plurality of processors perform different independent processing such as character recognition processing and character extraction processing. In this proposal, a processor which is generally in charge of the character recognition processing is proposed. There was a problem that the load on the server became large and the operating rate was unbalanced among the processors.
【０００５】本発明の目的は、高性能の文字認識手段を
必要とすることなく文字認識精度を向上する文字認識装
置を提供することにある。An object of the present invention is to provide a character recognition device which improves character recognition accuracy without requiring a high-performance character recognition means.
【０００６】[0006]
【課題を解決するための手段】上記目的を達成するた
め、本発明では、異なる文字認識手法で文字認識を行う
複数の文字認識手段を備え、一つの文字認識手段で認識
処理を行い、不読となった文字のみを他の文字認識手段
で認識処理を行う。In order to achieve the above object, the present invention comprises a plurality of character recognition means for performing character recognition by different character recognition methods, and the recognition processing is performed by one character recognition means, and unreadable. Only the characters that have become are recognized by another character recognition means.
【０００７】[0007]
【作用】本発明によれば、第１の認識手法で不読となっ
た文字に対してのみ、第２の認識手法で再度認識処理を
行い、認識結果を求める方法を用いることにより、第２
の認識手法で行う上記他の文字認識手段で認識処理を実
行する回数を減らすことができるため、上記他の文字認
識手段は第１の認識手法で行う文字認識手段に比べ安価
なハードウエアで、文字の認識率を向上させることが可
能となる。According to the present invention, the second recognition method is used to perform the recognition process again for only the characters that are unreadable by the first recognition method, and the second recognition method is used.
Since it is possible to reduce the number of times the recognition processing is performed by the other character recognition means performed by the other recognition method, the other character recognition means is less expensive hardware than the character recognition means performed by the first recognition method. It is possible to improve the character recognition rate.
【０００８】また、第２の認識手段で行う認識処理は、
処理の負荷が少ないため、上記他の文字認識手段では他
の処理との同時動作が可能となる。The recognition processing performed by the second recognition means is
Since the processing load is small, the other character recognition means can simultaneously operate with other processing.
【０００９】[0009]
【実施例】以下、本発明の一実施例を図面を参照して詳
細に説明する。図１は本発明の一実施例を全体のブロッ
ク図、図２は認識対象となる帳票の例を示す図、図３は
本発明の一実施例における処理の流れを示す図である。DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings. 1 is an overall block diagram of an embodiment of the present invention, FIG. 2 is a diagram showing an example of a form to be recognized, and FIG. 3 is a diagram showing a flow of processing in an embodiment of the present invention.
【００１０】図１において、文字認識装置１１は制御プ
ロセッサ１７を備え、文字認識装置１１全体の制御を行
う。文字認識装置１１は、認識フィールドの位置座標が
格納されているフォーマット情報１２を文字認識装置１
１内のワークメモリ１５にダウンロードする。次に、フ
ァクシミリ装置（ＦＡＸ）等の帳票入力装置により光電
変換された帳票のイメージデータ１３をインタフェース
（Ｉ／Ｆ）制御部１４を介して画像メモリ１６に格納す
る。画像メモリ１６に格納された帳票イメージデータ１
３と、フォーマット情報１２から、認識プロセッサＡ１
８および認識プロセッサＢ１９により、認識フィールド
位置を抽出し、文字切り出し処理、文字認識処理を行
い、認識結果をワークメモリ１５に格納後、Ｉ／Ｆ制御
部１４を介し上位装置に転送するものである。In FIG. 1, the character recognition device 11 includes a control processor 17 to control the entire character recognition device 11. The character recognition device 11 stores the format information 12 in which the position coordinates of the recognition field are stored in the character recognition device 1.
1 is downloaded to the work memory 15 inside. Next, the image data 13 of the form photoelectrically converted by the form input device such as a facsimile machine (FAX) is stored in the image memory 16 via the interface (I / F) control unit 14. Form image data 1 stored in the image memory 16
3 and the format information 12 from the recognition processor A1
8 and the recognition processor B19 extract the recognition field position, perform character cutting processing and character recognition processing, store the recognition result in the work memory 15, and then transfer it to the higher-level device via the I / F control unit 14. .
【００１１】次に、図２、図３を用い本装置の処理の流
れの例を述べる。ワークメモリ１５に格納された認識対
象の帳票イメージデータ１３に対し、フォーマット情報
１２に基づき認識対象文字２１を切り出す。フォーマッ
ト情報１２は帳票原点２０からの認識対象文字２１の座
標Ｘ、Ｙ及び文字幅Ｗ、文字高さＨ等が格納されてい
る。このフォーマット情報１２に基づき、１文字切り出
し処理（ステップ３１）を認識プロセッサＢ１９により
処理を行い、１文字単位のイメージデータをワークメモ
リ１５に格納する。この１文字単位のイメージデータに
対し、まず認識プロセッサＡ１８により文字認識処理
（３２）を行う。認識プロセッサＡ１８による文字認識
手法は、例えば文字の輪郭抽出による文字特徴を利用し
た特徴量マッチング法による。その結果、正常に認識し
（３３）、次の文字が有れば（３５）、ステップ３１で
次に切り出された次の文字の認識処理を行う。Next, an example of the processing flow of this apparatus will be described with reference to FIGS. Based on the format information 12, the recognition target character 21 is cut out from the recognition target form image data 13 stored in the work memory 15. The format information 12 stores the coordinates X, Y of the recognition target character 21 from the form origin 20, the character width W, the character height H, and the like. Based on the format information 12, the recognition processor B19 performs a one-character cutout process (step 31), and stores image data for each character in the work memory 15. First, the recognition processor A18 performs character recognition processing (32) on the image data in units of one character. The character recognition method by the recognition processor A18 is, for example, a feature amount matching method using character features by contour extraction of characters. As a result, it is recognized normally (33), and if there is a next character (35), the recognition processing of the next character cut out in step 31 is performed.
【００１２】ステップ３２の認識プロセッサＡ１８によ
る文字認識処理の結果が不読の場合（３３）、認識プロ
セッサＡ１８は上記と同様に次の文字の認識処理を行う
が、それと並行して認識プロセッサＢ１９で認識プロセ
ッサＡにおける文字認識処理（３２）で不読となった文
字について文字認識処理（３４）を行う。認識プロセッ
サＢ１９による文字認識手法は、認識プロセッサＡ１８
による文字認識手法とは異なる認識手法、例えば、文字
のストローク特徴を用いた認識手法にて文字認識処理
（３４）を行う。その結果、正常に認識すれば、その文
字コードを認識結果とするものである。When the result of the character recognition processing by the recognition processor A18 in step 32 is unreadable (33), the recognition processor A18 performs the recognition processing of the next character in the same manner as described above, and in parallel with this, the recognition processor B19 performs the recognition processing. The character recognition process (34) is performed on the character that has become unreadable in the character recognition process (32) in the recognition processor A. The character recognition method by the recognition processor B19 is the recognition processor A18.
The character recognition processing (34) is performed by a recognition method different from the character recognition method by, for example, a recognition method using the stroke feature of the character. As a result, if the character is recognized normally, the character code is used as the recognition result.
【００１３】認識プロセッサＢ１９は、通常、次文字の
文字切り出し処理（３１）を行っているが、文字切り出
し処理（３１）は文字認識処理に比べ処理量が少ないた
め、認識プロセッサＢ１９の空いている時間を利用し
て、文字認識処理３４を実行することが可能となる。ま
た、仮に認識プロセッサＢ１９が文字切り出し処理３１
の実行中であっても、例えば割込み処理により文字認識
処理（３４）を優先して実行させることも可能である。The recognition processor B19 normally performs a character cutting process (31) for the next character. However, since the character cutting process (31) has a smaller processing amount than the character recognition process, the recognition processor B19 is free. It is possible to execute the character recognition processing 34 by using the time. Further, if the recognition processor B19 performs the character segmentation process 31
It is also possible to preferentially execute the character recognition process (34) by, for example, an interrupt process even during execution of.
【００１４】本公知例では、２つの認識処理手段を用い
で認識処理を行う例を示したが、２つ以上の認識処理手
段を設け、不読文字に対して、複数段の認識処理を施す
ことも可能である。In this known example, the recognition processing is performed by using two recognition processing means. However, two or more recognition processing means are provided, and the recognition processing is performed in a plurality of stages for unread characters. It is also possible.
【００１５】また、文字認識処理３４の同時処理として
文字認識の前処理である文字切り出し処理の例を示した
が、文字認識処理に比べ、処理負荷の少ない処理、例え
ば単語照合等を用いることも可能である。Further, as an example of the character cutting process which is a pre-process of the character recognition as the simultaneous process of the character recognition process 34, a process having a smaller processing load than the character recognition process, for example, a word matching process may be used. It is possible.
【００１６】[0016]
【発明の効果】本発明によれば、従来の認識プロセッサ
に、安価な認識プロセッサと手法の異なる認識処理を追
加し、従来の認識プロセッサで不読になった文字につい
てのみ再認識を行うことにより、認識速度を犠牲にする
ことなく文字認識率の向上を図る事が可能となる。ま
た、不読文字のみ再認識処理を行ない認識処理の負荷を
軽減することにより、全体システムの処理時間を増やす
ことなく、他の処理を行うプロセッサを用いて認識処理
を行うことが可能となる。According to the present invention, a recognition process of a different method from that of an inexpensive recognition processor is added to a conventional recognition processor, and only a character which is unreadable by the conventional recognition processor is recognized again. It is possible to improve the character recognition rate without sacrificing the recognition speed. Further, by performing the re-recognition process only on the unreadable characters and reducing the load of the recognition process, it is possible to perform the recognition process using the processor that performs other processes without increasing the processing time of the entire system.
【図１】本発明の一実施例を示すブロック図。FIG. 1 is a block diagram showing an embodiment of the present invention.
【図２】認識対象帳票の例を示す図。FIG. 2 is a diagram showing an example of a recognition target form.
【図３】本発明の一実施例の処理の流れを示す図。FIG. 3 is a diagram showing a flow of processing according to an embodiment of the present invention.
１１…文字認識装置、１７…制御プロセッサ、１８…認
識プロセッサＡ、１９…認識プロセッサＢ、11 ... Character recognition device, 17 ... Control processor, 18 ... Recognition processor A, 19 ... Recognition processor B,
Claims (2)
を行う文字認識装置において、異なる文字認識手法で文
字認識を行う複数の文字認識手段を備え、一つの文字認
識手段で認識処理を行い、不読となった文字のみを他の
文字認識手段で認識処理を行うことを特徴とする文字認
識装置。1. A character recognition device for reading image data of a form and performing character recognition, comprising a plurality of character recognition means for performing character recognition by different character recognition methods, and performing recognition processing by one character recognition means, A character recognition device characterized in that only unreadable characters are recognized by another character recognition means.
であり、上記他の文字認識手段に対応するプロセッサ
で、文字認識のための前処理も実行することを特徴とす
る請求項１記載の文字認識装置。2. The character recognition according to claim 1, wherein each of the character recognition means is a processor, and the processor corresponding to the other character recognition means also executes preprocessing for character recognition. apparatus.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP7075010A JPH08272906A (en) | 1995-03-31 | 1995-03-31 | Character recognition device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP7075010A JPH08272906A (en) | 1995-03-31 | 1995-03-31 | Character recognition device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| JPH08272906A true JPH08272906A (en) | 1996-10-18 |
Family
ID=13563799
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP7075010A Pending JPH08272906A (en) | 1995-03-31 | 1995-03-31 | Character recognition device |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPH08272906A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105261110A (en) * | 2015-10-26 | 2016-01-20 | 江苏国光信息产业股份有限公司 | Efficient DSP banknote serial number recognizing method |
-
1995
- 1995-03-31 JP JP7075010A patent/JPH08272906A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105261110A (en) * | 2015-10-26 | 2016-01-20 | 江苏国光信息产业股份有限公司 | Efficient DSP banknote serial number recognizing method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP0176672B1 (en) | Method and system for line-thinning process with variable scanning areas | |
| JPH08272906A (en) | Character recognition device | |
| JP3904397B2 (en) | Table recognition method | |
| JPH0713994A (en) | Character recognition device | |
| JPH04367986A (en) | Rectangle extraction method and rectangle extraction device | |
| JP2940974B2 (en) | Matching method | |
| JPS62257583A (en) | Character recognizing system | |
| JP2838556B2 (en) | Image processing device | |
| JPH09179935A (en) | Character recognition device and control method therefor | |
| JPH1166232A (en) | Form definition method | |
| JP2746345B2 (en) | Post-processing method for character recognition | |
| JP2665226B2 (en) | Character recognition device | |
| JP4209511B2 (en) | Character recognition method, character recognition device, and computer-readable recording medium recording a program for causing a computer to execute the character recognition method | |
| JP4580520B2 (en) | Character recognition method and character recognition apparatus | |
| JPH11126216A (en) | Automatic drawing input device | |
| JP2789835B2 (en) | Table frame detection method and device | |
| JPH07104940B2 (en) | Figure recognition device | |
| JP2925239B2 (en) | Image processing method and apparatus | |
| JP2000048191A (en) | Image processing method and medium recording image processing program | |
| JPH01142890A (en) | Preprocessing unit | |
| JPH06259602A (en) | Candidate character extraction device | |
| JPH03212783A (en) | Matching method | |
| JPH0433074B2 (en) | ||
| JPS62224871A (en) | Document picture processing system | |
| JP2001126030A (en) | Pattern recognition method and pattern recognition device |