Document recognition apparatus

• US 5,621,818 A
• Filed: 07/09/1992
• Issued: 04/15/1997
• Est. Priority Date: 07/10/1991
• Status: Expired due to Term

• Current Assignee: Fuji Xerox Company Limited (Xerox Holdings Corp.)
• Original Assignee: Fuji Xerox Company Limited (Xerox Holdings Corp.)
• Inventors: Tashiro, Kiyoshi
• Primary Examiner(s): Razavi, Michael T.
• Assistant Examiner(s): Prikockis, Larry J.
• Application Number:
  US07/910,377
• Time in Patent Office: 1,741 Days
• Field of Search: 382/34, 382/1, 382/25, 382/27, 382/14, 382/37, 382/38, 382/40, 382/30, 382/218, 382/209, 382/227, 382/226, 382/229, 382/170, 382/203, 382/304
• US Class Current: 382/227
• CPC Class Codes: G06V 30/248 involving plural approaches, e.g. verification by template match; Resolving confusion among similar patterns, e.g. "O" versus "Q" G06V30/242 takes precedence

Abstract

A character recognition apparatus for recognizing characters in a document by sequentially detecting one-character portions of images in a document image and by recognizing the character images of the one-character portions detected. The character recognition apparatus includes a character detecting section for detecting an image of one character from the document image; a character recognizing section for recognizing the detected image of one character and outputting a character code; and a control section for effecting control such that the character image of one character recognized by the character recognizing means is stored in correspondence with the character code obtained as a result of recognition, a comparison is made between the stored character image and the image of a newly detected character, and the character-code stored in correspondence with the relevant character image is read as a result of recognition in a case where the similarity of the images is sufficiently large. Character recognition is effected in which a character image is detected from the document image, the detected character image is recognized, and the character code is outputted, thereby to sequentially recognize a plurality of characters in the document.

Claims

1. A document recognition apparatus in which character images are detected from a document image to be recognized, and character recognition for outputting a character code is effected so as to sequentially recognize a plurality of different characters, said document recognition apparatus comprising:

character detecting means for detecting character images;

character recognizing means for processing, when activated, each detected character image to generate a corresponding character code during a first time interval commencing from the activation to the generation of the character code, said character recognizing means includes means for outputting the generated corresponding character code at the end of the first time interval;

character comparing means for comparing, when activated, each detected character image with a stored character image during a second time interval commencing from activation until the reading of the character code corresponding to the stored character image, said character comparing means includes means for reading the character code corresponding to the stored character image in response to a favorable character image comparison, and means for outputting the read character code corresponding to the stored character image at the completion of the second time interval; and

control means for activating the character recognizing means and the character comparing means at substantially the same time for processing the same detected character image, said control means including means for storing each recognized character image together with its corresponding character code, and means responsive to the outputting of a character code for deactivating one of the character recognizing and character comparing means to discontinue the corresponding process prior to the completion of one of the first and second time intervals.

2. A document recognition apparatus as defined in claim 1 wherein when said character recognizing means recognizes a character image with a preset assurance, said storing means stores the recognized character image together with the corresponding character code.

3. A document recognition apparatus as defined in claim 1 wherein the character code corresponding to the stored character image is read as a result of the favorable character image comparison when the comparing means indicates that the similarity of the images meets a predefined threshold, and the character code as a result of recognition by said character recognizing means is outputted when the comparing means indicates that the similarity of the images does not meet a predefined threshold.

4. A document recognition apparatus as defined in claim 1 wherein the storing means stores a character image and the corresponding character code as a storage unit, the comparing means compares the character image of each storage unit and the image of a newly detected character, the reading means reads the character code corresponding to the stored character image as a result of the favorable character image comparison in the case where the similarity of the images meets a predefined threshold, and the storing means stores the recognized character image and the corresponding character code where the similarity of the images does not meet a predefined threshold.

5. A document recognition apparatus as defined in claim 4 wherein frequency information for registering the frequency with which a character code stored as a storage unit is read is stored with the storage unit and at times when the storage means exceeds a maximum capacity of storage, one storage unit is deleted from the storage means on the basis of the frequency information.

6. A document recognition apparatus as defined in claim 1 wherein the control means activates the character comparing means only at times when there is at least one character image stored in correspondence with a character code.

7. A document recognition apparatus as defined in claim 1 wherein the character code is independent of the size of the image of the character.

8. A document recognition apparatus in which kanji or kana character images are detected from a document image to be recognized, and character recognition for outputting a character code is effected so as to sequentially recognize a plurality of characters, said document recognition apparatus comprising:

character detecting means for detecting character images;

simple character image feature extracting means for extracting a simple character image feature from the detected character image;

character recognizing means for processing, when activated, each detected character to generate a character code corresponding to the extracted simple character image feature during a first time interval commencing from the activation to the generation of the character code, said character recognizing means includes means for outputting a character code corresponding to the extracted simple character image feature at the completion of the first time interval;

character comparing means for comparing, when activated, the stored simple character image feature with the extracted simple character image feature during a second time interval commencing from activation until the reading of the character code corresponding to the stored simple character image feature, said character comparing means including means for reading the character code corresponding to the stored simple character image feature, and means for outputting the character code corresponding to the stored simple character image feature in response to a favorable comparison at the end of the second time interval; and

control means for activating at substantially the same time for processing of the same character image the character recognizing means in response to each character image detected, and a character comparing means in response to each extracted simple character image feature from the character image, said control means including means for storing each simple character image feature together with its corresponding character code, and means responsive to the outputting of a character code for deactivating one of the character recognizing and character comparing means to discontinue the corresponding process prior to the completion of one of the first and second time intervals.

9. A document recognition apparatus as defined in claim 8 wherein the control means activates the character comparing means only at time when there is at least one simple character image feature stored in correspondence with a character code.

10. A document recognition apparatus as defined in claim 8 wherein the character code is independent of the size of the image of the character.

11. A document recognition apparatus as defined in claim 8 wherein when said character recognizing means recognizes a character image with a preset assurance, said character image feature storing means stores the simple character image feature of the recognized character image together with the corresponding character code obtained as a result of recognition of the recognized character image.

12. A document recognition apparatus as defined in claim 8 wherein the simple character image feature of a recognized character and the corresponding character code are stored in said simple character image feature storing means as a storage unit, a comparison is sequentially made between the simple character image feature stored as a storage unit and the simple character image feature of a newly detected character, the reading means reads the character code corresponding to the stored simple character image feature as a result of the favorable character image comparison in the case where the similarity of the simple character image features meets a predefined threshold, the reading means reads the character code recognized by said character recognizing means in a case where the similarity of the simple character image features does not meet a predefined threshold, and the storing means stores the recognized simple character image feature and the corresponding character code where the similarity of the images does not meet a predefined threshold.

13. A document recognition apparatus as defined in claim 8 wherein frequency information for registering the frequency with which a character code stored is a storage unit is read is stored with the storage unit, and if the storing means exceeds a maximum capacity of storage, one storage unit is deleted from the storing means on the basis of the frequency information.