Document revising system for use with document reading and translating system

• US 5,222,160 A
• Filed: 12/28/1990
• Issued: 06/22/1993
• Est. Priority Date: 12/28/1989
• Status: Expired due to Term

• Current Assignee: PFU Limited (Fujitsu Limited)
• Original Assignee: Fujitsu Limited
• Inventors: Sakai, Rieko, Oshima, Chieko, Kitajima, Naoko
• Primary Examiner(s): Moore, David K.
• Assistant Examiner(s): Cammarata, Michael
• Application Number:
  US07/635,077
• Time in Patent Office: 907 Days
• Field of Search: 382/1, 382/9, 382/57, 382/61, 382/10, 382/13, 382/48, 364/419, 340/723
• US Class Current: 382/185
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation
• CPC Class Codes: G06F 40/55 Rule-based translation
• CPC Class Codes: G06F 40/58 Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation
• CPC Class Codes: G06V 10/987 with the intervention of an operator

Abstract

An image-to-character-position-correspondence-table producing unit produces image-to-character-position-correspondence-table composed of a set comprising an-image-of-a-document, a character-recognized document and a translated document. A candidate character producing unit produces candidate characters for revising misrecognized characters. A Japanese-document-to-translated-document correspondence table stores a correspondence relationship between an original Japanese document and a translated document in the form of a table. When misrecognized characters are being revised, the image-to-character-position-correspondence-table is displayed by the image-to-character-position-correspondence-table. A revising unit prompts a user to specify a misrecognized portion in the translated document of the image to character-position-correspondence-table. Next, the revising unit refers to the Japanese-document-to-translated-document correspondence table to extract a portion of each of the-image-of-the-document and the recognized document that corresponds to the specified portion and causes the image-to-character-position-correspondence-table producing unit to display the corresponding portions. Subsequently, the revising unit refers to the candidate character producing unit to extract candidate characters as requested by the user and causes the image-to-character-position-correspondence-table producing unit to display these candidate characters. Candidate characters are selected by the user. The misrecognized portion in the recognized document is replaced with the selected candidate characters, a new character-recognized document is translated and a newly translated document is displayed. In this way even foreigners who have little knowledge of Japanese can carry out revision work on misrecognized characters with ease.

Claims

1. A document revising apparatus for use with a document reading and translating system for performing character recognition of an image-of-a-document to produce a recognized document, and for translating the recognized document to produce a translated document, comprising:

character recognition means for entering a document written in a first language as an image-of-a-document, segregating characters from said image-of-a-document and performing character recognition one each segregated character to produce a recognized document;

translating process means for translating said recognized document in said first language to a second language to produce the translated document, and for producing a correspondence relationship between the recognized document and the translated document;

correspondence table producing and displaying means for producing and displaying an image-to-character-position-correspondence-table in which a correspondence is established between said image-of-a-document, said recognized document and said translated document;

original-document-to-translated-document correspondence relationship storing means for storing the correspondence relationship between the recognized document and the translated document, from said translating process means;

candidate character producing means for producing candidate characters used for revising misrecognized characters; and

document revising means including

means for allowing a user to specify a misrecognized portion in said image to-character-position-correspondence-table displayed by said correspondence table producing and displaying means;

means for referring to said original-document-to-translated-document correspondence relationship storing means to extract portions of said image-of-the-document and said recognized document which correspond to said misrecognized portion specified and causing said correspondence table producing and displaying means to display said extracted portions;

means for referring to said candidate character producing means to extract candidate characters for said misrecognized portion in said recognized document as requested by the user and causing said correspondence table producing and displaying means to display said candidate characters;

means for allowing the user to select arbitrary characters from said candidate characters displayed and replacing said misrecognized portion in said recognized document with selected candidate characters; and

means for causing said translating means to retranslate a new document in which said misrecognized portion is replaced with said selected candidate characters to produce a new translated document and causing said correspondence table producing and displaying means to display said new translated document.

2. The document revising apparatus according to claim 1, wherein said character recognition means includes

image entry means for entering said document in said first language as an image-of-a-document,

image-of-the-document storage means for storing said image-of-the-document,

character segmentation means for segregating each character from said stored image-of-the-document,

character recognition means for performing character recognition on each of said characters segregated from said image-of-the-document, and

document storage means for storing said recognized character as the recognized document.

3. The document revising apparatus according to claim 2, wherein said translation means includes

translating means for translating the recognized document into the translated document, and

translated document storage means for storing the translated document.

4. The document revising apparatus according to claim 3,

wherein said character segmentation means generates position information corresponding to each segregated character, and

wherein said correspondence table producing and displaying means reads said image-of-the-document from said image-of-the-document storage means, said recognized document from said document storage means and said translated document from said translated document storage means based on position information read from said character segmentation means, to produce said displayed document.

5. The document revising apparatus according to claim 2, wherein said candidate character producing means refers to said character recognition means to produce said candidate characters.

6. The document revising apparatus according to claim 2, wherein said image entry means includes an optical character reader for entering a document written or printed on a sheet of paper as an image-of-a-document.

7. The document revising apparatus according to claim 1, wherein said original-document-to-translated-document correspondence relationship storage means stores a correspondence relationship between said original document and said translated document in the form of a table.

8. The document revising apparatus according to claim 1, wherein said first language is Japanese.

9. A method for revising a translated document of a first language based on an image-of-a-document and a recognized document of a second language using a document revising system, comprising the steps of:

a) displaying at least the translated document using a display;

b) user designating a misrecognized portion of the translated document based on a context of the translated document, using a revising unit;

c) displaying the misrecognized portion in correspondence with a portion of the recognized document and a portion of the image-of-a-document which corresponds to the misrecognized portion using at least the display and correspondence table producing and displaying unit;

d) selectively displaying at least one candidate character for the recognized document, corresponding to the misrecognized portion, using at least the display, the correspondence table producing and displaying unit, a candidate character producing unit and the revising unit;

e) user comparing the at least one candidate character of said sep (d) with a corresponding at least one character in the image-of-a-document;

f) user selecting the at least one candidate character based on said step (e) to replace the misrecognized portion with the at least one candidate character, using at least a character recognition process unit;

g) translating the at least one candidate character to modify the misrecognized portion of the translated document, using at least a translating process unit; and

h) displaying the at least one candidate character in correspondence with a portion of the translated document modified in said step (g) and the portion of the image-of-a-document, using at least the display and the correspondence table producing and displaying unit.

10. A method according to claim 9, further comprising the steps of:

i) user determining that the portion of the translated document is correct based on the context of the translated document, and

j) repeating said steps (b)-(i) when said determining of said step (i) establishes that the portion of the translated document is not correct.

11. A method for revising a translated document of a first language based on an image-of-a-document and a recognized document of a second language, comprising the steps of:

a) displaying at least the translated document;

b) designating a misrecognized portion of the translated document;

c) displaying a portion of the image-of-a-document and a portion of the recognized document corresponding to the misrecognized portion;

d) selectively displaying at least one candidate character corresponding to the portion of the recognized document;

e) comparing the at least one candidate character with a corresponding at least one character in the image-of-a-document;

f) selecting the at least one candidate character based on said step (e);

g) translating the at least one candidate character to modify the translated document; and

h) displaying at least the translated document modified in said step (g).

12. A method according to claim 11, further comprising the steps of:

i) determining whether the translated document displayed in said step (h) is correct; and

j) repeating said steps (b)-(i) when said determining of said step (i) establishes that the translated document modified in said step (g), is not correct.