US5818952A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Apparatus for assigning categories to words in a documents for databases
Assignee: Panasonic Holdings Corp
•
Filing: 1995-12-27
•
Granted: 1998-10-06
•
Expires: 2015-12-27
Abstract
Abstract An apparatus and a method are provided for assigning categories to words in document images. The apparatus and method extract characters and words from document images and assign categories to the words by referring to dictionaries installed in the apparatus. Furthermore, the apparatus and method correct or assign categories by referring to category correction rule dictionary.
Inventors
- 👤 Mariko Takenouchi
- 👤 Satoshi Emura
- 👤 Saki Takakura
- 👤 Ichiro Nakao
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-31 13:36
CPC Classifications
G; G06; G06V; G06V30/00; G06V30/40
Claims Analysis
Total Claims: 31
Claim 1
Independent Claim
226 words
An apparatus for assigning categories to words in a document image, comprising: a character row extracting means for extracting character row images from the document image; a character extracting means for extracting character images for each character row image in units of a character image; a character recognizing means for recognizing characters from the extracted character images and outputting recognition results; a word detecting means for detecting words for each character row by using positions of the extracted character images and the character recognition results; a keyword dictionary for prestoring categories to be assigned and groups of keyword character rows belonging to the categories; a category correction rule dictionary for prestoring rules used for correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row; and a category assigning means for comparing each character recognition result for each detected word with the groups of keyword character rows in the keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which the keyword character row belongs to the word, then, referring to the category correction rule dictionary, and, if any applicable rule is detected, any of correcting the category assigned to the word and assigning a category to the word according to the applicable rule.
Claim 2
Independent Claim
226 words
An apparatus for assigning categories to words in a document image, comprising: a character row extracting means for extracting character row images from the document image; a character extracting means for extracting character images for each character row image in units of a character image; a character recognizing means for recognizing characters from the extracted character images and outputting recognition results; a word detecting means for detecting words for each character row by using positions of the extracted character images and the character recognition results; a keyword dictionary for prestoring categories to be assigned and groups of keyword character rows belonging to the categories; a category correction rule dictionary for prestoring rules used for correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row; and a category assigning means for comparing each character recognition result for each detected word with the groups of keyword character rows in the keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which the keyword character row belongs to the word, then, referring to the category correction rule dictionary, and, if any applicable rule is detected, any of correcting the category assigned to the word and assigning a category to the word according to the applicable rule.
Claim 3
Dependent Claim
55 words
The apparatus of claim 1, wherein the category assigning means further comprises: a joining means for joining two of more than one category-undefined word detected in a same character row after the keyword dictionary has been referred to, the keyword dictionary being referred to again for the joined words regarded as a new undefined word.
Claim 4
Dependent Claim
140 words
The apparatus of claim 1, wherein the category assigning means comprises: a referring unit for referring to the keyword dictionary; and a category correcting unit for any of correcting and assigning a category by referring to the category correction rule dictionary; and wherein the referring unit comprises: a selecting unit for selecting from the keyword dictionary keyword character rows with the same number of characters as the detected word, keyword character rows having less characters than the detected word, and keyword character rows having more characters than the detected word; and a comparing unit for comparing characters of the detected word with those of the selected keyword character rows, the comparing unit, if a number of characters of the detected word differ from that of the keyword character row, comparing the characters by masking a part of a larger one.
Claim 5
Dependent Claim
67 words
The apparatus of claim 3, wherein the category correcting unit comprises: a numeral counting unit for counting numerals in one of an undefined word and more than one of consecutive undefined words in a same character row; and a numeral category assigning unit for assigning a category for numeral character rows to the word if the counted value of the numeral counting unit exceeds a certain value.
Claim 6
Dependent Claim
83 words
The apparatus of claim 1 further comprises: a category-word dictionary for prestoring words to be referred to for each category; and a word correcting means for, after the category assignment and correction by the category assigning means have been completed, selecting a part of the category-word dictionary suitable for a category for each word, comparing the word with words in the part of the category-word dictionary, and replacing the recognition result of the word by a matched word detected in the category-word dictionary.
Claim 7
Dependent Claim
55 words
The apparatus of claim 5, wherein the category assigning means further comprises: a joining means for joining two of more than one category-undefined word detected in a same character row after the keyword dictionary has been referred to, the keyword dictionary being referred to again for the joined words regarded as a new undefined word.
Claim 8
Dependent Claim
140 words
The apparatus of claim 6, wherein the category assigning means comprises: a referring unit for referring to the keyword dictionary; and a category correcting unit for any of correcting and assigning a category by referring to the category correction rule dictionary; and wherein the referring unit comprises: a selecting unit for selecting from the keyword dictionary keyword character rows with the same number of characters as the detected word, keyword character rows having less characters than the detected word, and keyword character rows having more characters than the detected word; and a comparing unit for comparing characters of the detected word with those of the selected keyword character rows, the comparing unit, if a number of characters of the detected word differ from that of the keyword character row, comparing the characters by masking a part of a larger one.
Claim 9
Dependent Claim
67 words
The apparatus of claim 7, wherein the category correcting unit comprises: a numeral counting unit for counting numerals in one of an undefined word and more than one of consecutive undefined words in a same character row; and a numeral category assigning unit for assigning a category for numeral character rows to the word if the counted value of the numeral counting unit exceeds a certain value.
Claim 10
Dependent Claim
17 words
The apparatus of claim 5, wherein the category-word dictionary and the keyword dictionary share the same part.
Claim 11
Independent Claim
280 words
An apparatus for assigning categories to words in a document image, comprising: a character row extracting means for extracting character row images from the document image; a block extracting means for extracting character rows in a range of certain distances as a block by comparing coordinates of each extracted character row; a character extracting means for extracting character images for each character row image in units of a character image; a character recognizing means for recognizing characters from the extracted character images and outputting recognition results; a word detecting means for detecting words for each character row by using positions of the extracted character images and the character recognition results; a keyword dictionary for prestoring categories to be assigned and groups of keyword character rows belonging to the categories; a first category correction rule dictionary for prestoring rules used for any of correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row; a second category correction rule dictionary for prestoring rules of relation between categories of words in different character rows in a same block; and a category assigning means for comparing each character recognition result for each detected word with the groups of keyword character rows in the keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which the keyword character row belongs to the word, then, referring to the first and second category correction rule dictionaries, and, if any applicable rule is detected, any of correcting the category assigned to the word and assigning a category to the word according to the applicable rule.
Claim 12
Independent Claim
280 words
An apparatus for assigning categories to words in a document image, comprising: a character row extracting means for extracting character row images from the document image; a block extracting means for extracting character rows in a range of certain distances as a block by comparing coordinates of each extracted character row; a character extracting means for extracting character images for each character row image in units of a character image; a character recognizing means for recognizing characters from the extracted character images and outputting recognition results; a word detecting means for detecting words for each character row by using positions of the extracted character images and the character recognition results; a keyword dictionary for prestoring categories to be assigned and groups of keyword character rows belonging to the categories; a first category correction rule dictionary for prestoring rules used for any of correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row; a second category correction rule dictionary for prestoring rules of relation between categories of words in different character rows in a same block; and a category assigning means for comparing each character recognition result for each detected word with the groups of keyword character rows in the keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which the keyword character row belongs to the word, then, referring to the first and second category correction rule dictionaries, and, if any applicable rule is detected, any of correcting the category assigned to the word and assigning a category to the word according to the applicable rule.
Claim 13
Dependent Claim
55 words
The apparatus of claim 10, wherein the category assigning means further comprises: a joining means for joining two of more than one category-undefined word detected in a same character row after the keyword dictionary has been referred to, the keyword dictionary being referred to again for the joined words regarded as a new undefined word.
Claim 14
Dependent Claim
140 words
The apparatus of claim 10, wherein the category assigning means comprises: a referring unit for referring to the keyword dictionary; and a category correcting unit for any of correcting and assigning a category by referring to the category correction rule dictionary; and wherein the referring unit comprises: a selecting unit for selecting from the keyword dictionary keyword character rows with the same number of characters as the detected word, keyword character rows having less characters than the detected word, and keyword character rows having more characters than the detected word; and a comparing unit for comparing characters of the detected word with those of the selected keyword character rows, the comparing unit, if a number of characters of the detected word differ from that of the keyword character row, comparing the characters by masking a part of a larger one.
Claim 15
Dependent Claim
67 words
The apparatus of claim 12, wherein the category correcting unit comprises: a numeral counting unit for counting numerals in one of an undefined word and more than one of consecutive undefined words in a same character row; and a numeral category assigning unit for assigning a category for numeral character rows to the word if the counted value of the numeral counting unit exceeds a certain value.
Claim 16
Dependent Claim
83 words
The apparatus of claim 10 further comprises: a category-word dictionary for prestoring words to be referred to for each category; and a word correcting means for, after the category assignment and correction by the category assigning means have been completed, selecting a part of the category-word dictionary suitable for a category for each word, comparing the word with words in the part of the category-word dictionary, and replacing the recognition result of the word by a matched word detected in the category-word dictionary.
Claim 17
Dependent Claim
17 words
The apparatus of claim 14, wherein the category-word dictionary and the keyword dictionary share the same part.
Claim 18
Independent Claim
226 words
A method for assigning categories to words in a document image, comprising: a first step for extracting character row images from the document image; a second step for extracting character images for each character row image in units of a character image; a third step for recognizing characters from the extracted character images and outputting recognition results; a fourth step for detecting words for each character row by using positions of the extracted character images and the character recognition results; a fifth step for comparing a detected word with groups of keyword character rows in a keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which a keyword character row belongs to the word, the keyword dictionary prestoring categories to be assigned and groups of keyword character rows belonging to the categories; a sixth step for referring to a category correction rule dictionary, and, if any applicable rule is detected, correcting the category assigned to the word according to the applicable rule, the category correction rule dictionary prestoring rules used for correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row; and a seventh step for outputting recognition results of words and categories related to the words obtained from the previous steps.
Claim 19
Independent Claim
226 words
A method for assigning categories to words in a document image, comprising: a first step for extracting character row images from the document image; a second step for extracting character images for each character row image in units of a character image; a third step for recognizing characters from the extracted character images and outputting recognition results; a fourth step for detecting words for each character row by using positions of the extracted character images and the character recognition results; a fifth step for comparing a detected word with groups of keyword character rows in a keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which a keyword character row belongs to the word, the keyword dictionary prestoring categories to be assigned and groups of keyword character rows belonging to the categories; a sixth step for referring to a category correction rule dictionary, and, if any applicable rule is detected, correcting the category assigned to the word according to the applicable rule, the category correction rule dictionary prestoring rules used for correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row; and a seventh step for outputting recognition results of words and categories related to the words obtained from the previous steps.
Claim 20
Dependent Claim
61 words
The method of claim 16, wherein the fifth step further comprises: a sub-step for detecting category-undefined words other than a category-undefined word in a same character row after the keyword dictionary has been referred to; and a sub-step for, if detected, joining two words and referring to the keyword dictionary again for the joined words regarded as a new undefined word.
Claim 21
Dependent Claim
103 words
The method of claim 16, wherein the fifth step comprises: a sub-step for selecting from the keyword dictionary keyword character rows with the same number of characters as the detected word, keyword character rows having less characters than the detected word, and keyword character rows having more characters than the detected word; and a sub-step for comparing characters of the detected word with those of the selected keyword character rows in a comparing unit and, if a number of characters of the detected word differ from that of the keyword character row, comparing the characters by masking a part of a larger one.
Claim 22
Dependent Claim
61 words
The method of claim 18, wherein the sixth step comprises: a sub-step for counting numerals in one of an undefined word and more than one of consecutive undefined words in a same character row; and a sub-step for assigning a category for numeral character rows to the word if the counted value of a numeral counting unit exceeds a certain value.
Claim 23
Dependent Claim
79 words
The method of claim 16 further comprises: a sub-step for, after the category assignment and correction by a category assigning means have been completed, selecting a part of a category-word dictionary suitable for a category for each word, comparing the word with words in the part of the category-word dictionary, and replacing the recognition result of the word by a matched word detected in the category-word dictionary, the category-word dictionary prestoring words to be referred to for each category.
Claim 24
Dependent Claim
61 words
The method of claim 20, wherein the fifth step further comprises: a sub-step for detecting category-undefined words other than a category-undefined word in a same character row after the keyword dictionary has been referred to; and a sub-step for, if detected, joining two words and referring to the keyword dictionary again for the joined words regarded as a new undefined word.
Claim 25
Dependent Claim
103 words
The method of claim 21, wherein the fifth step comprises: a sub-step for selecting from the keyword dictionary keyword character rows with the same number of characters as the detected word, keyword character rows having less characters than the detected word, and keyword character rows having more characters than the detected word; and a sub-step for comparing characters of the detected word with those of the selected keyword character rows in a comparing unit and, if a number of characters of the detected word differ from that of the keyword character row, comparing the characters by masking a part of a larger one.
Claim 26
Dependent Claim
61 words
The method of claim 22, wherein the sixth step comprises: a sub-step for counting numerals in one of an undefined word and more than one of consecutive undefined words in a same character row; and a sub-step for assigning a category for numeral character rows to the word if the counted value of a numeral counting unit exceeds a certain value.
Claim 27
Independent Claim
291 words
A method for assigning categories to words in a document image, comprising: a first step for extracting character row images from the document image; a second step for extracting character rows in a range of certain distances as a block by comparing coordinates of each extracted character row; a third step for extracting character images for each character row image in units of a character image; a fourth step for recognizing characters from the extracted character images and outputting recognition results; a fifth step for detecting words for each character row by using positions of the extracted character images and the character recognition results; a sixth step for comparing a detected word with groups of keyword character rows in a keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which the keyword character row belongs to the word, the keyword dictionary prestoring categories to be assigned and the groups of keyword character rows belonging to the categories; a seventh step for referring to a first category correction rule dictionary and a second category correction rule dictionary, and, if any applicable rule is detected, any of correcting the category assigned to the word and assigning a category according to the applicable rule, the first category correction rule dictionary prestoring rules used for any of correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row, and the second category correction rule dictionary prestoring rules of relation between categories of words in different character rows in a same block; and an eighth step for outputting recognition results of words and categories related to the words obtained from the previous steps.
Claim 28
Independent Claim
291 words
A method for assigning categories to words in a document image, comprising: a first step for extracting character row images from the document image; a second step for extracting character rows in a range of certain distances as a block by comparing coordinates of each extracted character row; a third step for extracting character images for each character row image in units of a character image; a fourth step for recognizing characters from the extracted character images and outputting recognition results; a fifth step for detecting words for each character row by using positions of the extracted character images and the character recognition results; a sixth step for comparing a detected word with groups of keyword character rows in a keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which the keyword character row belongs to the word, the keyword dictionary prestoring categories to be assigned and the groups of keyword character rows belonging to the categories; a seventh step for referring to a first category correction rule dictionary and a second category correction rule dictionary, and, if any applicable rule is detected, any of correcting the category assigned to the word and assigning a category according to the applicable rule, the first category correction rule dictionary prestoring rules used for any of correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row, and the second category correction rule dictionary prestoring rules of relation between categories of words in different character rows in a same block; and an eighth step for outputting recognition results of words and categories related to the words obtained from the previous steps.
Claim 29
Dependent Claim
79 words
The method of claim 24 further comprises: a step for, after the category assignment and correction by a category assigning means have been completed, selecting a part of a category-word dictionary suitable for a category for each word, comparing the word with words in the part of the category-word dictionary, and replacing the recognition result of the word by a matched word detected in the category-word dictionary, the category-word dictionary prestoring words to be referred to for each category.
Claim 30
Independent Claim
285 words
An apparatus for assigning categories to words in a document image, comprising: a character row extracting means for extracting character row images from the document image; a character extracting means for extracting character images for each character row image in units of a character image; a character recognizing means for recognizing characters from the extracted character images and outputting recognition results; a word detecting means for detecting words for each character row by using positions of the extracted character images and the character recognition results; a keyword dictionary for prestoring categories to be assigned and groups of keyword character rows belonging to the categories; category correction rule dictionary for prestoring rules used for correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row; and a category assigning means for comparing each character recognition result for each detected word with the groups of keyword character rows in the keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which the keyword character row belongs to the word, then, referring to the category correction rule dictionary, and, if any applicable rule is detected, any of correcting the category assigned to the word and assigning a category to the word according to the applicable rule including a numeral counting unit for counting the number of characters that are numerals and a telephone category assigning unit for assigning a category of telephone if three consecutive words in a same character row has not been assigned a category and the numeral counting unit determines that at least half of the characters in the character row are numerals.
Claim 31
Independent Claim
285 words
An apparatus for assigning categories to words in a document image, comprising: a character row extracting means for extracting character row images from the document image; a character extracting means for extracting character images for each character row image in units of a character image; a character recognizing means for recognizing characters from the extracted character images and outputting recognition results; a word detecting means for detecting words for each character row by using positions of the extracted character images and the character recognition results; a keyword dictionary for prestoring categories to be assigned and groups of keyword character rows belonging to the categories; category correction rule dictionary for prestoring rules used for correcting a category of a word and assigning a category to a word in terms of relation between the word and other words in a same character row; and a category assigning means for comparing each character recognition result for each detected word with the groups of keyword character rows in the keyword dictionary, and, if a matching keyword character row is detected, assigning a category to which the keyword character row belongs to the word, then, referring to the category correction rule dictionary, and, if any applicable rule is detected, any of correcting the category assigned to the word and assigning a category to the word according to the applicable rule including a numeral counting unit for counting the number of characters that are numerals and a telephone category assigning unit for assigning a category of telephone if three consecutive words in a same character row has not been assigned a category and the numeral counting unit determines that at least half of the characters in the character row are numerals.
References Cited (Backward Citations)
US4524453A
Cited
JPS6355624A
Cited
US5050218A
Cited
JPS63158678A
Cited
US4847912A
Cited
US5410611A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2001-10-05 | 2002-04-06 | 2002-10-06 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2005-10-05 | 2006-04-06 | 2006-10-06 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2009-10-05 | 2010-04-06 | 2010-10-06 | $7700.00 | Window Closed / Lapsed |