JPH10198688A - Fixed form document reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a reader capable of efficiently sorting/retrieving a fixed form document. SOLUTION: A character area is extracted from a fixed form document inputted from a picture input part 201 by a character string area extracting part 203 and its character parttern is recognized by a character recognizing part 206. Then the recognized character string is sorted in each item by an item sorting part 209, business category information for sorting is applied to the recognized character string by a business category applying part 210 and the character string of the item sorted by the item sorting part 209 is registered in a name plate database 212 as a fixed form document reading result together with the business category information. Thereby the read fixed form document can be efficiently sorted/retrieved.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a standard document reading apparatus for reading an image such as a business card.

[0002]

2. Description of the Related Art Conventionally, as a standard document reading technology,
One described in Japanese Patent Application Laid-Open No. 2-240780 is known.

According to this, a character string region is first extracted from an image of a standard document to read a standard document, and a recognized character string is output. Subsequently, the recognized character strings are classified into items mainly using a word dictionary. Subsequently, the character strings recognized and classified into items are stored in the fixed-form document database.

[0004]

However, in the above-mentioned conventional technique, the fixed-form document which has been read and registered in the fixed-form document database is used as information for classification other than the character code of the character string registered for each item. And classification and search were inconvenient.

There is another problem that it is very complicated to input additional information for classifying a standard document.

Another problem is that the accuracy rate of the additional information for classifying the standard document is low.

Further, there is a problem that storage capacity and search time are wasted in obtaining additional information for classifying a standard document.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to provide a fixed document reading apparatus for providing additional information for efficiently classifying fixed documents to a reading result.

[0009]

To solve the above-mentioned problems, the present invention has the following arrangement.

According to the first aspect, a character recognizing means for extracting a character string area from an image of a standard document and outputting a recognized character string, an item classifying means for adding an item to the recognized character string output by the character recognizing means, An additional information providing unit for providing additional information for classifying the standard document; and a registration unit for registering, for each item, a recognition character string output by the character recognition unit together with the additional information provided by the additional information providing unit. Prepare.

According to a second aspect of the present invention, a character recognizing means for extracting a character string area from an image of a standard document and outputting a recognized character string, an item classifying means for giving an item to the recognized character string output by the character recognizing means, A word dictionary storing words to which additional information for classifying a standard document is stored, word matching means for checking a recognized character string output by the character recognition means with the word dictionary and extracting a matching word, Additional information determining means for determining additional information to be added to the fixed form document from additional information added to words extracted by the matching means; and a recognition character output by the character recognition means together with the additional information determined by the additional information determining means Registration means for registering a column for each item.

According to a third aspect of the present invention, a fixed-form document database that records character strings described in the fixed-form document for each item together with additional information for classifying the fixed-form document, and a character string area is extracted from the fixed-form document input as an image. A character recognition unit that outputs a recognition character string, an item classification unit that adds an item to the recognition character string output by the character recognition unit, and a recognition character string output by the character recognition unit is recorded in the fixed-form document database. Character string collating means for extracting additional information recorded together with a character string that matches the matching character string, and additional information added to the standard document input as the image from the additional information extracted by the character string collating means An additional information determining means for determining information; and a recognition character string output by the character recognition means together with the additional information determined by the additional information determining means. And a registration means for registering in the book database.

According to a fourth aspect of the present invention, a character recognition means for extracting a character string region from an image of a standard document and outputting a recognized character string, and an item for adding an item to the character string together with additional information for classifying the standard document A word dictionary storing words to which classification information has been added, a word matching means for checking a recognized character string output by the character recognition means with the word dictionary and extracting a matching word, and a word extracted by the word matching means. An additional information determining unit that determines additional information to be added to the fixed form document from the added additional information; and a recognition character string output by the character recognition unit from item classification information added to a word extracted by the word matching unit. Item classification means for assigning items, and registration means for registering, for each item, a recognition character string output by the character recognition means together with the additional information determined by the additional information determination means.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS.

FIG. 1 is a block diagram of a device for reading business cards. In the drawing, reference numeral 201 denotes an image input unit for inputting a business card image; 202, an image memory for storing the business card image input by the image input unit 201; 203, a character string area extracted from the business card image stored in the image memory 202; A character string extracting unit 204 for extracting a character pattern from a character string region in the business card image extracted by the character string area extracting unit 203; 205, a recognition dictionary for character recognition;
Reference numeral 06 denotes a character recognition unit for collating and recognizing the character extracted by the character extraction unit 204 with the recognition dictionary 205. Reference numeral 207 denotes a recognition character memory (FIG. 3) for storing the character recognized by the character recognition unit 206 in units of character strings. Is a word dictionary (FIG. 4) in which words, item classification information, and industry information are recorded. 209 is an item classification unit that collates the recognized character strings with words recorded in the word dictionary 208 and classifies them as items. Business type information and item classification unit 2 assigned to the read result to which the character string of a specific item recorded in the business card database 212 belongs
09 is an industry information adding unit that adds industry information to the result of reading the business card image from the industry information corresponding to the word matched with the word dictionary 208, and 211 is the classification result of the item classification unit 209 and the industry that the industry information adding unit 210 has added. A business card database registration unit that registers the reading result of the business card image together with the information in the business card database, 212 is a business card database (FIG. 5) in which the business card database registration unit 211 records one or more business card image reading results, and 213 is a business card database registration unit. 211 is a user interface (FIGS. 6, 7, and 8) for confirming, correcting, and classifying registered contents.

The operation when the business card image P1 shown in FIG. 2 is input to the business card reading device configured as described above will be described below.

An image input unit 201 inputs a business card image P 1, and stores it in an image memory 202. Character string area extraction unit 2
03 extracts character string regions L1, L2 and L3 from the business card image P1 stored in the image memory 202. The character cutout unit 204 converts character patterns L1C1, L1
, L1C4, L2C1, ..., L2C4, L3
Cut out C1,..., L3C4. The character recognition unit 206 collates each cut-out character pattern with the recognition dictionary 205 to obtain candidate characters, and as shown in FIG.
To hold. The item classification unit 209 stores the recognition character memory 207
Are collated with the word dictionary 208 shown in FIG. 4 for each character string area to classify items. L for item classification
1, L2 and L3 will be described in order. When the recognition result of L1 and the words in the word dictionary 208 are sequentially compared, 1 in L1C3 is obtained.
The position candidate character "Large" and the first position candidate character "Gaku" of L1C4 are W
Business type “educational institution” that matches the word “university” of 3 and has 1 registered in the item classification information “company” and business type information of W3
Get. When the recognition result of L2 and the words in the word dictionary 208 are sequentially compared, the second-place candidate character “Kyo” of L2C1 and L2C2
Matches the word “Kyoto” in W5,
The business type “educational institution” in which 1 is registered in the item classification information “title” of W5 and the business type information is obtained. When the L3 recognition result and the words in the word dictionary 208 are sequentially collated, the first-place candidate character "yama" of L3C1 and the first-place candidate character "da" of L3C2 match the word "Yamada" of W6, and the second place character of L3C3. The candidate character "fat" and the first-place candidate character "ro" of L3C4 match the word "taro" of W7, and the item classification information "name" common to W6 and W7 is obtained. The business type information adding unit 210 counts the business type information obtained by the word matching performed by the item classification unit 209 to extract the maximum business type information “educational institution”, and also registers the business type information “educational institution” in the business card database 212 shown in FIG. The character string “Kadoshin University” of the company item of the data D1 is the first-place candidate character “Gate” of L1C1, the first-place candidate character “true” of L1C2, and 1 of L1C3.
Since the first-place candidate character "Large" matches the first-place candidate character "Gaku" of L1C4 and the business information of D1 is "educational institution", the business information of "educational institution" is added to the read result of the business card image P1. . The business card database registration unit 211 uses an item classification unit 209 based on the recognition result held in the recognition character memory 207.
From the words W3, W5, W6, and W7 collated by the item classification unit 209 with the first-place candidate character of the character pattern not collated in the above, the read character strings "Kadoshin University", "Kyoto", and "Taro Yamada" are constructed. Output to the UI section 213 together with the item classification results (in order, “company”, “title”, “name”) obtained by the classification section 209 and the business information “educational institution” provided by the business provision section 210 (FIG. 6). 5) The character string, the item classification result, and the business type information constructed in accordance with the input of the UI section are registered in the part D26 of the business card database 212 shown in FIG. The UI section 213 displays the output of the business card database registration section (701), returns the user's confirmation or correction result to the business card database registration section, and outputs the registered contents of the business card database by a specified classification method. FIG. 7 is a display when a company name is specified in the classification method of the business card database 212.
Classification is performed using the registered contents of 12 company items. FIG.
Is displayed when the business type is specified as the classification method of the business card database, and is classified using the business type information of the business card database 212.

As apparent from the above description, in reading a business card image, the present invention adds the business information to the read result with a small number of collations by adding the business information attribute to the word dictionary used for item classification. In addition, by enabling the classification display of the business card database other than the character code order, the user can easily search when using the business card database. In addition, the reliability of the business information to be provided is improved by collating with the company name or the like of the read result registered in the past and searching for the business information provided in the past, so that reliable classification can be performed.

Further, in view of the present invention, the present invention is not limited to business cards only, but can be applied to general fixed documents in general. Further, the additional information to be provided is not limited to the type of business, and can be applied to other information such as a region.

[0020]

As described above, according to the present invention, the following advantageous effects can be obtained.

According to the first aspect of the present invention, the user is provided with classification means other than the classification based on the read character code order by providing additional information for classifying the read standard document, and the read standard document is efficiently classified and searched. it can.

According to the second aspect of the present invention, a word dictionary is provided, and additional information is automatically given by collating with a read result, so that the work for the user to give additional information can be reduced.

According to the configuration of the third aspect, the reliability of the additional information to be provided by referring to the past read result from the standard document database can be enhanced, and the work required for the user to correct the additional information can be reduced.

In the configuration of the fourth aspect, the word dictionary used for classifying items of the standard document and the word dictionary for providing the additional information are made common, so that the storage capacity used in the apparatus and the time required for word matching can be reduced. .

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a business card reading device according to an embodiment of the present invention;

FIG. 2 illustrates an example of a business card image.

FIG. 3 is a diagram illustrating an example of a character recognition result of a business card image;

FIG. 4 is a configuration diagram of a word dictionary including item classification information and industry information.

FIG. 5 is a configuration diagram of a business card database.

FIG. 6 is a diagram showing an example of a reading result confirmation screen.

FIG. 7 is a diagram showing an example of an industry classification display of a business card database;

FIG. 8 is a diagram showing an example of a classification display based on company names in a business card database;

[Explanation of symbols]

 Reference Signs List 201 Image input unit 202 Image memory 203 Character string area extraction unit 204 Character cutout unit 205 Recognition dictionary 206 Character recognition unit 207 Recognition character memory 208 Word dictionary 209 Item classification unit 210 Business information addition unit 211 Business card database registration unit 212 Business card database 213 U・ I part 701 Reading result confirmation display

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Hoho Takakura 1006 Kadoma Kadoma, Kadoma, Osaka Prefecture Inside Matsushita Electric Industrial Co., Ltd. (72) Yoshihiro Yamamoto 1006 Kadoma Kadoma, Kadoma, Osaka Pref.

Claims (4)

[Claims] 1. A character recognition means for extracting a character string area from an image of a fixed form document and outputting a recognized character string, an item classifying means for adding an item to a recognized character string output by the character recognition means, and the fixed form document A fixed form document having additional information adding means for adding additional information for classifying the information, and registration means for registering, for each item, a recognition character string output by the character recognition means together with the additional information added by the additional information adding means. Reader. 2. A character recognition unit for extracting a character string region from an image of a standard document and outputting a recognition character string, an item classifying unit for adding an item to the recognition character string output by the character recognition unit, and A word dictionary storing words to which additional information for classification is stored, word matching means for checking a recognized character string output by the character recognition means with the word dictionary and extracting a matching word, An additional information determining means for determining additional information to be added to the fixed form document from additional information added to the word to be extracted; and a recognition character string output by the character recognition means together with the additional information determined by the additional information determining means. A fixed form document reading device having a registration unit for registering each document. 3. A fixed-form document database for recording a character string described in the fixed-form document for each item together with additional information for classifying the fixed-form document, a character string area extracted from the fixed-form document input as an image, and a recognition character A character recognizing unit that outputs a string, an item classifying unit that adds an item to the recognized character string output by the character recognizing unit, and a character that is output from the character recognizing unit and outputs the recognized character string in the fixed-form document database. A character string matching unit that extracts additional information recorded together with a matching character string by collating with a column, and determines additional information to be added to the standard document input as the image from the additional information extracted by the character string matching unit And a recognition character string output by the character recognition means together with the additional information determined by the additional information determination means for each item. A fixed form document reading device having a registration unit for registering a document in a document. 4. A character recognition means for extracting a character string region from an image of a standard document and outputting a recognition character string, and item classification information for adding an item to the character string together with additional information for classifying the standard document. A word dictionary storing the assigned words,
A word matching unit that matches a recognized character string output by the character recognition unit with the word dictionary to extract a matching word; and an addition that is added to the fixed-form document from additional information added to the word extracted by the word matching unit. Additional information determining means for determining information; item classifying means for assigning an item to a recognized character string output by the character recognizing means from item classification information assigned to a word extracted by the word matching means; A registration unit for registering, for each item, a recognition character string output by the character recognition unit together with the additional information determined by the unit.