JP5492666B2 - Judgment device, method and program - Google Patents

Description

  The present invention relates to a determination apparatus, method, and program, and in particular, a document in which a region appears and a position in the document by using a captured image of a partial region in a document in which a page break or line break position is fixed as a search query. The present invention relates to a determination apparatus, a method, and a program for supporting creation of a document and an index of each position in the document in response to a search request for acquiring the document.

  Specifically, this is applied to a document that may include the area in a document where the page break or line feed position has been determined, and a position in the document that is not comprehensively acquired, but is intended to uniquely identify the position. The present invention relates to a determination apparatus, method, and program.

  There are not a few scenes in which it is necessary to uniquely identify from which document a part of a document is included in which document or at which position in which document. For example, if there is a magazine cut out at hand, there may be a case where the original cut out magazine is searched and the continuation of the cut out is read. In this case, it is necessary to be able to uniquely identify which magazine the clipping was part of.

  The above example can be regarded as a search system that uses a partial area of a document as a query and inquires about a document name including the area or a document name and a position in the document from a huge amount of documents.

  In order to construct a system that responds to a search request for acquiring information from a document group, it is necessary to analyze the document group in advance and create a search index.

  For example, nouns are extracted from sentences in a document using morphological analysis, which is a general technique, and each noun is used as a search index key as shown in FIG. 23 to search for the appearance position (document name, page) of the noun. An index value method may be mentioned.

  In the above method, as shown in FIG. 24, when an area in which a character exists in a book (hereinafter referred to as “character area”) is shot with a camera-equipped mobile phone, the position (book title / page) of the shooting area is specified. The present invention can be applied to a system that presents a predetermined content in the area.

  In the present invention, this system is applied to a visually impaired person (a person who is blind or who cannot perceive a character area in a book due to low vision, etc., hereinafter referred to as “user”). That is, when a user captures a character area in a book with a camera-equipped mobile phone, an audio file that reads out the area is presented.

  At this time, since the user cannot perceive where the character is in the book, it is not a character region, but a non-character region, that is, a region where a figure / photo is present in the book or a region where nothing is printed. It may end up. In this case, the system needs to determine that the shooting area is a non-character area and notify the user that “the shooting area is a non-character area”.

  However, it is not easy to determine whether an image obtained by photographing a partial area in a book whose unknown non-character area exists is a character area.

  If the object to be photographed is a car or a bookshelf, there is a method for estimating the character area by utilizing the fact that the character area appears in a rectangle such as a license plate or a back cover (for example, Non-Patent Document 1, Non-Patent Document 2). reference). However, since the character area in the book is rarely surrounded by a rectangle, it is difficult to estimate the character area in the book.

  In addition, if the subject to be photographed is a signboard or the like in a landscape, there is a technique for specifying the character region by utilizing the fact that the character region has a large difference in shade between the background color and the character color (see, for example, Non-Patent Document 3). . However, there are non-character areas in the book that have large shade differences from the background color, such as line drawings and ruled lines.

Katsuyoshi Tabuchi, Harumi Kawashima, Eisaku Marubayashi, Tadashi Nakanishi, Akio Shio, Osuka Osamu License plate region extraction considering partial character placement rules IEICE Transactions, D-II, Information System, II-Information Processing J81 -D-2 (10), pp.2280-2287. Minako Sawaki, Hiroshi Murase, Norihiro Hirota “Character recognition in bookshelf images by automatic dictionary selection based on degradation estimation”, Journal of the Institute of Image Information and Television Engineers, 54 (6), pp.881-886. Liu Xiaomei, Satoshi Yamamura, Noboru Onishi, Noboru Sugie, “Extraction of character string regions in scenes” IEICE Transactions, D-II, Information and Systems, II-Information Processing, J81-D-2 (4) , Pp.641-650.

  As described above, since there are various types of books and various non-character areas appearing, it is difficult to determine one character area extraction method in advance. If a method corresponding to only a single type of non-character area is used, the accuracy of determination for a book including a non-type of non-character area that is not assumed is reduced. Although it is conceivable to use a plurality of character region extractions in combination, these methods alone are expensive in processing cost, and it is difficult to say that using a plurality of methods together is practical from the viewpoint of processing speed. In the first place, it is only necessary to know whether or not a character area is included in the image, and the above-described conventional method aiming at extracting the character area from the image with high accuracy is overspec.

  The present invention has been made in view of the above points, and a determination apparatus capable of determining whether an image obtained by capturing a partial area in a book whose unknown non-character area exists is a character area, and An object is to provide a method and a program.

In order to solve the above problems, the present invention (Claim 1) is directed to a document in which a character, a photograph, a figure, a table, a ruled line, or other non-character is described. A determination device that determines whether a whole area or a partial area is an input and the area is a character area including a certain percentage of characters,
An area input means for receiving an input of an area to be determined;
As a result of performing optical character recognition processing assuming that characters are described in the determination target area, at least a character area or a non-character based on the number of unknown words detected and the ratio of the number of words included in the area Character area determination means for determining whether the area;
Determination result output means for outputting either a character area or a non-character area based on the determination result of the character area determination means.

Further, according to the present invention (Claim 2), the character area determination means of Claim 1
Means for determining a character area when the number of detected unknown words is less than a predetermined value A and the area includes a word greater than or equal to a predetermined value B;

Further, according to the present invention (Claim 3), the character area determination means of Claim 1
The ratio of the number of detected unknown words that is less than a predetermined value A and the number of words in which the length of the detected word is less than or equal to a predetermined value C is less than a predetermined value D, and Means for determining that the area is a character area when the area includes a word equal to or greater than a predetermined value B;

Further, the present invention (Claim 4) provides a whole area or a partial area of a document in which a character, a photograph, a figure, a table, a ruled line, or other non-character is described. As an input, a determination device that determines whether or not the area is a character area including a certain percentage of characters,
An area input means for receiving an input of an area to be determined;
As a result of performing optical character recognition processing assuming that characters are described in the determination target area, the character area or non-character is based on at least the length of the detected word and the ratio of the number of words included in the area Character area determination means for determining whether the area;
Determination result output means for outputting either a character area or a non-character area based on the determination result of the character area determination means.

Further, according to the present invention (Claim 5), the character area determination means of Claim 4
And means for determining a character area when the number of words whose length is less than or equal to a predetermined value C is less than a predetermined value D and the area includes a word that is equal to or greater than the predetermined value B.

  In the present invention (Claim 6), the predetermined value C of the length of the word in the character area determination means of Claim 3 or Claim 5 is set to one character.

Further, according to the present invention (Claim 7), an entire area or a partial area of a document in which a character, a photograph, a figure, a table, a ruled line, or any other character is described. A determination method for determining whether the region is a character region including a certain percentage or more of characters as an input,
An area input step for receiving an input of an area to be determined;
As a result of performing optical character recognition processing assuming that characters are described in the determination target area, at least a character area or a non-character based on the number of unknown words detected and the ratio of the number of words included in the area A character region determination step for determining whether the region is a region;
A determination result output step of outputting either a character area or a non-character area based on the determination result of the character area determination step;

Moreover, the present invention (Claim 8) is characterized in that in the character region determination step of Claim 7,
When the number of detected unknown words is less than a predetermined value A and the area includes words greater than or equal to a predetermined value B, the character area is determined.

Further, the present invention (Claim 9) is characterized in that in the character region determination step of Claim 7,
The ratio of the number of detected unknown words that is less than a predetermined value A and the number of words in which the length of the detected word is less than or equal to a predetermined value C is less than a predetermined value D, and When the area includes a word of a predetermined value B or more, it is determined as a character area.

Further, according to the present invention (Claim 10), an entire area or a partial area of a document in which a character, a photograph, a figure, a table, a ruled line, or other non-character characters is described. A determination method for determining whether the region is a character region including a certain percentage or more of characters as an input,
An area input step for receiving an input of an area to be determined;
As a result of performing optical character recognition processing assuming that characters are described in the determination target area, the character area or non-character is based on at least the length of the detected word and the ratio of the number of words included in the area A character region determination step for determining whether the region is a region;
A determination result output step of outputting either a character area or a non-character area based on the determination result of the character area determination step;

The present invention (Claim 11) is characterized in that, in the character region determination step of Claim 10,
When the number of words whose length of the detected word is less than or equal to the predetermined value C is less than the predetermined value D and the area includes a word that is greater than or equal to the predetermined value B, it is determined as a character area.

Further, according to the present invention (Claim 12), in the character region determination step according to Claim 9 or 11, the predetermined value C of the length of the word is one character.

  Moreover, this invention (Claim 13) is a program for functioning a computer as each means which comprises the determination apparatus of any one of Claim 1 thru | or 6.

  When determining whether an image is a character area, the conventional method determines what kind of image is an image (car or bookcase image, landscape image including a signboard, etc.), and the character area suitable for that image In many cases, an extraction method is used, and it is difficult to analyze an image obtained by photographing a partial region in a book where it is not known what non-character region exists. If only a single type of image is defined, the accuracy of determining whether or not it is a character region will be reduced, and if a plurality of types of images are defined and a plurality of types of character region extraction methods are used, the processing cost will be greatly increased.

  On the other hand, the present invention performs optical character recognition without considering what kind of image it is, and determines whether it is a character region by analyzing the result. This is a simple method with low processing costs.

  In particular, according to the present invention, it is possible to determine by a simple method whether an image obtained by photographing a partial area in a book whose unknown non-character area exists is a character area.

It is a system configuration figure in a 1st embodiment of the present invention. It is an example of the document input in the 1st Embodiment of this invention. It is an example of word DB in the 1st Embodiment of this invention. It is a sequence chart of the operation | movement in the 1st Embodiment of this invention. It is an example of the character area extraction of the document in the 1st Embodiment of this invention. It is an example of the conversion from the image file in the 1st Embodiment of this invention to text data. It is an example of the morphological analysis in the word extraction part in the 1st Embodiment of this invention. It is a flowchart of the character area determination in the 1st Embodiment of this invention. It is a system block diagram in the 2nd Embodiment of this invention. It is an example of position DB in the 2nd Embodiment of this invention. It is an example of content DB in the 2nd Embodiment of this invention. It is a sequence chart of operation in a 2nd embodiment of the present invention. It is an example of the character area extraction of the document in the 2nd Embodiment of this invention. It is an example of the conversion from the image file in the 2nd Embodiment of this invention to text data. It is an example of the morphological analysis in the word extraction part in the 2nd Embodiment of this invention. It is a flowchart of the character area determination in the 2nd Embodiment of this invention. It is a flowchart of the process of a content inquiry in the 2nd Embodiment of this invention. It is a sequence chart of operation in a 3rd embodiment of the present invention. It is the total result of the position of each noun in the 3rd Embodiment of this invention. It is a system block diagram in the 4th Embodiment of this invention. It is a sequence chart of operation in a 4th embodiment of the present invention. It is the figure which image | photographed the wide character area in the 4th Embodiment of this invention. It is an example which indexes the appearance position of a noun. It is an example which provides a content based on the character area in a book.

  Embodiments of the present invention will be described below with reference to the drawings.

[First Embodiment]
In the present embodiment, character recognition of the photographed region is performed, and it is determined whether or not the character region is based on the number of unknown words included in the region.

  FIG. 1 shows a system configuration according to the first embodiment of the present invention.

  The system shown in the figure includes a server unit 300, a client unit 400, a word DB 201 connected to the server unit 300, and an optical character recognition device 200.

  The server unit 300 in the figure can be realized by a device such as a PC server, and includes a server-side data transmission / reception unit 301, a word extraction unit 302, and a character area determination unit 303. The word extraction unit 302 includes a word DB 201 and a server side. The optical character recognition device 200 is connected to the data transmitting / receiving unit 301.

  The client unit 400 in FIG. 1 can be realized by a camera-equipped mobile phone or the like, and includes a document photographing unit 401, a client side data transmitting / receiving unit 402, and a determination result presenting unit 403. The document 100 is input to the document photographing unit 401. .

  A document 100 in the figure is one page of a paper medium book including a character area and a non-character area as shown in FIG.

  The optical character recognition device 200 in FIG. 1 is an external device such as general OCR software, which takes an image file containing characters as input, and outputs the converted characters as electronic text data. To do. The word DB 201 is a word dictionary used by general OCR software or morphological analysis software, and it is assumed that almost all the words that we regularly use are stored in the format shown in FIG. The optical character recognition apparatus 200 can identify character units without using the word DB 201 as in general OCR software. However, the optical character recognition apparatus 200 can identify word units using the word DB 201, and as a result, Identification accuracy is improved.

  Next, processing in the present embodiment in the above configuration will be described.

  FIG. 4 is a sequence chart of processing in the first embodiment of the present invention.

  Step 100: Input Step) The document photographing unit 401 of the client unit 400 captures a partial image of a document to be determined, and transmits the partial image to the server unit 300 via the client side data transmitting / receiving unit 402. Detailed processing will be described in steps 101 to 103 below.

  Step 110: Word Extraction Step) The server-side data transmission / reception unit 301 of the server unit 300 acquires a partial image of the document from the client unit 400, and the character region determination unit 303 extracts words from the shooting region. Detailed processing will be described in step 104 below.

  Step 120: Character Area Determination Step) The character area determination unit 303 determines whether or not the shooting area is a character area. Detailed processing will be described in step 105 below.

  Step 130: Output Step) The character area determination unit 303 outputs a determination result as to whether or not the shooting area is a character area. Detailed processing will be described in steps 106 to 108 below.

  Details of the above processing will be described below.

  Step 101) The document photographing unit 401 photographs the character area of the document 100 and outputs it as an image file as shown in FIG.

  Step 102) The client-side data transmission / reception unit 402 receives the output of step 101 as an input, and outputs the image file to the server unit 300 through the network or the like as it is.

  Step 103) The server-side data transmission / reception unit 301 receives the output of Step 102 as input, recognizes the text in the image file using the optical character recognition device 200, and converts it into text data as shown in FIG. Output. As described above, the optical character recognition apparatus 200 uses the word DB 201 at the time of identification, and can accurately recognize words registered in the word DB 201, but the recognition accuracy of unregistered words / characters is lowered. At this time, since the shooting situation is not good, it is assumed that a character string that is correctly “type” is erroneously recognized as “type”. In the present embodiment, the character area is shown in the image file. However, it is not considered at all whether the object to be shown is a character area or a non-character area.

  Step 104) The word extraction unit 302 receives the output of Step 103 as an input, and extracts and outputs the word. Here, word extraction is performed using morphological analysis, which is a general technique. Morphological analysis is a general natural language processing technique for dividing a sentence into morphemes using a word dictionary as an information source and discriminating the part of speech of each morpheme. Words that are not stored in the word dictionary, such as new words and typographical errors, are unknown words because the part of speech cannot be determined. In this embodiment, the word dictionary uses the word DB 201, and words that are not stored here and whose part of speech cannot be determined are determined as unknown words. When a word is extracted from the input of step 103, it is as shown in FIG. 7, and this is output. For example, “tyf” misrecognized in step 103 and “tena” generated by the interruption of the word “antenna” at the end of the imaging region are unknown words because they do not exist.

  Step 105) The character area determination unit 303 receives the output of Step 104 as an input, and outputs a result of determining whether the shooting area is a character area or a non-character area.

  At the time of determination, as shown in FIG. 8, it is determined whether the proportion of words determined to be unknown words input from the word extraction unit 302 is less than a specified value X (step 1051). If so (step 1051, Yes), the determination result is a character area (step 1052), and if it is equal to or greater than the specified value X (step 1051, No), the determination result is a non-character area (step 1053). ). Thus, when the ratio of words determined as unknown words in the input falls below a specified value, it is determined as a character area, and when it exceeds a specified value, it is determined as a non-character area. In the present embodiment, the specified value is 50%, and the word determined as an unknown word in FIG.

  Step 106) The server-side data transmission / reception unit 301 receives the determination result of step 105 as an input, and outputs it to the client unit 400 through the network.

  Step 107) The client-side data transmission / reception unit 402 receives and outputs the determination result of step 105 as an input.

  Step 108) The determination result presentation unit 403 announces by voice that “it is a character area” if the determination result is a character area, and “non-character area” if it is a non-character area.

[Second Embodiment]
In this embodiment, when a character area is determined, character recognition is performed on the captured area, and the character area is determined when the number of unknown words and the number of one-character words are less than a predetermined value. .

  This embodiment shows an example of a visually impaired support system that reads the position by voice when a camera-equipped mobile phone is held over a book using the method described in the first embodiment and the background art.

  FIG. 9 shows a system configuration in the second embodiment of the present invention. In the figure, the same components as those in FIG.

  The server unit 300 in FIG. 9 can be realized by a device such as a PC server, and includes a server-side data transmission / reception unit 301, a word extraction unit 302, a character area determination unit 303, a search inquiry unit 304, a content search unit 305, a position DB 306, and a content DB 307. Consists of In the position DB 306, data as shown in FIG. 10 is created and stored by the method described in the background art. In the content DB 307, it is assumed that an audio file that reads out each area is stored as shown in FIG. 11 for each area in the book.

  The client unit 400 in FIG. 9 can be realized by a camera-equipped mobile phone or the like, and includes a document photographing unit 401, a client-side data transmission / reception unit 402, and a content presentation unit 403.

  The document 100, the optical character recognition device 200, and the word DB 201 in the figure are the same as those in the first embodiment.

  The operation in the above configuration will be described below.

  FIG. 12 is a sequence chart of the operation in the second embodiment of the present invention.

  Step 200: Input Step) The document photographing unit 401 of the client unit 400 photographs a partial image of a document to be determined and transmits the partial image to the server unit 300 via the client side data transmitting / receiving unit 402. Detailed processing will be described in steps 201 to 203 below.

  Step 210: Word Extraction Step) The server-side data transmission / reception unit 301 of the server unit 300 acquires a partial image of a document from the client unit 400, and the character region determination unit 303 extracts words from the shooting region. Detailed processing will be described in step 204 below.

  Step 220: Character Area Determination Step) The character area determination unit 303 determines whether or not the shooting area is a character area. Detailed processing will be described in step 205 below.

  Step 230: Output Step The character area determination unit 303 outputs the content based on the determination result as to whether or not the shooting area is a character area. Detailed processing will be described in steps 206 to 209 below.

  Details of the above processing will be described below.

  Step 201) The same processing as in the above-described step 101 is performed, and an image file is output as shown in FIG.

  Step 202) The same processing as in Step 102 described above is performed.

  Step 203) The same processing as step 103 is performed. Since it is not considered at all whether the object to be reflected is a character area or a non-character area, the output is as shown in FIG. This is because if a typical OCR software recognizes a figure area that does not contain characters as a character and recognizes it as an unknown word, it can be recognized as an unknown word or from one character such as “.”, “•”, or “1”. This is because it may be recognized as an enumerated word sequence.

  Step 204) The word extraction unit 302 receives the recognition of Step 203 as an input, and extracts and outputs the word. Here, as in step 103, word extraction is performed using morphological analysis. Also in the present embodiment, the word dictionary is stored in the word DB 201 in the format of FIG. 3, and a word that is not stored here and whose part of speech cannot be determined is determined as an unknown word. When a word is extracted from the input in step 203, it is as shown in FIG. 15, and this is output to the character area determination unit 303.

  Step 205) The character area determination unit 303 receives the output of step 204 as an input, and outputs the result of determining whether the shooting area is a character area or a non-character area and the word group extracted in step 204. The determination is made according to the flowchart of FIG. First, in the case where the ratio of words determined to be unknown words is less than the prescribed value X (step 2051, Yes), the character area (provisional), if it is greater than the prescribed value X ( Step 2051, No) It determines with a non-character area | region. In the present embodiment, the specified value X is 50%, and the number of words determined as unknown words in FIG. 15 is less than 50%. Next, if the ratio of words consisting of one character is less than the prescribed value Y (step 2052, Yes), the character area is set (step 2053). If the ratio is greater than the prescribed value Y (step 2052, No), the non-character area is Determination is made (step 2054). In the present embodiment, the prescribed value Y is 80%, and the words that are determined to be one-character words in FIG. 15 are 80% or more, and thus are determined as “non-character areas”.

  Step 206) The search inquiry unit 304 accepts the determination result of step 205 as an input, and inquires the content according to the flowchart of FIG. If the input determination result is “character region” (step 2061, Yes), the same input word group is input to the content search unit 305 to acquire content (an audio file reading the shooting region), The content is output (step 2062). On the other hand, if the input determination result is “non-character area” (step 2061, No), an audio file (content) read out as “no character is there” is output (step 2063). Here, since the determination result of the input in step 206 is “non-character area”, an audio file read out as “there is no character there” is output as the content.

  Step 207) The server-side data transmission / reception unit 301 receives the searched content or the content indicating the non-character area in Step 206, and outputs it to the client unit 400 through the network.

  Step 208) The client-side data transmission / reception unit 402 receives the content of step 207 and outputs it to the content presentation unit 403.

  Step 209) The content presentation unit 403 receives the output of step 208 as an input, and reproduces the audio file that is the content. Here, since an audio file “There is no character there” is reproduced, the user can make a determination to shoot another area.

[Third Embodiment]
In this embodiment, when a character area is determined in the second embodiment, a content search is performed using a word (noun) included in the character area.

  In this embodiment, a case where a character area is photographed with the same configuration as that of the second embodiment will be described.

  FIG. 18 is a sequence chart of operations in the third embodiment of the present invention.

  Step 301) The same processing as in step 201 in the second embodiment described above is performed, but here, since the character area is photographed, it is assumed that the output as shown in FIG. 5 is performed.

  Step 302) The same processing as step 202 is performed.

  Step 303) The same processing as step 203 is performed. The output is as shown in FIG.

  Step 304) The same processing as step 204 is performed. The output is as shown in FIG.

  Step 305) The same processing as step 205 is performed. In the present embodiment, the prescribed value X is 50% and the prescribed value Y is 80%. In FIG. 7, the word determined as an unknown word is less than 50%, and the word determined as a single character is 80%. Therefore, it is determined as a “character area”.

  Step 306) Since the input determination result in step 305 is “character region”, the same input word group is input to the content search unit 305. The content search unit 305 makes an inquiry to the position DB 306 using each noun included in the input word group as a key, and totals the obtained results (positions of each noun) as shown in FIG. (In this case, book A3 page).

  Step 307) Next, the content DB 307 is inquired by using the specified position as a key, and the obtained result (in this case, “sound file reading the book A3 page”) is output.

  Step 308) The same processing as step 207 is performed.

  Step 309) The same processing as step 208 is performed. Here, an audio file that reads out the captured area is reproduced, so that the user can understand the contents of the area.

[Fourth Embodiment]
This embodiment is a partial modification of the second embodiment. Reference 1 “Mazunori Mano, Hideyuki Mizuno, Hideharu Nakajima, Noboru Miyazaki, Akihiro Yoshida: Text-to-speech synthesis technology that realizes real voice responses to customers. "Cralinet" Telecommunications Association NTT Technical Journal 18 (11), pp.19-22, November 2006. This is a visually impaired person support system that reads out the position by voice when a camera-equipped mobile phone is held over a book using a technique such as "".

  FIG. 20 shows a system configuration in the fourth embodiment of the present invention. In the figure, the same components as those in FIG. 9 are denoted by the same reference numerals, and the description thereof is omitted.

  FIG. 21 is an operation sequence chart according to the fourth embodiment of the present invention.

  Step 401) The same processing as in Step 201 in the second embodiment described above is performed, but here, since a wide character area is photographed, it is assumed that an output as shown in FIG. 22 is performed.

  Step 402) The same processing as step 202 is performed.

  Step 403) The same processing as step 203 is performed.

  Step 404) The same processing as step 204 is performed.

  Step 405) The same processing as step 205 is performed. In this embodiment, it is determined as a “character area”.

  Step 406) The content creation unit 306 receives the output of step 405 as an input, and outputs the content created based on the input. If the input determination result is “character region”, the content of the text obtained by concatenating the input word groups is converted into an audio file read out as speech using the technique of the above-mentioned document 1 or the like. On the other hand, if the input determination result is “non-character area”, an audio file that reads out “There is no character there” is output.

  Step 407) The same processing as step 207 is performed.

  Step 408) The same processing as step 208 is performed. Here, an audio file that reads out the captured area is reproduced, so that the user can understand the contents of the area.

  In addition, it is possible to construct | assemble the operation | movement of each component in said 1st-4th embodiment as a program, to install in the computer utilized as a determination apparatus, or to distribute | circulate via a network.

  In addition, the constructed program can be stored in a portable storage medium such as a hard disk, a flexible disk, or a CD-ROM, and can be installed or distributed in a computer.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

100 Document 200 Optical Character Recognition Device 201 Word DB
300 server unit 301 server side data transmission / reception unit 302 word extraction unit 303 character recognition determination unit 304 content inquiry unit 306 content creation unit 305 content search unit 400 client unit 401 document photographing unit 402 client side data transmission / reception unit 403 determination result presentation unit

Claims (13)

Input a whole area or a partial area of a document that contains text, a photograph, a figure, a table, a ruled line, or other non-character text. A determination device that determines whether a character area is included,
An area input means for receiving an input of an area to be determined;
As a result of performing optical character recognition processing assuming that characters are described in the determination target area, at least a character area or a non-character based on the number of unknown words detected and the ratio of the number of words included in the area Character area determination means for determining whether the area;
A determination result output means for outputting either a character area or a non-character area based on the determination result of the character area determination means;
The determination apparatus characterized by having. The character area determination means includes:
Means for determining a character area when the number of detected unknown words is less than a predetermined value A and the area includes a word greater than or equal to a predetermined value B;
The determination apparatus according to claim 1. The character area determination means includes:
The ratio of the number of detected unknown words that is less than a predetermined value A and the number of words in which the length of the detected word is less than or equal to a predetermined value C is less than a predetermined value D, and Means for determining a character area when the area includes a word equal to or greater than a predetermined value B;
The determination apparatus according to claim 1. Input a whole area or a partial area of a document that contains text, a photograph, a figure, a table, a ruled line, or other non-character text. A determination device that determines whether a character area is included,
An area input means for receiving an input of an area to be determined;
As a result of performing optical character recognition processing assuming that characters are described in the determination target area, the character area or non-character is based on at least the length of the detected word and the ratio of the number of words included in the area Character area determination means for determining whether the area;
A determination result output means for outputting either a character area or a non-character area based on the determination result of the character area determination means;
The determination apparatus characterized by having. The character area determination means includes:
The number of words whose length of the detected word is less than or equal to a predetermined value C is less than a predetermined value D, and includes means for determining a character area when the area includes a word greater than or equal to a predetermined value B. The determination apparatus described. 6. The determination apparatus according to claim 3, wherein the predetermined value C of the word length is one character. Input a whole area or a partial area of a document that contains text, a photograph, a figure, a table, a ruled line, or other non-character text. A determination method for determining whether a character area is included,
An area input step for receiving an input of an area to be determined;
As a result of performing optical character recognition processing assuming that characters are described in the determination target area, at least a character area or a non-character based on the number of unknown words detected and the ratio of the number of words included in the area A character region determination step for determining whether the region is a region;
A determination result output step for outputting either a character area or a non-character area based on the determination result of the character area determination step;
The determination method characterized by performing. In the character region determination step,
When the number of detected unknown words is less than a predetermined value A and the area includes a word that is greater than or equal to a predetermined value B, the character area is determined.
The determination method according to claim 7. In the character region determination step,
The ratio of the number of detected unknown words that is less than a predetermined value A and the number of words in which the length of the detected word is less than or equal to a predetermined value C is less than a predetermined value D, and When the area includes a word that is equal to or greater than a predetermined value B, it is determined as a character area.
The determination method according to claim 7. Input a whole area or a partial area of a document that contains text, a photograph, a figure, a table, a ruled line, or other non-character text. A determination method for determining whether a character area is included,
An area input step for receiving an input of an area to be determined;
As a result of performing optical character recognition processing assuming that characters are described in the determination target area, the character area or non-character is based on at least the length of the detected word and the ratio of the number of words included in the area A character region determination step for determining whether the region is a region;
A determination result output step for outputting either a character area or a non-character area based on the determination result of the character area determination step;
The determination method characterized by performing. In the character region determination step,
The determination according to claim 10, wherein the number of words having a length of the detected word equal to or less than a predetermined value C is less than a predetermined value D, and the character area is determined when the area includes a word having a predetermined value B or more. Method. The determination method according to claim 9 or 11, wherein the predetermined value C of the length of the word is one character.   The program for functioning a computer as each means which comprises the determination apparatus of any one of Claims 1 thru | or 6.

Images