JP3669832B2 - Character recognition device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a character recognition device that extracts a character or a character string selected by a selection mark as a recognition result.
[0002]
[Prior art]
2A and 2B are explanatory diagrams of a mark field of a conventional form.
In a character recognition apparatus that extracts a character or character string selected by a mark as a recognition result, as shown in FIG. 2A, a form (drop) in which a mark entry frame 1 and a guidance character 2 are printed in a dropout color in advance. Out-of-color forms) were often used. The mark entry frame 1 designates a place where a mark is to be entered, and a guidance character 2 is printed inside the mark entry frame 1. The guidance character 2 is used to display a selection item, and there are a case where it is singular and a case where it is plural. As the dropout color, blue and red colors are generally adopted. A mark 3 such as “◯” or “ν” is attached to the mark field of the form on which the character string of the mark entry frame 1 and the guidance character 2 is printed by a black writing tool. The character recognition device extracts “Showa” of the character string of the guidance character 2 in the mark entry frame 1 with the mark 3 as a recognition result.
[0003]
For example, a portion printed in a red-colored dropout color does not remain in the input image by setting the light source of the image input unit constituting the character recognition device to be red, and the mark 3 is a mark entry frame 1. And touching the guidance character 2 had little effect on recognition.
However, in recent years, the number of character recognition devices that use a facsimile machine or a general-purpose image scanner as an image input unit has increased. In such a character recognition device, it is difficult to adopt a dropout color due to an optical problem. Therefore, as shown in FIG. Dropout color form) is used. Since a non-dropout color form is laid out by a word processor or the like and printed by a general printer, not only the mark entry frame 4 but also the guidance characters 5 are printed in black. Therefore, in consideration of the influence on recognition, the guidance character 5 cannot be printed inside the mark entry frame 4. Therefore, the guidance character 5 and the surrounding character frame 6 are printed outside the mark entry frame 4. It was. The character recognition device extracts, for example, “Showa” of the guidance character 5 outside the mark entry frame 4 with the mark 3 as a recognition result.
[0004]
[Problems to be solved by the invention]
  However, the conventional character recognition device that recognizes a non-dropout color form has the following problems.
  In non-dropout color forms, guidance is provided outside the mark entry frame 4.characterSince 5 is printed, it occupies more space on the form than in the case of a dropout color form. In addition, a character frame 6 is required for each item, and there are many selection items or guidance for each selection item.characterWhen the number of characters of 5 is different, the field setting on the form is less flexible. Therefore, a large restriction is imposed when the form is laid out.
[0005]
[Means for Solving the Problems]
  In order to solve the above-described problems, a first invention of the present invention is configured as follows in a character recognition device.
  In other words, a plurality of selection items each composed of an arbitrary number of guidance characters are separated into a plurality of cells and printed, and a form image selectively marked is input to the plurality of selection items. A selection mark definition information storage memory for preliminarily storing mark information in which position information of the plurality of selection items in the form and character strings of the guidance characters constituting the selection items are respectively defined; A character string of the guidance character on the image of the form for each cellCut out the guidance characterIt is determined whether or not the character string definition for each selection item can be configured based on the recognition result for each cell obtained by the character recognition unit and the character recognition unit obtained from the character recognition unit, and based on the determination result Selection item with the markThe string ofAnd a selected character string extracting unit for extracting.
[0006]
  By adopting such a configuration, the character recognition unit recognizes the character string of the guidance character in each cell. In this recognition, the recognition result corresponding to the selection item with the mark becomes misread or unrecognizable. Therefore, the selected character string extraction unit determines whether or not the definition of the character string can be constituted by the recognition result, thereby selecting the marked selection item.StringCan be extracted.
[0007]
  According to a second aspect of the present invention, a character recognition device has the following configuration.
  That is, a plurality of selection items each composed of an arbitrary number of guidance characters are printed separately in a common cell, and an image of a form that is selectively marked is displayed on the plurality of selection items. An image input unit for inputting; a selection mark definition information storage memory for storing in advance mark information in which position information of the plurality of selection items in the form and character strings of the guidance characters constituting the selection items are respectively defined; , The character string of the guidance character on the image of the form for each lineCut out the guidance characterA character recognition unit for recognizing and determining whether or not the line-by-line recognition result obtained by the character recognition unit can constitute the definition of the character string for each of the selection items, and based on the determination result, the mark Selection items marked withThe string ofAnd a selected character string extracting unit for extracting.
[0008]
  By adopting such a configuration, the guidance character strings in the common cell areeveryIs recognized by the character recognition unit. In this recognition, the recognition result corresponding to the selection item with the mark becomes misread or unrecognizable. Therefore, the selected character string extraction unit determines whether or not the definition of the character string can be constituted by the recognition result, thereby selecting the marked selection item.StringCan be extracted.
[0009]
  According to a third aspect of the present invention, a character recognition apparatus has the following configuration.
  That is, a plurality of selection items each composed of an arbitrary number of guidance characters are arranged and printed in a common cell with a delimiter between them, and the selection items are selectively marked. An image input unit for inputting an image of the image, and mark information in which the position information of the plurality of selection items in the form, the location of the delimiters, and the character strings of the guidance characters constituting the selection items are defined in advance. A selection mark definition information storage memory to store, a character string of the guidance character on the image of the form, and the delimiterCut out the guidance character and the delimiterRecognizing the character recognition unit, collating the recognition result in the cell obtained by the character recognition unit with the location of the delimiter and the definition of the character string for each selection item, and selecting each of the selections in the image It is determined in turn whether or not the character string of the guidance character of the item has been recognized, and based on the result of the determination, the selection item with the markThe string ofAnd a selected character string extracting unit for extracting.
[0010]
  By adopting such a configuration, a character string of guidance characters in a common cell is recognized by the character recognition unit. In this recognition, the recognition result corresponding to the selection item and the delimiter with the mark is misread or unrecognizable. Therefore, the selected character string extraction unit collates the location of the delimiter and the definition of the character string with the recognition result, thereby selecting the marked selection item.StringCan be extracted.
[0011]
  According to a fourth aspect of the present invention, a character recognition apparatus has the following configuration.
  That is, a plurality of selection items each composed of an arbitrary number of guidance characters are arranged and printed in a common cell with a delimiter between them, and the selection items are selectively marked. An image input unit for inputting an image of the image, and mark information in which the position information of the plurality of selection items in the form, the location of the delimiters, and the character strings of the guidance characters constituting the selection items are defined in advance. Selection mark definition information storage memory to be stored, a character string of the guidance character on the form image, and the delimiterAnd the guidance character and the delimiterA character recognition unit for recognizing the character recognition unit, a result of recognition in the cell obtained by the character recognition unit and a definition of the character string for each selection item, and a guidance character of each selection item in the image It is determined in order whether or not the character string has been recognized, and based on the result of the determination, the selection item with the markThe string ofAnd a selected character string extracting unit for extracting.
[0012]
  By adopting such a configuration, a character string of guidance characters in a common cell is recognized by the character recognition unit. In this recognition, the recognition result corresponding to the selection item with the mark becomes misread or unrecognizable. Therefore, the selected character string extraction unit collates the definition of the character string and the recognition result in order, thereby selecting the marked selection item.StringCan be extracted.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
First embodiment
  FIG. 1 is a block diagram of a character recognition apparatus showing a first embodiment of the present invention.
  This character recognition device extracts a selection item with a mark from among a plurality of selection items printed in a selection mark field of a form in a non-dropout color, and is constituted by a scanner or a facsimile device. An image input unit 10 for inputting a form image; and an image data storage memory 11 for storing a form image given from the image input unit 10 via the bus B;TheHave. The bus B further includes selection mark definition information in which necessary information such as the position information of the selection mark field in the form to be read, the number of selection items, and the definition of the character string of the guidance character constituting each selection item is stored in advance. A storage memory 12, a character recognition unit 13, a recognition result storage memory 14, a selected character string extraction unit 15, and a selection mark character string extraction result storage memory 16 are connected.
[0014]
  The character recognition unit 13 cuts out a line in a specified area in the selection mark field, cuts out a character from the line, and performs character recognition of the character. The recognition result storage memory 14 stores the result recognized by the character recognition unit 13. The selected character string extraction unit 15 determines from the recognition result stored in the recognition result storage memory 14 whether the definition of the character string of the guidance character can be configured for each selected item, and based on the determination result, which selection item Marked withButThe selection item with the mark is extracted by determining whether it has been attached. The selection mark character string extraction result storage memory 16 stores the character string of the guidance character of the selection item extracted by the selection character string extraction unit 15, and externally from the selection mark character string extraction result storage memory 16 A character string is read out.
[0015]
FIG. 3 is an explanatory diagram illustrating an example of extraction of the first selection item performed by the character recognition device of FIG. FIG. 4 is an explanatory diagram illustrating an example of extracting the second selection item performed by the character recognition device of FIG. With reference to FIG. 3 and FIG. 4, the operation when the character recognition device of FIG. 1 extracts selection items with marks will be described.
As shown in FIGS. 3 and 4, the image input unit 10 of the character recognition apparatus in FIG. 1 receives a form image on which, for example, four cells 21 surrounded by a non-dropout color black line are printed. The In each cell 21 of the form, for example, selection items 23 of “Meiji”, “Taisho”, “Showa”, and “Heisei” composed of a character string of guidance characters 22 are printed in black. The form image is stored in the image data storage memory 11. The character recognition unit 13 extracts the position information of the selection mark field from the selection mark definition information storage memory 12 and performs character recognition, and outputs a character recognition result for each cell 21 based on the position information.
[0016]
For example, as shown in FIG. 3, when an image of a form in which “24” is added to “Showa” among the selection items 23 is input, the “Meiji” of “Meiji” is first recognized by the character recognition unit 13. ", The recognition result of the first candidate" sum "and the recognition result of the second candidate" bright "are obtained, and" ji "is the recognition result of" ji ". These recognition results are stored in the recognition result storage memory 14. Similarly, “Taisho” and “Major” recognition results are obtained, and the recognition results are stored in the recognition result storage memory 14. For “Showa”, since both the guidance characters 22 are overlaid, the “Showa” is cut out as one character, and the recognition result becomes an unrecognizable “?”, Which is also stored in the recognition result. Stored in the memory 14. In “Heisei”, recognition results of “flat” and “sei” are obtained, and the recognition results are stored in the recognition result storage memory 14.
[0017]
The selected character string extraction unit 15 reads the number of selection items 23 (= 4) and the definition of the character string of the guidance character 22 for each selection item 23 from the selection mark definition information storage memory 12 and stores them in the recognition result storage memory 14. The recognition results for each stored cell 21 are collated. As the definition of the guidance character string, information in the order of “Meiji”, “Taisho”, “Showa”, and “Heisei” is stored in advance. With respect to “Meiji” of the selection item 23, the definition of “Meiji” of the selection item 23 can be configured from the “sum”, “Meiji”, and “Oji” recognition results, and therefore, it is determined that there is no mark. With regard to the selection items 23 “Taisho” and “Heisei”, the character string of the defined guidance character 22 can be formed from the recognition result, so it is determined that there is no mark. Regarding the selection item 23 “Showa”, since the recognition result is one “?”, The number is insufficient, and the character string of the defined guidance character cannot be formed. Therefore, the character string of the guidance character 22 of “Showa” is determined to have a mark and stored in the selected mark character string extraction result storage memory 16.
[0018]
As shown in FIG. 4, for example, when an image of a form with a mark 24 biased to “sum” of the guidance character “Showa” is input, the character recognition result of “Akira” of the guidance character Although “Akira” is obtained, the character recognition result of “Japanese” cannot be obtained, so that the character string of the defined guidance character 22 cannot be formed. Therefore, the character string of the “Showa” guidance character 22 is determined to be marked and stored in the selected mark character string extraction result storage memory 16.
[0019]
As described above, in the first embodiment, the character recognition unit 13, the selection mark definition information storage memory 14 that stores the definition of the character string of the guidance character 22 for each selection item 23, and the selected character string extraction unit 15 and the selection item 23 with the mark 24 is extracted from the recognition result of the character string of the guidance character 22 for each cell 21 depending on whether or not the definition can be configured. Even if the guidance character 22 is printed in the non-dropout color, the selection item 23 with the mark 24 can be extracted sufficiently. Therefore, the layout of the form becomes easy. Furthermore, it becomes possible to flexibly adapt to fluctuations when the mark 24 is written.
[0020]
Second embodiment
FIG. 5 is a block diagram of a character recognition device showing a second embodiment of the present invention.
This character recognition device extracts a selection item with a mark from among a plurality of selection items printed in a selection mark field of a form in a non-dropout color, and is constituted by a scanner or a facsimile device. An image input unit 30 for inputting a form image and an image data storage memory 31 for storing the form image given from the image input unit 30 via the bus B are provided. The bus B further includes selection mark definition information in which necessary information such as the position information of the selection mark field in the form to be read, the number of selection items, and the definition of the character string of the guidance character constituting each selection item is stored in advance. A storage memory 32, a character recognition unit 33, a recognition result storage memory 34, a selected character string extraction unit 35, and a selection mark character string extraction result storage memory 36 are connected.
[0021]
The character recognition unit 33 cuts out a line in a designated area in the selection mark field, cuts out a character from the line, and recognizes the character. The recognition result storage memory 34 stores the recognition result in units of lines recognized by the character recognition unit 33. The selected character string extraction unit 35 determines whether the definition of the character string of the guidance character can be configured for each selection item from the recognition result stored in the recognition result storage memory 34, and based on the determination result, which selection item Is selected, and the selection item with the mark is extracted. The selection mark character string extraction result storage memory 36 stores the character string of the guidance character of the selection item extracted by the selection character string extraction unit 35, and externally from the selection mark character string extraction result storage memory 36. A character string is read out.
[0022]
FIG. 6 is an explanatory diagram illustrating an example of selection item extraction performed by the character recognition device of FIG. 5. With reference to FIG. 6, the operation when the character recognition device of FIG. 5 extracts selection items with marks will be described.
As shown in FIG. 6, the image input unit 30 of the character recognition apparatus shown in FIG. 5 receives an image of a form on which, for example, a cell 41 surrounded by a non-dropout color black line is printed. In the cell 41 of the form, for example, “Meiji”, “Taisho”, “Showa”, and “Heisei” selection items 43 made up of character strings of guidance characters 42 are printed separately in black. . The form image is stored in the image data storage memory 31. The character recognition unit 33 extracts the position information of the selected mark field from the selection mark definition information storage memory 32, performs character recognition, outputs the recognition result in the cell 41 based on the position information, and stores the recognition result. Store in the memory 34.
[0023]
The selected character string extraction unit 35 reads the number of selection items 43 (= 4) and the definition of the character string of the guidance character 42 for each selection item 43 from the selection mark definition information storage memory 32, and the recognition result storage memory 34. The recognition result of the cell 41 stored in is checked. As the definition of the character string of the guidance character 42, information in the order of “Meiji”, “Taisho”, “Showa”, “Heisei” is stored as in the first embodiment. As shown in FIG. 6, when the mark 44 is added to “Showa” in the third row, “Meiji” in the selection item 43 in the first row is changed from “Meiji” in the recognition result to “ Since the definition of “Meiji” can be configured, it is determined that there is no mark. Regarding the selection items 43 “Taisho” and “Heisei”, the character string of the defined guidance character can be constructed from the recognition result, so that it is determined that there is no mark. With respect to “Showa” of the selection item 43, since the recognition result is one “?”, It is determined that the character string of the guidance character of the defined selection item cannot be configured due to the lack of the number. Accordingly, the character string of the “Showa” guidance character 42 is extracted as a selection item with the mark 44, and the character string of the “Showa” guidance character 42 is stored in the selection mark character string extraction result storage memory 16. The
[0024]
As described above, in the second embodiment, the character recognition unit 33, the selection mark definition information storage memory 34 that stores the definition of the character string of the guidance character 42 for each selection item 43 separated from the line, and the selection A character string extraction unit 35, and the selection item 43 with the mark 44 is extracted depending on whether or not the definition can be configured by the character recognition result of the character string of the guidance character 42 for each line in the cell 41. Therefore, the marked selection item 43 can be extracted from the form in which the selection items 43 of a plurality of lines are printed in the cell 41 in a non-dropout color.
[0025]
Third embodiment
FIG. 7 is a block diagram of a character recognition device showing a third embodiment of the present invention.
This character recognition device extracts a selection item with a mark from among a plurality of selection items printed in a selection mark field of a form in a non-dropout color, and is constituted by a scanner or a facsimile device. An image input unit 50 for inputting a form image and an image data storage memory 51 for storing a form image given from the image input unit 50 via the bus B are provided. The bus B further stores in advance necessary information such as the position information of the selection mark field in the form to be read, the number of selection items, the location of the delimiter symbols, and the definition of the character strings of the guidance characters that constitute each selection item. The selection mark definition information storage memory 52, the character recognition unit 53, the recognition result storage memory 54, the selection character string extraction unit 55, and the selection mark character string extraction result storage memory 56 are connected.
[0026]
The character recognition unit 53 cuts out a line in a specified area in the selection mark field, cuts out a character from the line, and performs character recognition of the character. The recognition result storage memory 54 stores the result recognized by the character recognition unit 53. The selected character string extraction unit 55 determines whether the definition of the character string of the guidance character can be configured for each selection item from the recognition result stored in the recognition result storage memory 54, and based on the determination result, which selection item Is selected, and the selection item with the mark is extracted. The selection mark character string extraction result storage memory 56 stores the character string of the guidance character of the selection item extracted by the selection character string extraction unit 55, and externally from the selection mark character string extraction result storage memory 56. A character string is read out.
[0027]
FIG. 8 is an explanatory diagram illustrating an example of extraction of the first selection item performed by the character recognition device of FIG.
As shown in FIG. 8, the image input unit 50 of the character recognition apparatus shown in FIG. 7 receives a form image on which, for example, a cell 61 surrounded by a non-dropout color black line is printed. In the cell 61 of the form, for example, a selection item 63 of “Meiji”, “Taisho”, “Showa” and “Heisei” composed of a character string of guidance characters 62 is placed on one line with a delimiter “•” in between. Is printed in black. The form image is stored in the image data storage memory 51. The character recognition unit 53 extracts the position information of the selection mark field from the selection mark definition information storage memory 52, performs character recognition, and performs character recognition for each line based on the position information.
[0028]
For example, as shown in FIG. 8, when an image of a form in which a mark 64 is added to “· Showa” in the selection items 63 is input, the character recognition unit 53 recognizes the first “Meiji” “ In “Ming”, the character recognition result of “Ming” is obtained. The character recognition result of “Oji” is obtained for “Oji”. A character recognition result of “•” is obtained for the delimiter “•”. Similarly, when character recognition is performed, “Meiji, Taisho? Here, “· Showa” is cut out as one character by entering the mark 64, and the character recognition result is “?” Which cannot be recognized. These recognition results are stored in the recognition result storage memory 54.
The selected character string extraction unit 55 receives information from the selection mark definition information storage memory 52 that the number of selection items 63 (= 4) and the mark method, that is, each selection item separated by a delimiter is marked. And the definition of the character string of the guidance character 62 including the delimiter “·” of the selection item 63 are read out and collated in order with the recognition result of the cell 61 stored in the recognition result storage memory 54.
[0029]
In this collation, “Meiji” and “Taisho” of the selection items 63 coincide with the recognition results “Meiji” and “Taisho”, so it is determined that there is no mark. The next recognition result of “Meiji” and “Taisho” is “? · Pair”, but “·” is a delimiter, so that the recognition result that is actually valid is “Pair”. Here, it can be determined from the definition that the “hira” of the recognition result is the first character of “Heisei” in the selection item 63, so that “Showa” that should exist between them cannot be recognized. Therefore, it is determined that the mark 64 is attached to “Showa” of the selection item 63, and the character string of the guidance character 62 corresponding to “Showa” of the selection item 63 is stored in the selected character string extraction result over-storage memory 56. Stored.
[0030]
  FIG. 9 is an explanatory diagram illustrating an example of extraction of the second selection item performed by the character recognition device of FIG.
  FIG. 8 illustrates an example in which the selected character string extraction unit 55 extracts the selection item 63 with the mark 64 based on the collation including the delimiter “·”. You may perform collation without referring to "." Hereinafter, an extraction example of the selection item 63 in this case will be described.
  For example, as shown in FIG. 9, when a mark is added to “sum” of “Showa” in the selection item 63, the recognition result in the cell 61 is “Meiji / Taisho / “Aki? ・ Heisei”. These results are stored in the recognition result storage memory 54. The selected character string extraction unit 55 reads the number of selection items 63 (= 4) and the definition of the character string of the guidance character 62 of each selection item 63 from the selection mark definition information storage memory 52, and the recognition result storage memory 54. The recognition results of the cells 61 stored in are sequentially verified. In this way, only “Akira” is obtained as the recognition result for “Showa” of the selection item 63, and the recognition result for “sum” is “?”. Therefore, the character string “Showa” of the defined guidance character 62 is composed.KiSince there is no mark, it is determined that “Showa” was marked. The selection item 63 “Heisei” cannot be composed of “?” Next to “Aki”, but can be composed of recognition results “Pair” “Nari” corresponding to the next two characters by skipping that character. Therefore, it can be determined that there is no mark.
[0031]
As described above, in the third embodiment, the recognition result in the cell 61 obtained by the character recognition unit 53 is collated with the location of the delimiter “·” and the definition of the character string for each selection item 63. Then, it is determined in order whether or not the character string of the guidance character 62 for each selection item 63 has been recognized in the image, or the separator “·” is excluded from the target of collation, and the guidance character 62 for each selection item 63 is determined. Since the selected character string extraction unit 55 that sequentially determines whether or not the character strings can be recognized is provided, a selection item with a mark is added to a form in which a plurality of selection items 63 are printed in a common cell 61. 63 can be extracted. Therefore, the layout of the form becomes easy.
In addition, this invention is not limited to the said embodiment, A various deformation | transformation is possible.
For example, in the first to third embodiments, the marks 24, 44, 64 are “◯”, and the selection items 23, 43, 63 to which the marks 24-64 are added have been described. There are also forms in which the guidance characters 22, 42, 62 are erased by the marks 24 to 64 of “=”. Also in this case, selection items can be easily extracted by applying the techniques of the first to third embodiments. For example, instead of the selection item determined after marking, the character string of the selection item determined to have no mark may be stored in the selection mark character string extraction result storage memories 16, 36, and 56.
[0032]
【The invention's effect】
  As described above in detail, according to the first invention, the selection mark that stores in advance the mark information in which the position information of the plurality of selection items in the form and the character strings of the guidance characters constituting the selection items are respectively defined. A selection character string that determines whether the definition of the character string for each selection item can be configured based on the recognition result for each cell, the definition information storage memory, the character recognition unit, and extracts the marked selection item Since there is an extractor, a selection item with a mark is attached even if a drop-out color is not used for a form in which multiple selection items are separated into multiple cells and printed.StringCan be extracted, and the layout of the form becomes easy.
  According to the second invention, the selection mark definition information storage memory and the line-by-line recognition result on the form image determine whether or not the character string definition for each selection item can be configured, and based on the determination result Since there is a selection character string extraction unit that extracts selection items with a mark, dropout color is not used for forms in which multiple selection items are printed in separate cells in a common cell But the selection item markedStringCan be extracted, and the layout of the form becomes easy.
[0033]
  According to the third invention, the selection mark definition information storage memory, the recognition result in the cell, the location of the delimiter and the definition of the character string for each selection item are collated, and the guidance character of each selection item in the image Since a selection character string extraction unit that sequentially determines whether or not a character string can be recognized and extracts a selection item with a mark is provided, a plurality of selection items are arranged and printed in a common cell. Even if you do not use the dropout color for selected forms, the selected items are markedStringCan be extracted, and the layout of the form becomes easy.
  According to the fourth invention, the selection mark definition information storage memory is compared with the recognition result in the cell and the definition of the character string for each selection item, and the character string of the guidance character of each selection item is recognized in the image. Since a selection character string extraction unit that sequentially determines whether or not the selection has been made and extracts selection items with marks is provided, a plurality of selection items are arranged in a common cell and printed. Even if you do not use the dropout color, the selected item is markedStringCan be extracted, and the layout of the form becomes easy.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a character recognition device according to a first embodiment of the present invention.
FIG. 2 is an explanatory diagram of a mark field of a conventional form.
FIG. 3 is an explanatory diagram illustrating an example of extraction of a first selection item performed by the character recognition device of FIG. 1;
FIG. 4 is an explanatory diagram illustrating an example of extraction of second selection items performed by the character recognition device of FIG. 1;
FIG. 5 is a block diagram of a character recognition device showing a second embodiment of the present invention.
6 is an explanatory diagram showing an example of selection item extraction performed by the character recognition device of FIG. 5; FIG.
FIG. 7 is a block diagram of a character recognition device showing a third embodiment of the present invention.
8 is an explanatory diagram illustrating an example of extraction of a first selection item performed by the character recognition device of FIG. 7;
FIG. 9 is an explanatory diagram illustrating an example of extracting a second selection item performed by the character recognition device of FIG. 7;
[Explanation of symbols]
10, 30, 50 Image input section
11, 31, 51 Image data storage memory
12, 32, 52 Selection mark definition information storage memory
13, 33, 53 Character recognition unit
15, 35, 55 Selected character string extraction unit
21, 41, 61 cells
22, 42, 62 Guidance characters
23, 43, 63 selection items

Claims (4)

A plurality of selection items each composed of an arbitrary number of guidance characters are separated into a plurality of cells and printed, and an image of a form in which a plurality of selection items are selectively marked is input An input section;
A selection mark definition information storage memory for storing in advance mark information in which position information of the plurality of selection items in the form and character strings of the guidance characters constituting the selection items are respectively defined;
A character recognition unit that cuts out a character string of the guidance character on the image of the form for each cell and recognizes the guidance character ;
It is determined whether the definition of the character string for each selection item can be configured based on the recognition result for each cell obtained by the character recognition unit, and the selection with the mark is made based on the determination result A character recognition apparatus comprising: a selected character string extraction unit that extracts the character string of an item. A plurality of selection items each composed of an arbitrary number of guidance characters are printed separately in a common cell, and a form image selectively marked is input to the plurality of selection items. An image input unit;
A selection mark definition information storage memory for storing in advance mark information in which position information of the plurality of selection items in the form and character strings of the guidance characters constituting the selection items are respectively defined;
A character recognition unit that recognizes the guidance character by cutting out the character string of the guidance character on the image of the form in units of lines;
The recognition result of the row obtained by the character recognition unit determines whether it can constitute a definition of the character string of each of the selection item, the selected item attached with the mark on the basis of the result of the determination A character recognition apparatus comprising: a selected character string extraction unit that extracts the character string. A plurality of selection items each composed of an arbitrary number of guidance characters are arranged and printed in a common cell with a delimiter between them, and the plurality of selection items are selectively marked with a form image An image input unit for inputting
A selection mark definition information storage memory for preliminarily storing mark information in which the position information of the plurality of selection items in the form, the location of the separators, and the character strings of the guidance characters constituting the selection items are respectively defined;
A character recognition unit that cuts out the character string of the guidance character and the separator on the image of the form and recognizes the guidance character and the separator;
The result of recognition in the cell obtained by the character recognition unit is collated with the location of the separator and the definition of the character string for each selection item, and the character string of the guidance character of each selection item in the image A character recognition device , comprising: a selection character string extraction unit that sequentially determines whether or not a character string has been recognized and extracts the character string of the selection item with the mark based on the determination result . A plurality of selection items each composed of an arbitrary number of guidance characters are arranged and printed in a common cell with a delimiter between them, and the plurality of selection items are selectively marked with a form image An image input unit for inputting
A selection mark definition information storage memory for preliminarily storing mark information in which the position information of the plurality of selection items in the form, the location of the separators, and the character strings of the guidance characters constituting the selection items are respectively defined;
A character recognition unit that cuts out the character string of the guidance character and the separator on the image of the form and recognizes the guidance character and the separator;
The result of recognition in the cell obtained by the character recognition unit is collated with the definition of the character string for each selection item, and whether or not the guidance character string of each selection item has been recognized in the image A character recognition apparatus comprising: a selected character string extraction unit that sequentially determines the character string of the selection item with the mark based on a result of the determination.

Images