JP3805161B2 - Character reading device, character reading method, and computer-readable information storage medium recording character reading program - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a character reader that optically reads a form or the like on which written characters and non-filled characters (such as character frames) are written, and in particular, binary using a user's arbitrarily set reference regarding color. The present invention relates to a character reading apparatus, a character reading method, and a computer-readable information storage medium storing a character reading program for efficiently reading a necessary part of a form by generating a binarized image.
[0002]
[Prior art]
As an example of an apparatus for reading a form in color and recognizing a necessary part, there is an invention described in JP-A-10-51655. The means described in the publication extracts brightness and saturation from the values read in the three primary colors (colors), and changes the conditions for binarizing the brightness with these extracted values. As a result, the character portions necessary for character recognition are binarized except for chromatic portions such as character frames that are not necessary for character recognition.
[0003]
This means is a configuration for extracting a black character portion excluding a chromatic color portion and binarizing it to recognize a character. For example, recognition of a non-black character, acquisition of a binarized image including a chromatic color portion, and It is not intended for recognition.
[0004]
By the way, when reading and binarizing or recognizing a form, there are cases where it is desired to binarize the description included in each field (partial area to be binarized or recognized) on various forms under various conditions. For example, binarization only for handwritten black characters, binarization only for printed characters (for example, chromatic colors), binarization of non-character parts such as character frames, and binarization described by combining these. Furthermore, even in a single form, there are cases where it is desired to change the binarization or recognition conditions for each field. By realizing these functions, detailed binarization is possible in the selection of descriptions and selection of fields, and the functionality as a character reading device is greatly improved.
[0005]
[Problems to be solved by the invention]
As described above, conventionally, when a form is read and binarized or recognized, the description included in each field on the form cannot be binarized under various conditions in terms of color. In addition, in a single form, binarization or recognition conditions cannot be changed for each field.
[0006]
The present invention relates to a character reading apparatus and a character reading method capable of binarizing a description included in each field on a form under various conditions in color when the form is read and binarized or recognized. Another object of the present invention is to provide a computer-readable information storage medium in which a character reading program is recorded.
[0007]
It is another object of the present invention to provide a character reading device, a character reading method, and a computer-readable information storage medium in which a character reading program is recorded that can change binarization or recognition conditions for each field in a single form. Objective.
[0008]
In addition, when a form is read and binarized or recognized, the validity of the condition is indicated by displaying the result of binarization under various conditions in terms of color for the description contained in each field on the form. An object of the present invention is to provide a character reading device, a character reading method, and a computer-readable information storage medium recording a character reading program.
[0009]
In addition, a character reading device, a character reading method, and character reading that reliably captures the color of a part that is not necessary for binarization or recognition, such as a character frame, and binarizes or recognizes the necessary part with high reliability. An object of the present invention is to provide a computer-readable information storage medium storing a program.
[0010]
It is another object of the present invention to provide a character reading device capable of binarizing a character and a character frame, a character reading method, and a computer-readable information storage medium storing a character reading program.
[0011]
[Means for Solving the Problems]
In order to solve the above problems, a character reading apparatus according to the present invention includes a unit that divides a form into pixels and captures it in color by photoelectric conversion, a unit that specifies a point on the color space, and an origin of the color space. Means for calculating a projection value of the point in the color space corresponding to each captured pixel to the binarization axis with a straight line passing through the designated point as a binarization axis; And a means for binarizing the projected value with a predetermined threshold and a means for processing an image obtained by the binarization.
[0012]
By specifying a point on the color space, the form is binarized based on a specific color standard.
[0013]
The point on the color space is designated for each field constituting the form, and the projection value is obtained for pixels belonging to the field among the captured pixels. And
[0014]
Binarization is performed for each field of the form according to a specific color standard that is appropriate.
[0015]
Moreover, further comprising means for displaying a color image of a form taken by the collar, and a means for specifying a particular position or region in the displayed on the color image, the point on the color space The designation is made by referring to color information possessed by the pixel at the designated specific position or region.
[0016]
A point on the color space for the color standard is more easily designated by referring to the color information of the pixel at the designated specific position or region.
[0017]
Further, the designated point on the color space, performed for a plurality of points, said projected values are worth determined binarization axis each of which is obtained by the designation of the plurality of points, processing the image The means includes a process for obtaining a logical sum or a logical product for a plurality of images obtained by the binarization.
[0018]
Drop out multiple colors or obtain as a binary image.
[0019]
The image processing apparatus further comprises means for displaying an image obtained by the binarization and means for updating and designating a point on the designated color space.
[0020]
By displaying the image obtained by binarization, the validity of the binarization standard can be confirmed. If this is not appropriate, the point specification on the color space is updated.
[0021]
In addition, the color space is a space having hue, saturation, and lightness as components .
[0022]
Specify a point on the color space that matches the human color sensation.
[0023]
The character reading method according to the present invention and the computer-readable information storage medium storing the character reading program include steps for realizing the functions of the above-described means.
[0024]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
[0025]
FIG. 1 is a configuration diagram showing the configuration of a character reading apparatus according to an embodiment of the present invention (FIG. 1A) and the configuration diagram showing the configuration of an OCR (Optical Character Reader) system provided in the character reading apparatus. ((B) of the figure).
[0026]
As shown in FIG. 1A, the character reading apparatus of this embodiment includes a scanner 1 and a PC (personal computer) 2, and an OCR system 3 is provided inside the PC 2. The OCR system 3 includes a form definition program 5, an OCR reading program 4, and an FC (format control) table 6 as shown in FIG.
[0027]
The scanner 1 optically reads a form in color under the control of the PC 2, converts the read information into an electrical signal, and supplies it to the PC 2. The PC 2 is a hardware resource for the OCR system 3 to process the supplied read information. The OCR system 3 processes user input information in addition to the supplied read information, and generates a binary image, character recognition, and the like.
[0028]
The form definition program 5 sets color conditions for reading for each field of the form, and supplies the set conditions to the FC table 6. The OCR reading program 4 refers to the contents of the FC table 6 and performs generation of a binary image, character recognition, and the like using supplied reading information. The FC table 6 becomes a storage destination of the color conditions for reading for each field of the form by combining with the above other elements, and becomes a reference destination for generating a binary image and recognizing characters. is there.
[0029]
Next, the operation of this embodiment will be described.
[0030]
First, a form to be read is placed on the scanner 1, and a color image is acquired by scanning, and information on the form definition is set, that is, a color condition for reading is set. The form definition information setting screen is displayed on the display of the PC 2.
[0031]
FIG. 2 is a diagram illustrating an example of a form definition information setting screen. As shown in the figure, the form definition information setting screen 10 displays a color image 20 of the read form and a field information definition window 30.
[0032]
The color image 20 has a plurality of fields 21, 22, 23, and 24 to which characters are to be read. A field to be defined and input here is selected from the fields, and a rectangular display indicating that the field is being selected. . In this figure, the field 21 is currently selected.
[0033]
The field information definition window 30 includes a field number 31, a reading field selection reception display 32, an image field selection reception display 33, a field start point (start coordinate) display 34, a field end point (end coordinate) display 35, and a font. Display 36, character type display 37, special attribute display 38, binary axis selection display 39, saturation display 40 of the selected binary axis, also hue display 41, brightness display 42, and binary threshold display 43 There is a test binarization acceptance display 44.
[0034]
Here, when a reading field is selected by inputting to the reading field selection acceptance display 32, the field is binarized based on a predetermined standard, and the binarized image is recognized. On the other hand, when an image field is selected by inputting to the image field selection acceptance display 33, the field is binarized based on a predetermined standard, but the character recognition process is not performed. That is, this selection is a selection between processing the field as a simple image or recognizing and processing it as a character.
[0035]
The displays other than the reading field selection reception display 32 and the image field selection reception display 33 will be described below in the description of the information setting of the form definition.
[0036]
The procedure for setting the form definition information will be described with reference to the flowchart of FIG.
[0037]
First, as described above, a form image is captured (step 81), and the image is displayed by displaying the form definition information setting screen (step 82).
[0038]
Next, a field is designated by a rectangle. This designation can be made, for example, by designating / dragging from the start point to the end point of the field on the displayed image. Thereby, in the start point display 34 and the end point display 35 in FIG. 2, actual numerical values are displayed, for example, in units of mm or dots according to the designated / dragged positions. Alternatively, for example, numerical values may be directly input to the start point display 34 and the end point display 35. In this case, a rectangle is displayed on the color image 20 according to the input numerical value.
[0039]
Next, reading information setting is performed (step 84). Here, the font, character type, and special attributes are set first (step 85). These are performed by inputting to the font display 36, the character type display 37, and the special attribute display 38 in FIG. Here, the font is, for example, a handwritten character or a printed character, the character type is, for example, a number, an alphabetic character, a kana character, a kanji, or the like, and a special attribute is, for example, further dividing the field. Whether binarization or character recognition is performed for each divided area.
[0040]
The reading information setting (step 84) further includes a procedure for determining a binarized axis (step 86). The determination of the binarization axis means that an axis that passes through the origin (black) in the color space and a coordinate point indicating a certain color is uniquely determined. When this axis is determined, the values of the axis components can be calculated for points indicating all colors in the color space. That is, the value is a value on the axis indicated by a point projected from a certain point to the axis in the color space. Further, by determining the binarization axis, a binarization threshold on the axis is set. These pieces of information are stored in the FC table 6 described with reference to FIG. 1 (step 87).
[0041]
Such determination of the binarization axis and setting of the threshold value, for example, appropriately captures the dropout color (a color that should be ignored in binarization and recognition), ignores the dropout color, and makes each color highly reliable. This is important for binarization or recognition. If the binarization axis is determined so that it passes through the coordinates of the dropout color in the color space, the presence or absence of the dropout color can be detected with the highest sensitivity, so it becomes easier to distinguish from the colors other than the dropout color. is there. In the present embodiment , the determination of the binarization axis and the setting of the threshold value can be performed for each field (from step 88 to step 83). As a result, it is possible to set the reading information most appropriate for the processing of the field.
[0042]
Here, an example of a specific procedure for determining the binarization axis and setting the threshold will be described with reference to FIG. This figure is a flowchart showing a specific procedure for determining a binarization axis and setting a threshold value.
[0043]
As shown in the figure, in order to determine a binarization axis, first, representative axes (for example, R axis, G axis, and B axis which are three primary colors are prepared) are selected. Is selected (step 91). When a representative axis is selected (step 95), the binarization threshold on that axis is presented as a default, and the user can further change this value (step 96).
[0044]
Such selection of the binarization axis performs binarization and recognition from the viewpoint of the presence or absence of typical color components. For example, when the dropout color is clearly reddish or greenish, it can be used simply. The binarization threshold is presented as a default for the convenience of the user, but the user can change the threshold to a more appropriate threshold. This is because the test binarization execution (step 97) flows from step 96 to the Y side, and the binarization result is confirmed by the procedures of step 98, step 99, step 91, step 95, and step 96. (Step 98) can be done.
[0045]
These selections, presentation / changes are reflected in the selection display of the R axis, the G axis, and the B axis in the binarization axis selection display 39 in FIG.
[0046]
Further, as described above, when execution is selected from step 96 to selection step 97 of test binarization execution (this is done by clicking the test binarization display 44 in the field information definition window 30 in FIG. 2). In step 98, the selected field is subjected to test binarization using the selected binarization axis and binarization threshold (step 98). This result is displayed on the screen of the PC 1 described in FIG. For example, in the case of the field 21 of the color image 20 in FIG. 2, the result of the test binarization is as shown in FIG. 4, for example. If this result is acceptable, the binarization axis determination and threshold setting routine ends, and if the result is not satisfactory, the process is repeated from step 91.
[0047]
On the other hand, when the representative axis is not used in determining the binarized axis, the binarized axis setting screen is displayed (step 92), and the binarized axis and the threshold on the axis are set. Is done.
[0048]
That is, on the binarization axis setting screen, as described above, for example, coordinates in the color space of the dropout color are designated and input. Here, the coordinates are coordinates on a color space composed of hue, saturation, and lightness (step 93), which is easy to understand for humans.
[0049]
When the coordinates in the color space are designated / input, the binarization axis is uniquely determined. This presents a binarization threshold on that axis, and the user can also change that value (step 94). The binarization threshold is first presented in order to provide convenience to the user by first presenting the default value. As a specific method of presenting the default value, a table indicating the default threshold value on the specified binarized axis is prepared, and this table can be referred to. The change is made from step 94 through step 97 to the Y side, while confirming the binarization result (step 98) according to the procedure of step 98, step 99, step 91, step 92, step 93, and step 94. be able to.
[0050]
These are reflected in the display of the “other setting” display, the saturation display 40, the hue display 41, the lightness display 42, and the binarization threshold display 43 in the binary axis selection display 39 in FIG. .
[0051]
Further, the procedures after the test binarization execution selection step 97 have been described in the case where the representative axis is selected, and thus the description thereof will be omitted.
[0052]
Next, an example of the binarization axis setting screen (step 92) that appeared in the description of FIG. 6 will be described with reference to FIG. FIG. 3 is a diagram illustrating an example of a binarized axis setting screen.
[0053]
As shown in FIG. 3, the binarization axis setting screen 51 includes a color plane 52 having the hue 53 and the saturation 54 as the y axis and the x axis, respectively, and a display for indicating the brightness 58. That is, for example, the coordinates in the color space of the dropout color for determining the binarization axis are set according to the position of the cross pointer 55 on the color plane 52 for the hue 53 and saturation 54, and the position pointer 57 for the brightness 58. It can be indicated by the height of each.
[0054]
In this case, if the user sets these instructions without any clues, the burden is high and it cannot be said to be practical. For the improvement, for example, a case where the field 21 of the color image 20 of the form in the form definition information setting screen 10 shown in FIG. 2 is currently set will be described as an example.
[0055]
Now, when the handwritten description is the object of binarization and recognition, and the chromatic character frame is not the object of binarization and recognition, the cursor is moved to any part of the character frame in the field 21 specify. As a result, the hue, saturation, and lightness of the portion designated by the cursor are automatically reflected in the positions of the cross pointer 55 and the lightness instruction pointer 57 on the color plane 52 of the binarization axis setting screen 51. Thereby, hue, saturation, and brightness can be presented as defaults.
[0056]
As another method, an average of hue, saturation, and lightness is obtained for a chromatic color portion (pixel) that is not black in the field 21, and these values are cross-pointed on the color plane 52 of the binarization axis setting screen 51. 55. It can also be reflected in the position of the lightness instruction pointer 57. In this case, there is an advantage that the default value is not influenced by some pixels.
[0057]
Furthermore, on the binarization axis setting screen 51, a plurality of colors may be designated on the color plane 52 of hue 53 and saturation 54. For example, when two colors are selected / designated, a binary image based on the first designation can be obtained for the field, and a binarized image based on the second designation can also be obtained. Therefore, it is possible to drop out two colors or obtain a binarized image by calculating the logical sum or logical product of the two binarized images. Similarly, for three or more selections / designations, the same effect can be obtained by increasing the color.
[0058]
Further, an area on the color plane 52 may be specified for specifying the hue 53 and the saturation 54. The designation in this case is performed like a designation frame 56, for example. The effect in this case is essentially the same as the case where the plurality of colors can be specified. That is, when designated as an area, it can be considered that each color included in the area is designated, and a binarized image can be obtained for each color designation. Accordingly, by calculating the logical sum or logical product of these binarized images, it is possible to drop out the color of the designated area or obtain a binarized image.
[0059]
In this case, obtaining a binarized image for specifying all the colors (colors corresponding to each digital value) included in the specified area significantly increases the processing burden, so that a predetermined representative included in the area is determined. Processing may be performed on the assumption that a color is designated for a point.
[0060]
As described above, when the form definition information setting is completed, an actual form reading process is performed based on this information.
[0061]
【The invention's effect】
As described above in detail, according to the character reading device, the character reading method, and the computer-readable storage medium storing the character reading program according to the present invention, a point on the color space is designated, and the origin of the color space is designated. Since the straight line passing through the point is taken as the binarization axis, the projection value to the binarization axis of each pixel is obtained, and the obtained projection value is binarized by a predetermined threshold value. Is binarized according to a specific color standard. Thereby, it is possible to perform processing such as character recognition by obtaining a binarized image optimal for each form and each field while ensuring the user's degree of freedom.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing the configuration of a character reader according to an embodiment of the present invention (FIG. 1A) and the configuration of an OCR (Optical Character Reader) system provided in the character reader. .
FIG. 2 is a diagram showing an example of an information setting screen for form definition.
FIG. 3 is a diagram showing an example of a binarized axis setting screen.
FIG. 4 is a diagram showing an example of a test binarization result display.
FIG. 5 is a flowchart showing a procedure for setting information in a form definition.
FIG. 6 is a flowchart showing a specific procedure for determining a binarization axis and setting a threshold value.
[Explanation of symbols]
1 Scanner 2 PC
3 OCR System 4 OCR Reading Program 5 Form Definition Program 10 Form Definition Information Setting Screen 20 Color Image 20
21, 22, 23, 24 Field 30 Field information definition window 31 Field number 32 Reading field selection reception display 33 Image field selection reception display 34 Field start point (start coordinate) display 35 Field end point (end coordinate) display 36 Font display 37 Character type display 38 Special attribute display 39 Binary axis selection display 40 Saturation display of selected binary axis 41 Hue display of selected binary axis 42 Brightness of selected binary axis Display 43 Binary threshold display 44 Test binarization acceptance display 51 Binary axis setting screen 52 Color plane 53 Hue 54 Saturation 55 Cross pointer 56 Designation frame 57 Position pointer 58 Lightness

Claims (18)

Means for dividing a form into pixels and capturing it in color by photoelectric conversion;
A means of specifying a point on the color space;
Using a straight line passing through the origin of the color space and the designated point as a binarization axis, a projection value of the point in the color space corresponding to each captured pixel to the binarization axis is obtained. Means,
Means for binarizing the calculated projection value by a predetermined threshold;
And a means for processing the image obtained by the binarization. The point on the color space is designated for each field constituting the form,
The character reading device according to claim 1, wherein the calculation of the projection value is performed for pixels belonging to the fields among the captured pixels. And means for displaying a color image of a form that has been taken in the previous Symbol color,
Further comprising a means for specifying a particular position or region in the displayed on the color image,
The designation of a point on the color space, the character reader according to claim 1, wherein to be made with reference to the color information of pixels of the designated specific position or regions have. The designation of a point on the color space, performed for a plurality of points,
The projection value is obtained for each of a plurality of binarization axes obtained by designating the plurality of points,
4. The character according to claim 1, wherein the means for processing the image includes a process of calculating a logical sum or a logical product of the plurality of images obtained by the binarization. 5. Reader. And means for displaying the image obtained by the previous Symbol binary,
The character reading apparatus according to claim 1, further comprising : a unit that updates and designates a point on the designated color space. The character reading apparatus according to claim 1, wherein the color space is a space having hue, saturation, and lightness as components . Dividing the form into pixels and taking it in color by photoelectric conversion;
Specifying a point on the color space;
Using a straight line passing through the origin of the color space and the designated point as a binarization axis, a projection value of the point in the color space corresponding to each captured pixel to the binarization axis is obtained. Steps,
Binarizing the calculated projection value with a predetermined threshold;
And a step of processing an image obtained by the binarization. The point on the color space is designated for each field constituting the form,
The character reading method according to claim 7, wherein the calculation of the projection value is performed for pixels belonging to the fields among the captured pixels. And the step of displaying a color image of a form that has been taken in the previous Symbol color,
Further comprising the step of accepting designation of a specific location or region in the displayed on the color image,
The designation of a point on the color space, character reading method according to claim 7 or 8, wherein to be made with reference to the color information of pixels of the designated specific position or regions have. The designation of a point on the color space, performed for a plurality of points,
The projection value is obtained for each of a plurality of binarization axes obtained by designating the plurality of points,
10. The character according to claim 7, wherein the step of processing the image includes a process of calculating a logical sum or a logical product of the plurality of images obtained by the binarization. Reader. And displaying an image obtained by the pre-Symbol binarization,
The character reading method according to claim 7, further comprising a step of updating and designating a point on the designated color space. The character reading method according to claim 7, wherein the color space is a space having hue, saturation, and lightness as components . Dividing the form into pixels and taking it in color by photoelectric conversion;
Specifying a point on the color space;
Using a straight line passing through the origin of the color space and the designated point as a binarization axis, a projection value of the point in the color space corresponding to each captured pixel to the binarization axis is obtained. Steps,
Binarizing the calculated projection value with a predetermined threshold;
A computer-readable information storage medium recording a character reading program that causes a computer to execute the step of processing an image obtained by binarization. The point on the color space is designated for each field constituting the form,
14. The computer-readable information storage medium having a character reading program recorded thereon according to claim 13, wherein the calculation of the projection value is performed for pixels belonging to each field among the captured pixels. Before Symbol character reading program further comprises a step of displaying a color image of a form taken by the collar, and a step of accepting designation of a specific location or area in the displayed on the color image,
The designation of a point on the color space were recorded claim 13 or 14, wherein the character reading program characterized in that it is made with reference to the color information of pixels of the designated specific position or regions having A computer-readable information storage medium. The designation of a point on the color space, performed for a plurality of points,
The projection value is obtained for each of a plurality of binarization axes obtained by designating the plurality of points,
The character processing method according to any one of claims 13 to 15, wherein the step of processing the image includes a process of obtaining a logical sum or a logical product of the plurality of images obtained by the binarization. A computer-readable information storage medium having a reading program recorded thereon. Before Symbol character reading program, and displaying an image obtained by the binarization to 13 claims, characterized in that it further comprises the step of updating specify a point on the designated color space 16 A computer-readable information storage medium recording the character reading program according to any one of the above. 18. The computer-readable information storage medium storing a character reading program according to claim 13, wherein the color space is a space having hue, saturation, and lightness as components .

Images