JP2001195543A - Device and method for processing document and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make accurately extractable only contents, which are described on a document later, from the image of the document even without designating the color of the document form of the document or color of the contents of characters or the like described later, beforehand. SOLUTION: When the color image of a document is read by a color image input means 2, an image classifying means classifies the color image into an image in the color comprising the document form previously described on the document and an image in the color comprising a character described on the document later, and these two images are extracted from the color image reading the document. On the basis of size comparison between the respective components of two images extracted by the image classifying means 3, one of two images is identified as the image of a character described later by an image identifying means 4. Concretely, the largest link component of pixels is extracted from two images, areas are found, the areas are compared, an image containing a larger circumscribed rectangle is defined as the image of the document form of the document and the other image is defined as the image of characters described later.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a form processing apparatus and method, and a storage medium.

[0002]

2. Description of the Related Art Japanese Patent Application Laid-Open No. Hei 10-187882 discloses a technique for recognizing only a character entered later on a form in which a document form including frame lines and characters is described in advance. I have.

[0003] In general, even in a form that does not use a dropout color, in consideration of the easiness of discrimination between a document form described in advance on the form and characters to be entered later on the form, both of them are considered. Often have different colors. From the image read such a form, without previously specifying the color of the document form described in the form, in order to extract the characters to be entered later on the form, the JP-A-10-187882 Performs the following processing.

[0004] That is, it is defined that the characters to be entered later on the form are black, and in the image of the form read by the color scanner, the color of the document form of the form is lighter than the color of the character entered later. Utilizing the fact that is higher, threshold processing is performed to separate and identify only characters that are entered later.

[0005]

SUMMARY OF THE INVENTION However, Japanese Patent Application Laid-Open
The technique disclosed in Japanese Patent Application Laid-Open No. 187882 has a problem that, when the character color is not black, it is not possible to extract only a character entered later on the form. Some forms use several sheets of carbon or non-carbon copy paper, but the color of the copied paper is not always black. In that case, blue is often used, but since the saturation and brightness of the color and the like differ depending on the form, it is often not possible to specify the color in advance.

SUMMARY OF THE INVENTION An object of the present invention is to make it possible to accurately analyze only the contents of a form from a form image later without specifying the color of a document form of the form or the color of contents such as characters entered later. It is to be able to extract well.

[0007] Another object of the present invention is to be able to accurately extract only the contents that are later entered in a form from a form image, for example, when a frame is described in the form as a document form. is there.

[0008] Another object of the present invention is to be able to accurately extract only the contents that have been entered in a form later from an image of the form when a ruled line is described in the form as a document form. .

Another object of the present invention is to make it possible to recognize only the characters entered on the form which are entered later.

[0010]

According to a first aspect of the present invention, there is provided an image processing apparatus comprising: a color image forming a document form described in a form in advance; and a color image forming a content entered later in the form. Image classifying means for extracting the two images from the image obtained by reading the form, and based on a magnitude comparison between respective components of the two images extracted by the image classifying means. An image identification means for identifying one of the two images as an image of the content entered later.

Therefore, it is possible to accurately extract only the contents of a form afterward from an image of the form without specifying the color of the document form of the form or the color of the contents such as characters entered later. Can be.

According to a second aspect of the present invention, in the form processing apparatus according to the first aspect, the image identification unit determines a size of a connected portion of pixels included in each of the two images as a magnitude comparison of the specific portion. This is to make a comparison.

Therefore, when a frame is described in a form as a document form, for example, it is possible to accurately extract only the contents later entered in the form from the image of the form.

According to a third aspect of the present invention, in the form processing apparatus according to the first aspect, the image identification unit determines a length of a straight line included in each of the two images as a magnitude comparison between the constituent elements. This is to make a comparison.

Therefore, not only when a frame is described on a form as a document form, but also when a ruled line is described in place of the frame, the contents entered later on the form from the form image Can be accurately extracted.

According to a fourth aspect of the present invention, there is provided the form processing apparatus according to any one of the first to third aspects, wherein the form identification unit is configured to register the image identified by the image identification unit as the content described later. The image processing apparatus includes character extracting means for extracting an image forming a character, and character recognizing means for performing character recognition of the image forming the extracted character.

[0017] Therefore, it is possible to recognize only the later entered characters among the characters described in the form.

According to a fifth aspect of the present invention, a color image constituting a document form described in advance on a form and a color image constituting contents which are later entered in the form are classified into: An image classification step of extracting the two images from the image obtained by reading the form, and based on a magnitude comparison between respective constituent elements of the two images extracted by the image classification means, An image identification step of identifying one as an image of the content entered later.

Therefore, it is possible to accurately extract only the contents of a form from a form image later without specifying the color of the form of the form or the color of contents such as characters entered later. Can be.

According to a sixth aspect of the present invention, in the form reading method according to the fifth aspect, the image identifying step includes a step of comparing a connected portion of pixels included in each of the two images as a magnitude comparison between the constituent elements. This is to compare the sizes.

Therefore, in the case where a frame line is described in a form as a document form, for example, it is possible to accurately extract only the contents later entered in the form from the image of the form.

According to a seventh aspect of the present invention, in the form reading method according to the fifth aspect, the image identification step includes determining a length of a straight line included in each of the two images as a magnitude comparison between the constituent elements. This is to make a comparison.

Therefore, not only when a frame is described in a form as a document form, but also when a ruled line is described in place of the frame, etc., the contents entered later on the form from the form image Can be accurately extracted.

According to an eighth aspect of the present invention, there is provided the form reading method according to any one of the first to third aspects, wherein an image identified as the content later entered in the image identifying step is used. The method includes a character extracting step of extracting an image forming a character, and a character recognizing step of performing character recognition of an image forming the extracted character.

Therefore, it is possible to recognize only the characters entered later among the characters described in the form.

According to a ninth aspect of the present invention, a color image constituting a document form described in advance on a form and a color image constituting contents to be later entered in the form are classified into: An image classification step of extracting the two images from the image obtained by reading the form, and based on a magnitude comparison between respective constituent elements of the two images extracted by the image classification means, And a computer-readable storage medium storing a program for causing the computer to execute an image identification step of identifying one of the contents as an image of the content entered later.

Therefore, it is possible to accurately extract only the contents of a form later written from a form image without specifying the color of a document form of the form or the color of contents such as characters entered later. Can be.

According to a tenth aspect of the present invention, in the storage medium according to the ninth aspect, in the image identifying step, a size of a connected portion of a pixel included in each of the two images is compared as a magnitude comparison between the constituent elements. This is to make a comparison.

Therefore, when a frame is described in a document as a document form, it is possible to accurately extract only the content that is later entered in the document from the image of the document.

According to an eleventh aspect of the present invention, in the storage medium according to the ninth aspect, the image identifying step compares the length of a straight line included in each of the two images as a magnitude comparison between the constituent elements. Is what you do.

Therefore, not only when a frame line is described on a form as a document form, but also when a ruled line is described in place of the frame line, the content entered later on the form from the image of the form. Can be accurately extracted.

According to the twelfth aspect of the present invention, there is provided the ninth to the first aspects.
1. The storage medium according to claim 1, wherein the program further includes: a character extracting step of extracting an image forming a character from an image identified by the image identifying step as the content entered later. And a character recognition step of performing character recognition of an image constituting the extracted character.

Therefore, it is possible to recognize only the characters entered later among the characters described in the form.

[0034]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [First Embodiment of the Invention] One embodiment of the present invention will be described as a first embodiment of the present invention.

FIG. 1 is a functional block diagram of a form processing apparatus 1 according to the first embodiment of the present invention. As shown in FIG. 1, the form processing apparatus 1 includes a color image input unit 2 for inputting a color image of a form, and a frame line, a character, An image classifying means 3 for classifying into a document form composed of figures and the like (a template prepared in advance for document creation and the like) and a color of a character entered later and extracting each of them; An image identification means 4 for determining which of the classified image is a document form pre-filled in a form and a character entered later, and a character for each form. Referring to a character area information dictionary 5 storing information indicating an area to be filled, a character extracting means 6 for extracting a character described in a form, a character image and its character code are associated with each other. Remember for each character That with reference to the character recognition dictionary 7, and recognizes the character recognition means 8 the extracted character by character extraction means 6, constituted by an output unit 9 for outputting a result of the character recognition.

Next, the processing performed by the form processing apparatus 1 will be described. FIG. 2 is a flowchart illustrating a flow of a process performed by the form processing apparatus 1. In the following, a description will be given using the form 11 as shown in FIG. 3 as an example. As shown in FIG. 2, a color image of a form is input by the color image input means 2 (step S1). The image data of this color image is
Each pixel has a multi-valued pixel value for each of the R (red), G (green), and B (blue) color components. Based on this pixel value, the color of the document form of the form is first determined. Pixels, pixels having the character color written later, and pixels having the background color are classified by the image classification means 3 (step S2). This can be realized by classifying pixel values into a plurality of sets by a generally well-known statistical method, for example, a discriminant analysis method or the like. Then, the pixels are classified into the pixels having the color of the document form of the form and the pixels having the color of the character entered later, and the two classified pixels are respectively mapped on two different images. The read image of the form 11 shown in FIG. 3 is composed of an image 12 composed of pixels having the color of the document form of the form 11 shown in FIG. 4 and a pixel having the color of a character written in the form 11 later shown in FIG. And an image 13 composed of The step S2 implements the image classification step.

However, at this point, only the images can be classified into two, and the image of either color is the image 12 of the document form of the form 11, and the image of the character entered later on the form 11 Since it is unclear whether the image is 13 or not, the images 12 and 13 are identified by the image identifying means 4 (step S3). As a means, two images 1
By performing the comparison based on the magnitude comparison between the constituent elements 2 and 13, the form 1 can be used without specifying the color of the document form of the form 11 or the color of the character entered later.
The image of the first document form and the image of the character entered later on the form 11 can be identified with high accuracy. The image identification step is realized by step S3.

Specifically, the processing is performed as follows. In general,
The document form of the form 11 includes many figures that are graphically larger than characters to be entered later. Utilizing this, the largest one of the connected components of pixels is extracted from each of the images 12 and 13 to obtain the area thereof, and the areas are compared.
Image 12 of the document form, and the image 13 of the character entered later. That is, as shown in FIG. 6, a circumscribed rectangle 14 of the largest connected component of the image 12 is extracted, and as shown in FIG. 7, a circumscribed rectangle 15 of the largest connected component of the image 13 is extracted. When 14 and 15 are compared, in this example, since the circumscribed rectangle 14 is larger, it is understood that the image 12 including the circumscribed rectangle 14 is an image composed of pixels having the color of the document form of the form 11.

Then, referring to the character area information dictionary 5,
An area where a character is to be written in the form 11 is specified, and a character image is extracted from the area in the character image 13 written later in the form 11 by the character extracting means 6 (step S4), and the extracted character image is extracted. The character recognition is executed by acquiring the character code corresponding to the character image by the character recognition unit 8 with reference to the character recognition dictionary 7 (step S5), and the result is output by the output unit 9 (step S5). S6). The character extracting step is realized by step S4, and the character recognizing step is realized by step S5.

FIG. 8 is a block diagram showing the electrical connection of the form processing apparatus 1. The form processing apparatus 1 includes a CPU 21 that performs various calculations and performs various controls, a ROM 22 that stores a BIOS and the like, and a RAM 23 that stores various data in a rewritable manner and is a work area of the CPU 21. Connected by The bus 24 further includes, via various interfaces 25, an input device 26 such as a keyboard and a mouse, an output device 27 such as a display and a printer, a hard disk 28 as an external storage device, and a CD-ROM 29 as a storage medium. A CD-ROM drive 30 for reading an image, an image scanner 31 such as a flatbed type, and a communication control device 32 for communicating with the Internet or the like are connected.

Various programs are stored in the CD-ROM 29. The programs are read by the CD-ROM drive 30 and installed on the hard disk 30, whereby the above-mentioned various processes can be executed. That is, the form is read by the image scanner 31 to realize the color image input unit 2, and the internal processing of the form processing apparatus 1 is realized by the image classification unit 3, the image identification unit 4, the character extraction unit 6, and the character recognition unit 8. ,
The output unit 9 is realized by outputting the results of these processes to the output device 27. An area for storing the character area information dictionary 5 and the character recognition dictionary 7 is also secured on the hard disk 28.

The storage medium is not limited to the CD-ROM 29, and various types of recording media such as CD-R, CD-RAM, and other types of CDs, DVDs, MOs, and FDs can be used. Further, the program may be downloaded from the Internet or the like via the communication control device 32 and installed on the hard disk 30. In this case, the program is used for a transmission-side device connected to the Internet or the like. A storage device that stores a program so that it can be transmitted is also a storage medium of the present invention. Note that the program may be operated on a predetermined OS.

According to the form processing apparatus 1 described above, when an image of a form in which the same type of form is different in color is input, such as the second and third sheets of a plurality of spelled copy sheets. However, it is possible to correctly extract and recognize characters without having to specify in advance the colors of the document form of the form and the colors of the characters that are later entered in the form. In addition, since the color of the character does not need to be black, it is possible to correctly extract and recognize a character having a color other than black, which is often used for a copy sheet composed of a plurality of sheets.

[Second Embodiment of the Invention] Another embodiment of the present invention will be described as a second embodiment of the present invention.

In the second embodiment of the present invention, items common to the first embodiment of the present invention are denoted by the same reference numerals, and detailed description is omitted. The difference between the second embodiment of the present invention and the first embodiment of the present invention lies in the content of the processing of step S3 executed by the image identifying means 4.

That is, when the document form of the form is not a rectangular frame like the form 11 shown in FIG. 3 but is composed of ruled lines, the area of the circumscribed rectangle of the connected component becomes small. In some cases, it may not be possible to correctly identify the color of the document form and characters that are later entered in the form. For example, in the case of the form image as shown in FIG. 5, the area of the largest circumscribed rectangle in the connected component of the document form of the form is not large compared to the area of the character image, and may be erroneously identified.

In order to prevent this, instead of judging the size of the area of the circumscribed rectangle of the connected component, the maximum length of the straight line included in each image is calculated, and these are compared to determine the maximum length of the straight line. May be an image of the document form of the form, and the other may be an image of characters entered later.

That is, in general, ruled lines included in a document form of a form are often straight lines having a certain length in the vertical or horizontal direction. The frequency becomes extremely large at a certain position. The frequency distribution of the image of the character entered later does not have such a value. By distinguishing both images using this feature, it is possible to correctly extract and recognize characters without erroneous identification of an image of a document form of a form without a frame.

For example, in the case of the form 41 shown in FIG. 9, the image of the document form of the form 41 is the image 4 shown in FIG.
2, and the image of the character entered later in the form 41 is
It is an image 43 shown in FIG. In this example, the frequency distribution of pixels having a value in the vertical or horizontal direction is obtained for each of the images 42 and 43, and the larger one of the respective frequencies is defined as the image 42 of the document document form. That is, FIG.
In the example shown in FIG. 11 to FIG.
The frequency distribution of pixels having values in the horizontal direction of the image of FIG. 12A is obtained (FIG. 12B), and the image 43 of FIG. The frequency distribution of pixels having values in the horizontal direction is obtained (FIG. 13B), and when the maximum values of the frequencies are compared, the maximum value of the image 42 is larger than that of the image 42 (FIG. 12B). 13 (b)), image 42
Can be determined to be the image of the document form of the form 41.

[0050]

According to the first aspect of the present invention, it is possible to fill out a form from an image of a form later without specifying a color of a document form of the form or a color of contents such as characters entered later. Only the contents that have been extracted can be accurately extracted.

According to a second aspect of the present invention, in the form processing apparatus according to the first aspect, when a frame line is described in the form as a document form, the form is later filled out from the form image. Only the contents can be accurately extracted.

According to a third aspect of the present invention, in the form processing apparatus according to the first aspect, not only a case where a frame is described as a document form but also a ruled line is described instead of the frame line. In such a case, it is possible to accurately extract only the contents later entered in the form from the form image.

According to a fourth aspect of the present invention, in the form processing apparatus according to any one of the first to third aspects, it is possible to recognize only a character entered later in the form. it can.

According to the fifth aspect of the present invention, it is possible to provide a content that is later entered into a form from an image of the form without specifying the color of the document form of the form or the color of the content such as characters entered later. Can be accurately extracted.

According to a sixth aspect of the present invention, in the form reading method according to the fifth aspect, when a frame line is described in the form as a document form, the form is later entered from the form image into the form. Only the contents can be accurately extracted.

According to a seventh aspect of the present invention, in the form reading method according to the fifth aspect, not only a case where a frame is described in a form as a document form but also a ruled line is described in place of the frame. In such a case, it is possible to accurately extract only the contents later entered in the form from the form image.

According to an eighth aspect of the present invention, in the form reading method according to any one of the first to third aspects, it is possible to recognize only a character entered later in the form. it can.

According to a ninth aspect of the present invention, the contents of a form that is later entered from a form image without specifying the color of the document form of the form or the color of the contents such as characters entered later. Can be accurately extracted.

According to a tenth aspect of the present invention, in the storage medium according to the ninth aspect, when a frame line is described in a form as a document form, the content that is later entered in the form from the form image Can be accurately extracted.

According to an eleventh aspect of the present invention, in the storage medium according to the ninth aspect, not only a case where a frame is described in a form as a document form, but also a ruled line is described in place of the frame line. Even in such cases, it is possible to accurately extract only the contents later entered in the form from the form image.

The twelfth aspect of the present invention relates to the ninth to the first aspects.
In the storage medium according to any one of (1) and (3), only the characters written later in the form can be recognized.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a form processing apparatus according to a first embodiment of the present invention.

FIG. 2 is a flowchart illustrating processing of the form processing device.

FIG. 3 is a plan view showing an example of a form read by the form processing apparatus.

FIG. 4 is a plan view showing an image of a document form which is extracted from the image of the form and which is previously entered in the form.

FIG. 5 is a plan view showing an image of a character written on a form after being extracted from the form image.

FIG. 6 is a plan view showing a circumscribed rectangle in which a connected component of a pixel is maximum in the image of the document form.

FIG. 7 is a plan view showing a circumscribed rectangle in which a connected component of a pixel is maximum in the image of the character.

FIG. 8 is a block diagram showing an electrical connection of the form processing device.

FIG. 9 is a plan view showing an example of a form read by a form processing apparatus according to Embodiment 2 of the present invention;

FIG. 10 is a plan view showing an image of a document form which is extracted from the image of the form and which is previously written in the form.

FIG. 11 is a plan view showing an image of a character entered in a form after being extracted from the form image.

FIG. 12 is a plan view showing a frequency distribution of pixels in the horizontal direction in the image of the document form.

FIG. 13 is a plan view showing a frequency distribution of horizontal pixels in the character image.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Form processing device 3 Image classification means 4 Image identification means 6 Character extraction means 7 Character recognition means

Claims (12)

[Claims] 1. A color image forming a document form previously described in a form and a color image forming a content which is later entered in the form, and the two images are classified into the form. An image classification unit that extracts from the read image, and one of the two images is later filled in based on a magnitude comparison between respective components of the two images extracted by the image classification unit. A form processing device comprising: an image identification unit that identifies the image as the content. 2. The form processing apparatus according to claim 1, wherein the image identification unit compares the size of a connected portion of pixels included in each of the two images as a magnitude comparison between the constituent elements. . 3. The form processing apparatus according to claim 1, wherein the image identification unit compares the lengths of straight lines included in the two images as a magnitude comparison between the constituent elements. 4. A character extracting means for extracting an image constituting a character from an image which has been subsequently identified by said image identifying means as said contents, and character recognition of an image constituting said extracted character. The form processing apparatus according to any one of claims 1 to 3, further comprising: a character recognition unit that performs the following. 5. A color image constituting a document form previously described on a form and a color image constituting contents which are later entered on the form are classified, and the two images are classified into the form. An image classification step of extracting from the read image, and one of the two images is filled in later based on a magnitude comparison between respective components of the two images extracted by the image classification means. An image identification step of identifying the image as the image of the content. 6. The form reading method according to claim 5, wherein the image identification step compares the size of a connected portion of pixels included in each of the two images as a magnitude comparison between the constituent elements. . 7. The form reading method according to claim 5, wherein the image identification step compares the lengths of straight lines included in the two images as a magnitude comparison between the constituent elements. 8. A character extracting step of extracting an image constituting a character from an image identified as having the content entered later in the image identifying step, and character recognition of the image constituting the extracted character The form reading method according to any one of claims 1 to 3, further comprising: 9. A color image forming a document form previously described in a form and a color image forming a content which is later entered in the form, and the two images are classified into the form. An image classification step of extracting from the read image, and one of the two images is filled in later based on a magnitude comparison between respective components of the two images extracted by the image classification means. A computer-readable storage medium storing a program for causing a computer to execute an image identification step of identifying the image as the content. 10. The image identification step includes comparing the sizes of connected portions of pixels included in the two images as a magnitude comparison between the constituent elements.
A storage medium according to claim 1. 11. The storage medium according to claim 9, wherein the image identification step compares the lengths of straight lines included in the two images as a magnitude comparison between the constituent elements. 12. The program further comprises: a character extracting step of extracting an image constituting a character from an image identified in the image identifying step as the content of the content entered later; The storage medium according to any one of claims 9 to 11, which causes a computer to execute a character recognition step of performing character recognition of an image to be formed.