JPH11120260A - Method and device for processing document image, and storage medium stored with document image processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the method and device for document image processing which enable accurate image cutting whatever kind of segment is entered into an image entry field by making accurate tilt corrections of a document and eliminating an error in format discrimination even if a straight line similar to a ruled line used for the tilt correction and format discrimination is present in various drawings such as a map and a layout figure entered into the image entry field, and the storage medium where the document image processing program is stored. SOLUTION: When a document into which characters, ruled lines showing the image entry field, and a mark indicating the position of the image entry field are entered in a color that a facsimile or scanner can sense is inputted (S1), a document image is stored in a memory (S21), lateral ruled lines of the document image and the mark on the lateral ruled line or on the prolongation of the lateral ruled line are detected (S3), and a horizontal ruled line and a mark on the horizontal ruled line or its prolongation are detected in a document image after tilt correction is made (S4) by using the lateral ruled line where the mark is present or a lateral ruled line outside a vertical scanning range sandwiched between the prolongations of the lateral ruled lines to discriminate, the document by using a horizontal ruled line outside the vertical scanning range sandwiched between marks (S6).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a form image processing method and apparatus, and a storage medium storing a form image processing program, and in particular, what kind of image is described in an image entry box indicated by a ruled line. The present invention also relates to a form image processing method and apparatus capable of performing highly accurate tilt correction, form identification, and image area cutout processing, and a storage medium storing a form image processing program.

[0002]

2. Description of the Related Art Conventionally, in a form reading system using ruled lines, inclination correction is performed using long ruled lines (Japanese Patent Application No. 7-24 / 1990).
7644 "Image data input processing method and its apparatus"), and a plurality of types of form image data are identified by using the relative positions and the number of long ruled lines.
115924 "Method and apparatus for identifying form format"), and detects and cuts out character entry fields and image entry fields.

[0003]

However, in the above-mentioned conventional technique, the image entry field is a small area where a sign or a simple illustration is described. That is, if a map or a sketch drawn on a long line segment is described using a large image entry column, a straight line similar to a ruled line used for inclination correction and format identification may be described in the image entry column. Depending on the length and inclination of the straight line, there is a problem that the inclination of the form cannot be corrected correctly, or an error occurs in the format identification, and the image entry column cannot be taken out correctly.

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and includes various types of drawings, such as maps and floor plans, described in image entry columns, in which lines similar to ruled lines used for inclination correction and format identification exist. In addition, a form image processing method and apparatus and form that perform accurate inclination correction of a form, eliminate format identification errors, and enable accurate image cutout regardless of what kind of line segment is described in an image entry column. It is an object to provide a storage medium storing an image processing program.

[0005]

FIG. 1 is a diagram for explaining the principle of the present invention. The present invention relates to a form image processing method for recognizing a character or image input by a facsimile or scanner on a form on which a character or image is described, and showing a ruled line indicating a character and image entry field of the form and a position of the image entry field. When a form in which the mark is described in a color that can be detected by a facsimile or a scanner is input (step 1), the form image of the input form is stored in a memory (step 2), and a horizontal ruled line of the form image is detected and the horizontal line is detected. A mark on the ruled line or on an extension of the horizontal ruled line is detected (step 3), and inclination is performed using a horizontal ruled line on which the mark exists or a horizontal ruled line other than the sub-scanning area sandwiched by the extended lines of the horizontal ruled line. After the correction (step 4), the horizontal ruled line is detected and the mark on the horizontal ruled line or on the extension of the horizontal ruled line is detected from the form image on which the tilt correction has been performed. It detected (Step 5), and identifies the form by using the horizontal ruled lines other than the sub-scanning range between the mark (Step 6).

Further, the present invention detects a vertical ruled line from a form image on which tilt correction has been performed, detects a mark on the vertical ruled line or an extension of the vertical ruled line, and performs main scanning sandwiched between the marks. The form is identified using vertical ruled lines outside the range. Further, according to the present invention, a mark indicating the position of the image entry column can be registered. FIG. 2 is a diagram illustrating the principle of the present invention.

The present invention relates to a form image processing apparatus for recognizing a character and an image cut out by inputting a form on which a character or an image is described with a facsimile or a scanner. Form input means 8 for inputting a form in which a mark indicating a position of a column is described in a color which can be detected by a facsimile or a scanner.
A form image storage means 1 for storing a form image of the form inputted by the form input means 8, a form image obtained from the form image storage means 1, detection of a horizontal ruled line of the form image, A mark that detects a mark above or on an extension of the horizontal ruled line, and performs a tilt correction using a horizontal ruled line other than the sub-scanning range sandwiched between the horizontal ruled line where the mark exists or the extension of the horizontal ruled line. From the correction means 2 and the form image on which the tilt correction has been performed by the tilt correction means 2,
Format identification means 3 for detecting a horizontal ruled line, detecting a mark on the horizontal ruled line or an extension of the horizontal ruled line, and identifying a form using a horizontal ruled line other than the sub-scanning range sandwiched between the marks; Having.

The format identification means 3 detects a vertical ruled line from a form image on which the inclination has been corrected, detects a mark on the vertical ruled line or an extension of the vertical ruled line, and detects the mark between the marks. Means for identifying a form using a vertical ruled line other than the main scanning range is included. Further, the form image processing apparatus of the present invention further has a mark registration unit for registering a mark indicating the position of the image entry column.

The present invention is a storage medium storing a form image processing program for recognizing characters and inputting a form containing characters and images by facsimile or scanner, and showing a character and image entry field of the form. A form input process for inputting a form in which a ruled line and a mark indicating the position of the image entry field are described in a color that can be detected by a facsimile or a scanner, and a form image storage process for storing a form image of the form input by the form input process. Acquiring a form image from a form image storage process, detecting a horizontal ruled line of the form image, detecting a mark on the horizontal ruled line or on an extension of the horizontal ruled line, and detecting a horizontal ruled line where the mark exists or A tilt correction that performs tilt correction using a horizontal ruled line other than the sub-scanning range sandwiched between extensions of the horizontal ruled line. A horizontal ruled line, and a mark on the horizontal ruled line or an extension of the horizontal ruled line is detected from the form image on which the tilt is corrected by the tilt correction process, and a sub-scanning range sandwiched between the marks is detected. And a form identification process for identifying a form using horizontal ruled lines other than the above.

In the format identification process, a vertical ruled line is detected from the form image on which the inclination has been corrected, and a mark on the vertical ruled line or an extension of the vertical ruled line is detected.
The process includes a process of identifying a form using a vertical ruled line other than the main scanning range sandwiched between the marks. The image processing apparatus further includes a mark registration process for registering a mark indicating the position of the image entry column.

As described above, the present invention first stores the received form image data in the memory. Then, a ruled line used for tilt detection is detected from the form image data, and a mark on the ruled line or an extension of the ruled line is detected. Then, inclination correction is performed using ruled lines outside the sub-scanning range sandwiched by the marks. Next, while detecting the horizontal and vertical ruled lines of the form image data whose inclination has been corrected,
Detects marks on horizontal / vertical ruled lines or on extensions of horizontal / vertical ruled lines. Then, identification of a plurality of formats is performed using the position and the number of ruled lines other than the main scanning range and the sub-scanning range sandwiched by the marks. After the format identification, the character entry column and the image entry column are detected based on the information described in the identified format, and characters are cut out and the image is cut out. In this way, even if a line segment similar to the ruled line is described in the form image entry field using the ruled line, accurate inclination correction and format identification can be performed, and the image area can be cut out.

As described above, a form in which a mark is described on a ruled line indicating an image entry column or an extension of the ruled line is used, and this mark is detected, and a horizontal ruled line in a sub-scanning range sandwiched between the marks is removed. , Using a horizontal ruled line, and performing format identification of a form using a ruled line excluding a vertical ruled line existing in a main scanning range and a horizontal ruled line existing in a sub-scanning range sandwiched between marks. In addition, it is possible to describe a large arbitrary drawing or the like without generating a form identification NG or a character / image cutout error.

[0013]

FIG. 3 shows the configuration of a form image processing apparatus according to the present invention. The form image processing apparatus shown in FIG. 1 uses an image memory unit 1 for storing received form image data, an inclination correction unit 2 for correcting the form image inclination using horizontal ruled lines, and horizontal and vertical ruled lines of the form image data. A format identification unit 3 for selecting a specific format from a plurality of registered formats, a mark storage unit 4 for storing a mark indicating the position of an image entry column, and a character entry column using the selected format information. A character / image cutout unit 5 for cutting out character data and image data in an image entry column,
It comprises a character recognition unit 6 for recognizing the cut-out character data, and a result editing unit 7 for editing the character recognition result and the cut-out image data together with an external device.

Next, the operation of the form image processing apparatus will be described. 1. When the form image data is received, it is stored in the image memory unit 1. 2. When the storage of the form image data is completed, the image memory unit 1 instructs the tilt correction unit 2 to start tilt correction. 3. Receiving the instruction to start the inclination correction, the inclination correction unit 2 detects a horizontal ruled line having a length equal to or more than a specified length and an inclination, and checks whether a mark exists on the detected horizontal ruled line.

4. Then, the inclination correction is performed using the inclination of the horizontal ruled line excluding the horizontal ruled line in the sub-scanning range sandwiched between the horizontal ruled lines where the marks are present, among the detected horizontal ruled lines. 5. The form image data after the inclination correction is written to the image memory 1. 6. When the inclination correction is completed, the inclination correction unit 2 instructs the format identification unit 3 to start format identification.

[7] The format identification unit 3 that has received the instruction to start the format identification detects horizontal and vertical ruled lines of a specified length or more from the form image data in the image memory 1. 8. The presence / absence of a mark on the detected horizontal ruled line is checked and compared with the ruled lines of a plurality of registered formats except for the horizontal ruled line in the sub-scanning range sandwiched by the horizontal ruled lines where the mark exists. 9. A format that matches the ruled line of the registration format is selected as the format of the form image data.

10. When a format is selected, format identification unit 3
Instructs the character / image cutout unit 5 to cut out characters and images. 11. The character / image cut-out portion that has received the character / image cut-out instruction is set to a character region by the character
In the case of the image entry column, the image area is transferred to the result editing unit 7.

[12] Character recognition unit 6 that received the character area
Transfers the character code resulting from the character recognition to the result editing unit 7. 13. The result editing unit 7, which has acquired the image area and the character recognition result, edits these together with the external interface and outputs them.

[0019]

Embodiments of the present invention will be described below with reference to the drawings. First, a method of tilt correction by the tilt correction unit 2 will be described. FIG. 4 is a diagram for explaining a method of tilt correction in the tilt correction unit according to one embodiment of the present invention.

In FIG. 1, a vertical arrow indicates a sub-scanning direction, and a horizontal arrow indicates a main scanning direction. a is the form image data in the image memory 1, b is a form effective area, c is a character entry field, and d to h are horizontal line segment detection scanning coordinates of a specified length or more. The marks e to h indicating the four corners of the character entry column c and the image entry column d in the form are those which are stored in the mark storage unit 4 as squares with one side being x and read into the inclination correction unit 2. And

Further, in the present embodiment, it is assumed that the form image data is input at an inclination of 0 degree, and the straight line j is parallel to the horizontal ruled line of the image entry column, and the straight line i is the horizontal ruled line of the image entry column d. On the other hand, a description will be given of a case where the angle is described with an inclination of β degrees. Next, the operation in FIG. 4 will be described. First, as disclosed in Japanese Patent Application No. 7-247644, "Image Data Input Processing Method and Apparatus Thereof", an angle of scanning from 0 to ± α (allowable tilt angle | α | ≧ | β | ), While counting the number of black pixels on each scanning line. As a result, the horizontal line scanning lines L1 to L
5, L7, L6 can be detected, and the horizontal line segment detection scanning line L6 can be detected at an inclination β degrees. Next, a shape similar to a square whose one side on the detected horizontal line segment detection scanning lines L1 to L9 is x is 2
Check if there are any. Then, the horizontal line segment detection scanning lines L5 and L5 which have detected two shapes similar to a square with one side x.
The horizontal line detection scanning lines L6 and L7 sandwiched between 8 are determined to be straight lines described in the image entry column d, and are not used for inclination correction.

When there are a plurality of image entry columns d, a horizontal line detection scanning line sandwiched between an odd-numbered ruled line and an even-numbered ruled line where a mark is detected from the top of the form is used for inclination correction. Good if not. For this reason, the inclinations of the detected horizontal line segment detection scanning lines L1 to L5, L8, and L9 and the sub-scanning coordinates Y0 to Y4, Y7, and Y8 are used, and the Japanese Patent Application No. Hei 7-24764 is used.
4 As disclosed in “Image Data Input Processing Method and Apparatus”, tilt correction can be performed by rotating image data according to each ruled line tilt. As a result, the inclination correction can be performed without being affected by the straight line i having the specified length or more described in the image entry column d.

In this embodiment, a shape similar to a square is used as a mark indicating the four corners of the image entry column d. However, the present invention is not limited to this example, and if another shape is registered, it is used. You can also. Further, in this embodiment, the marks indicating the ruled lines in the image entry column d are four corners, but may be on the extension of the horizontal ruled line and the vertical ruled line, and the number of marks is two. Alternatively, there may be one.

Further, in order to improve the reliability of the image entry field d, it is possible to compare the distance or the position between the two marks so as to correspond to two ruled lines. Next, the operation of the format identification unit 3 will be described. FIG. 5 is a diagram for explaining a format identification method in the format identification unit according to one embodiment of the present invention.

In the figure, TH1 is the horizontal ruled line threshold of the black pixel distribution, B1 to B8 are the black pixel distribution in the main scanning direction, T1
H2 is the vertical ruled line threshold of the black pixel distribution, A1 and A2 are the black pixel distributions in the sub-scanning direction, S0 to S7 are the sub-scanning positions of the black pixel distributions B1 to B8 in the main scanning direction, T0, T
Reference numeral 1 denotes a main scanning position of the black pixel distributions A1 and A2 in the sub-scanning direction.

The marks e to h indicating the four corners of the image entry field 13 are squares each having a side of x.
Is stored in the format identification unit 3. Next, the operation of the format identification unit 3 in FIG. 5 will be described. First, in order to detect ruled lines required for format identification, black pixel distributions are taken in the main scanning direction and the sub-scanning direction.
Black pixel distributions B1 to B8 equal to or larger than the threshold value TH1 of the horizontal ruled line and sub-scanning positions S0 to S7 are detected. Similarly, the vertical ruled line threshold T
H2 or more black pixel distributions A1, A2 and main scanning positions T0,
Detect T1. At this time, the straight line i is not detected because it is inclined.

Next, square marks e to h on the main scanning lines of the detected black pixel distributions B1 to B8 are detected by a method similar to the method described with reference to FIG. Accordingly, in the present embodiment, the marks e and f are set at the sub-scanning position S6 at the sub-scanning position S4.
Marks g and h can be detected. That is, it can be seen that the sub-scanning positions S4 and S6 indicate horizontal ruled lines in the image entry column d. Since the black pixel distribution B6 in the sub-scanning range sandwiched between the sub-scanning positions S4 and S6 is found to be a straight line j in the image entry column d, it can be determined that it is not ruled line information used for format identification.

When there are a plurality of image entry fields d, a horizontal line detection scanning line sandwiched between an odd-numbered ruled line and an even-numbered ruled line where a mark is detected from the top of the form must be used for format identification. I just need. Then, the number of ruled lines in the format definition information registered in advance is compared with the relative interval to identify the format. In the present embodiment, there are seven horizontal ruled lines, which are S1-S
0, S2-S1, S3-S2, S4-S3, S6-S
4, two vertical ruled lines at intervals of S7-S6, and this interval is T
Since 1−T0, format definition information close to this value may be searched for.

In the present embodiment, a shape similar to a square is used as a mark indicating the four corners of the image entry column d. However, if another shape is registered, it can be used.
Further, in the present embodiment, the marks indicating the ruled lines in the image entry column d are four corners, but may be on an extension of the horizontal ruled line, and the number of marks is two on each ruled line.
One is possible.

Further, in order to improve the reliability of the image entry column d, the distance or the position between the two marks can be compared to obtain a correspondence between two ruled lines. Furthermore, in the format identification, the number of ruled lines and the ruled line interval are simply compared with a plurality of registered format definition information to determine the format definition information.
In order to perform format identification in consideration of expansion and contraction of a form image and lack of ruled line information, it is preferable to perform format identification using a method as disclosed in Japanese Patent Application No. 7-115924 “Form form identification method and apparatus”.

The straight line described in the image entry column d is a horizontal line. However, even when a vertical straight line is described, a vertical straight line between main scans sandwiched by marks e to h is similarly used. It is sufficient to perform format identification without using. Next, the operation of the character / image cutout unit 5 will be described. FIG.
FIG. 4 is a diagram for explaining a method of extracting a character / image area in a character / image extraction unit according to one embodiment of the present invention.

FIG. 7A is a diagram for explaining a method of cutting out a character area, where k is a ruled line detection area indicating a character entry box c, and TH3 is a horizontal ruled line indicating a character entry box c. The horizontal ruled line threshold, m is the horizontal ruled line threshold T in the detection area k.
H3 or more, horizontal ruled line black pixel distribution, TH4 is a vertical ruled line threshold for detecting a vertical ruled line indicating the character entry column c, p is a vertical ruled line black pixel distribution in the detection area k, which is a vertical ruled line threshold TH4 or more,
n is a character area cut out from the character entry field c, C0 to C
2 is the sub-scanning coordinates of the black pixel distribution m, and D0 to D7 are the sub-scanning coordinates of the black pixel distribution p.

FIG. 7B is a diagram showing a method of cutting out the image area, where q is a ruled line detection area indicating the image entry field d, and t is a horizontal ruled line black pixel in the detection area q that is equal to or larger than the horizontal ruled line threshold TH5. Distribution, TH6 is a vertical ruled line detection threshold for detecting a vertical ruled line in the message entry field d, s is a vertical ruled line black pixel distribution equal to or greater than the vertical ruled line threshold TH6, r is an image area cut out from the image entry field d, E0, E1 Is the sub-scanning coordinates of the black pixel distribution t, and F0 and F1 are the main scanning coordinates of the black pixel distribution s.

Next, the operation in FIG. 6 will be described. FIG.
As shown in (A), in order to detect the character entry column c, the selected format definition information (first (A) is a fixed position from the ruled line coordinates S0, T0 in the character entry column c of 2 rows and 7 columns. The second rule (B) exists in the detection area c located at a certain position from the ruled line coordinates S0, T0 in the image entry field d. There is a black pixel distribution in the main scanning direction and the sub-scanning direction of the detection area k which exists and is described in the horizontal ruled line threshold TH5 and the vertical ruled line threshold TH6. Then, the sub-scanning coordinates C0 to C2 of the horizontal ruled line black pixel distribution m equal to or more than the horizontal ruled line threshold TH3, and the main scanning coordinates D0 to D of the vertical ruled line black pixel distribution p equal to or more than the vertical ruled line threshold TH4.
Get 7. However, at this time, the horizontal ruled line threshold value TH3 is registered in the format definition information as a value proportional to the number of characters, and the vertical ruled line threshold value TH4 is registered in the format definition information. Thus, the detected sub-scanning coordinates C
0 to C3 and main scanning coordinates D0 to D7 indicate the coordinates of the corners of each character entry frame in the character entry box c. Then, the character area n can be cut out by cutting out the inner area in consideration of the ruled line width from the coordinates of the corners of each character entry frame.

Next, as shown in FIG. 3B, in order to detect the image entry field d, the black pixel distribution in the main scanning direction and the sub-scanning direction is performed for the second detection area q of the format definition information. Take. Then, the sub-scanning coordinates E0 and E1 of the horizontal ruled line black pixel distribution t having the horizontal ruled line threshold TH5 or more and the main scanning coordinates F0 to F1 of the vertical ruled line black pixel distribution s having the vertical ruled line threshold TH6 or more are obtained.
The coordinates of the corners of the image entry field d are indicated by the sub-scanning coordinates E0 and E1 and the sub-scanning coordinates F0 and F1 thus detected.
Then, the image area r of the form image data 10 can be cut out by cutting out the image area r corresponding to the inner area in consideration of the ruled line width from the corner coordinates from the image memory unit 1.

In addition, the present invention constructs the configuration shown in FIG. 3 as a program and stores it in a portable storage medium such as a disk device connected to the processing device, a floppy disk, or a CD-ROM. It can be used for general purposes. It should be noted that the present invention is not limited to the above-described embodiment, but can be variously modified and applied within the scope of the claims.

[0037]

As described above, according to the present invention, even if the same length and straight line as the ruled line used for inclination correction and format identification are described in the form image filled-in column using the ruled line, the image can be stored in the image filled-in column. , The inclination can be corrected and the format can be identified without being affected by the written image, and an accurate image can be cut out.

Further, since the shape of the mark indicating the image entry column can be changed, it is possible to use a shape that is unlikely to appear depending on the service. Further, the present invention can be processed even when there is no image entry column or when there is no mark in the image entry column, and can maintain compatibility with the conventional form processing.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the principle of the present invention.

FIG. 2 is a principle configuration diagram of the present invention.

FIG. 3 is a configuration diagram of a form image processing apparatus of the present invention.

FIG. 4 is a diagram for explaining a method of tilt correction in a tilt correction unit according to one embodiment of the present invention.

FIG. 5 is a diagram for explaining a format identification method in a format identification unit according to one embodiment of the present invention.

FIG. 6 is a diagram for explaining a method of extracting a character / image area in a character / image extraction unit according to one embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image memory part, form image storage means 2 Inclination correction part, inclination correction means 3 Format identification part, format identification means 4 Mark storage part 5 Character / image cutout part 6 Character recognition part 7 Result editing part 8 Form input means

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Hiroshi Sekiya, Inventor Hiroshi 3-chome, Nishishinjuku 3-chome, Shinjuku-ku, Tokyo Japan (72) Kenji Ogura, Inventor Kenji 3-192-1, Nishishinjuku, Shinjuku-ku, Tokyo No. Japan Telegraph and Telephone Corporation

Claims (9)

[Claims] 1. A form image processing method for recognizing a character and an image cut out by inputting a form on which a character or an image is described by a facsimile or a scanner, wherein a ruled line indicating a character and image entry field of the form and a position of the image entry field When a form in which the mark indicating is described in a color that can be detected by a facsimile or a scanner is input, the form image of the input form is stored in a memory, and the detection of the horizontal ruled line of the form image and the horizontal ruled line or
Detecting a mark on an extension of the horizontal ruled line, performing tilt correction using a horizontal ruled line in which the mark is present or a horizontal ruled line other than the sub-scanning range sandwiched between the extended lines of the horizontal ruled line; Detection of horizontal ruled lines and a mark on the horizontal ruled line or an extension of the horizontal ruled line from the image of the form, and identifying the form using horizontal ruled lines other than the sub-scanning range sandwiched between the marks. And a form image processing method. 2. A vertical ruled line is detected from the form image on which the tilt correction has been performed, a mark on the vertical ruled line or an extension of the vertical ruled line is detected, and a mark other than the main scanning range sandwiched by the mark is detected. 2. The form image processing method according to claim 1, wherein the form is identified using a vertical ruled line. 3. The form image processing method according to claim 1, wherein a mark indicating the position of the image entry field can be registered. 4. A form image processing apparatus for recognizing a character and an image cut out by inputting a form on which a character or an image is described by a facsimile or a scanner, and comprising a ruled line indicating a character and image entry field of the form and an image entry field Form input means for inputting a form in which a mark indicating the position of the form is described in a color that can be detected by a facsimile or a scanner, form image storage means for storing a form image of the form input by the form input means, and the form image The form image is acquired from the storage means, the horizontal ruled line of the form image is detected, and a mark on the horizontal ruled line or on an extension of the horizontal ruled line is detected, and the horizontal ruled line where the mark exists or the horizontal ruled line is detected. An inclination correction unit that performs inclination correction using a horizontal ruled line other than the sub-scanning range sandwiched by extension lines of the ruled line; From the form image subjected to the inclination correction, horizontal ruled lines are detected, and marks on the horizontal ruled lines or on extensions of the horizontal ruled lines are detected, and horizontal ruled lines other than the sub-scanning range sandwiched between the marks are used. A form identifying means for identifying the form. 5. The format identification means detects a vertical ruled line from a form image on which the tilt correction has been performed, detects a mark on the vertical ruled line or an extension of the vertical ruled line, and sandwiches the mark. 5. The form image processing apparatus according to claim 4, further comprising means for identifying the form using vertical ruled lines other than the main scanning range. 6. The form image processing apparatus according to claim 5, further comprising a mark registration unit for registering a mark indicating a position of the image entry column. 7. A storage medium storing a form image processing program for recognizing characters and inputting a form on which a character or image is described by a facsimile or a scanner, and a ruled line indicating a character and image entry field of the form And a form input process for inputting a form in which a mark indicating the position of the image entry field is described in a color that can be detected by a facsimile or a scanner, and a form image storing process for storing a form image of the form input by the form input process. Acquiring a form image from the form image storage process, detecting a horizontal ruled line of the form image, detecting a mark on the horizontal ruled line or an extension of the horizontal ruled line, Tilt correction is performed using a ruled line or a horizontal ruled line other than the sub-scanning range sandwiched by extensions of the horizontal ruled line. A horizontal ruled line, a mark on the horizontal ruled line or an extension of the horizontal ruled line is detected from the form image on which the tilt correction has been performed by the tilt correction process. A form identification process for identifying a form using horizontal ruled lines other than the specified sub-scanning range. 8. The format identification process includes detecting a vertical ruled line from a form image on which the tilt correction has been performed, detecting a mark on the vertical ruled line or an extension of the vertical ruled line, and sandwiching the mark. 8. A storage medium storing a form image processing program according to claim 7, including a process of identifying the form using vertical ruled lines other than the main scanning range. 9. The storage medium storing the form image processing program according to claim 7, further comprising a mark registration process for registering a mark indicating a position of the image entry field.