JP3122476B2 - Automatic document copy machine - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic document copying apparatus for automatically copying a document.

[0003]

2. Description of the Related Art Conventionally, as described in Japanese Patent Application Laid-Open Nos. 2-127775 and 2-127776, this type of document refining apparatus has been disclosed in Japanese Patent Application Laid-Open No. HEI 2-127776. According to the vertical and horizontal sizes of the data, the calculation is performed so that a portion lacking in the display area does not occur and the aspect ratio of the image data is 1: 1. Therefore, when the operator inputs the data for copy-writing such as the vertical size and the horizontal size of the image data display area, the document copy-writing apparatus prevents the display area from lacking, and the image data has an aspect ratio of 1: 1. The character string can be copied by calculating so as to be 1.

[0004]

However, in the above-described conventional document copy apparatus, the calculation is performed based on the vertical size and the horizontal size of the image data display area. Another problem is that it is not possible to automatically copy a handwritten character document.

The present invention has been made in view of the above-mentioned conventional problems, and does not require the operator to input any data for clean copy and preserves handwriting.
It is an object of the present invention to provide an automatic document copying apparatus capable of automatically copying a document of handwritten characters while holding it .

[0006]

In order to achieve the above object, a first means comprises a reading means for optically reading a handwritten character string, and a first means for retrieving a handwritten character string read by the reading means. Search means for searching for each line, cutout means for cutting out the handwritten character string of each line extracted by the search means for each character, position, size and vector of each handwritten character cut out by the cutout means
Character component detecting means for detecting data;
The size of each character detected by the component
Classify by Gori so that the average size for each category
Scaling means for scaling the length of the vector while holding the handwriting, and distributing such that the distance between the handwritten character strings on each line scaled by the scaling means and the position of each line are the same. An arrangement means for arranging the character strings and an output means for outputting the character strings of each line arranged by the arrangement means are provided.

[0007]

[0008]

[0009]

[Action] by the upper Symbol configuration, each the size of each character category
And the average size for each category and
Scale the length of the vector while keeping the handwriting
In, or without an operator to any input a fair copy for the data
A handwritten character document can be automatically copied while holding one handwriting .

[0010]

[0011]

[0012]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of the automatic document copying apparatus according to the present invention, FIG. 2 is an explanatory diagram showing a handwritten document read by the automatic document copying apparatus of FIG. 1 and its histogram, and FIG. FIG. 4 is an explanatory diagram showing a line image extracted by the line search unit, FIG. 4 is an explanatory diagram showing a character cut out by the character region extracting unit in FIG. 1, and FIG. 5 is an explanatory diagram showing constituent elements of the character in FIG. 6 is a flowchart for explaining the operation of the character component detection unit in FIG. 1, FIG. 7 is a flowchart for explaining the operation of the character area size changing unit and the position correction unit in FIG. 1, and FIG. FIG. 9 is an explanatory diagram showing the scaling operation of the area size changing unit, FIG. 9 is an explanatory diagram showing vector data before and after scaling by the character area size changing unit of FIG. 1, and FIG. 10 is a document image shaping and reconstruction of FIG. In the explanatory diagram showing the operation of the unit That.

In FIG. 1, an image of a handwritten document or the like is optically read by a reading unit 1 and converted into an electric signal. This electric signal is converted into a digital signal by an A / D conversion unit 2 and stored in a memory 3. Is done. The left side of FIG. 2 shows an example of the handwritten document P. The handwritten document P has a horizontally written handwritten character 13 and a rectangular image 14 such as a photograph. The right side of FIG. 2 shows a histogram 15 in which the handwritten characters 13 and the rectangular image 14 of the handwritten document P are accumulated in the horizontal direction.
Since it is considered that they are added uniformly, the handwritten character 1
The break with 3 is saved.

The character line search section 4 shown in FIG. 1 determines a point at which the cumulative frequency of the histogram 15 changes extremely in the vertical direction based on the image data stored in the memory 3 and regards the point as a character line cutout point. Extract the image. The upper part of FIG. 3 shows a line image 17 extracted by the character line search unit 4,
This line image 17 has each horizontally written character area 16. The lower part of FIG. 3 shows a histogram 18 in which black pixels of each character area 16 are accumulated in the vertical direction.

The character area cutout section 5 shown in FIG. 1 determines each character area 16 by determining a point at which the cumulative frequency of the histogram 18 changes extremely in the horizontal direction. Then, the position of the character portion and the length and width are calculated by removing the margins. FIG. 4 shows a character “a” cut out by the character area cutout unit 5. The character “a” is an x coordinate Xs at a lower left corner of the character area 16, a y coordinate Ys at an upper right corner, and a horizontal length x length. , Vertical length y length
Etc. FIG. 5 shows the character components of the character "A", each character component comprising a plurality of pixels 25,
It is represented by a stroke end point 23 and a vector 24.

As shown in FIG. 6, the character component detector 6 shown in FIG. 1 thins the current character image to be processed consisting of the plurality of pixels 25 (step 19), and then the character image line width. Are sampled, and the average line width is calculated (step 20). Next, the character component detection unit 6
The end point 23 of the character stroke is detected (step 2).
1) A character image is vectorized by calculating a vector 24 based on the intersection of the end point 23 and the stroke (step 22), and the position, size, and vector data of the character area 16 are stored in the data storage unit 7.

As shown in FIG. 7, the character area size changing unit 8 shown in FIG. 1 first classifies the heights of the extracted characters into a plurality of categories (step 26). That is, the characters are classified into categories such as relatively large characters such as a headline, characters of a normal size of a sentence, and relatively small characters such as a prompt sound "tsu" which are characters of a sentence. Next, the average height of the characters representing each category is calculated, and it is determined whether each character belongs to the category by comparing the average height with the height of each character. The magnification is changed to the average height (step 27). That is, as shown in FIG. 8, for example, the size of the cut-out character 32 is changed to the character 34 of the height y ave by the height y length of the character 32 and the average height y ave of the category. Next, as shown in FIGS. 7 and 8, the size changing unit 8 compares the vertical and horizontal ratios of the character (step 28), and increases the horizontal length x length at the same ratio as the vertical scaling ratio. The magnification is changed to x reform (step 29). Note that the horizontal length x reform can be obtained by equation (1). x reform = (x length / y length) * y ave (1)

As shown in FIG. 9, the starting point and the ending point of the vector data of the character "A" are indicated by the addresses in the image area 39 before scaling and the image area 40 after scaling. Therefore, the amount of displacement in each of the vertical and horizontal directions of the vector is calculated based on the vertical and horizontal scaling factors of the cut-out character area 39 before scaling, and the position address of the vector is calculated in accordance with the position correction of the cut-out area 39. Is similarly corrected, and each stroke of the character can be scaled to obtain each stroke of the character area 40 after scaling.

Next, the character area position correcting section 9 shown in FIG. 1 calculates the average length of the blank in the horizontal direction between the characters as shown in FIG. Correction is made using blanks so that the characters are at appropriate positions and the number of characters in one line is the same before and after scaling (step 31). This correction is
The average length x reform of each area after scaling is totaled, and the total value is subtracted from the length of the extracted line image 17 to calculate a blank length that can exist in the line image 17. By dividing the blank length by the number of characters, the average blank length between the respective character areas is calculated.

Next, the document image shaping / reconstructing unit 10 shown in FIG. 1 converts a character string 36 whose vertical direction does not coincide with a character string 36 having a predetermined interval below the line 35 as shown in FIG. Reconfigure. In addition, the distance 37 between the characters in the horizontal direction
Is the average blank length calculated by the position correction unit 9. The image data corrected in this way is temporarily stored in the memory 11, printed and output by the output unit 12, and output.

Therefore, according to the above embodiment, the character line search unit 4 cuts out the handwritten character string for each line, and the character area cutout unit 5 cuts out each character of the handwritten character string in each line. , A line segment element of each character is detected, the size of each character is determined by the size changing unit 8, and the vertical and horizontal lengths 38 are scaled so as to have an appropriate size. The horizontal position of each character is corrected so that the horizontal spacing of
Since the correction is made so that the interval between the character strings on each line is constant by 0, the operator can automatically copy the handwritten character document without inputting any data for the copywriting.

In this case, the number of characters included in one line is the same, and the characters handwritten in a conscious size are scaled at a ratio corresponding to the size. Is output. In addition, since the line segment of the handwritten character is vectorized, it is possible to prevent the character from being blurred, cut off, or to add noise, and to further maintain the handwriting of the scribe.

FIG. 11 is a flow chart for explaining the operation of the character component detecting section 61 in the second embodiment of the automatic document copy apparatus according to the present invention, and FIG. 12 is an explanatory diagram showing the scaling operation. . In FIG. 11, a character component detection unit 61 matches each character extracted by the character area extraction unit 5 in the first embodiment with a dictionary pattern, performs structural analysis using character components and the meaning of characters in a sentence. It is recognized by analysis or the like (step 41). Next, the outline of the font data is vectorized with reference to the character font of the dictionary pattern (step 42). Therefore, as shown in FIG. 12, when the size changing unit 9 enlarges the font data 43 of the dictionary pattern, the characters 44 such as prints are output, so that the handwritten character string can be completely copied.

[0024]

According to the present invention, as described above ,
Classify font size by category and average by category
The vector so that it is the size of
Since the length of the tor is changed, it is possible for the operator to automatically copy the handwritten character document without inputting any data for clear copy and holding the handwriting.

[0025]

[0026]

[Brief description of the drawings]

FIG. 1 is a block diagram showing one embodiment of an automatic document clean-up device according to the present invention.

FIG. 2 is an explanatory diagram showing a handwritten document read by the automatic document copying apparatus of FIG. 1 and a histogram thereof;

FIG. 3 is an explanatory diagram showing a line image extracted by a character line search unit in FIG. 1;

FIG. 4 is an explanatory view showing characters cut out by a character area cut-out unit in FIG. 1;

FIG. 5 is an explanatory diagram showing components of the character in FIG. 4;

FIG. 6 is a flowchart for explaining the operation of the character component detection unit in FIG. 1;

FIG. 7 is a flowchart for explaining operations of a character area size changing unit and a position correcting unit in FIG. 1;

FIG. 8 is an explanatory diagram showing a scaling operation of the character area size changing unit in FIG. 1;

FIG. 9 is an explanatory diagram showing vector data before and after scaling by a character area size changing unit in FIG. 1;

FIG. 10 is an explanatory diagram showing an operation of the document image shaping / reconstructing unit in FIG. 1;

FIG. 11 is a flowchart for explaining an operation of a character component detecting unit in the second embodiment of the automatic document clean-up device according to the present invention.

FIG. 12 is an explanatory diagram showing the scaling operation.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Reading part 4 Character line search part 5 Character area cutout part 6 Character component detection part 8 Size change part 9 Position correction part 10 Document image shaping and reconstruction part 12 Output part 61 Character component detection part

Continuation of front page (58) Fields investigated (Int.Cl. ⁷ , DB name) G06T 1/00 G06T 11/60-17/50 G06K 9 / 00,9 / 82 H04N 1/38-1/393

Claims (1)

(57) [Claims] A reading means for optically reading a handwritten character string; a search means for searching for a handwritten character string read by the reading means and searching for each line; A cutout means for cutting out a handwritten character string for each character, and a position of each handwritten character cut out by the cutout means.
Character component to detect position, size and vector data
Detecting means, and a support for each character detected by the character component detecting means.
Are classified by category and the average size of each category
Of the vector so that
Scaling means for changing the length ; arrangement means for arranging the line so that the distance between the handwritten character strings of each line scaled by the magnification means and the position of each line are the same; and An output means for outputting a character string of each line;