JP2843638B2 - Character image alignment method - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a character image alignment method, and more particularly, to an apparatus for reading a character image on a document by a scanner and aligning and outputting each character image. A variable output position of the character image.

[Conventional technology]

Generally, in image processing, a method is employed in which a character image read by a scanner is output with its inclination, size, etc. adjusted. In this case, conventionally, the bottom coordinate of the character image read by the scanner is aligned with the baseline,
Or they were aligned with the center line. FIG. 6 (a)
FIG. 6B shows an output character string in the case where the bottom coordinate of the character image is aligned with the base line after extracting the character from the character string on the second line of the original in FIG. 2, and FIG. This shows the output character string when aligned.

Note that the basic concept of outputting a read character image with the inclination, size, etc. adjusted is described in, for example,
No. 54-67425.

[Problems to be solved by the invention]

Just aligning the bottom coordinate of the character image read by the above-mentioned conventional scanner with the base line or aligning the center line only requires the repetition of “ょ”, “ょ”, punctuation “.”, “,” And square brackets. ""","""like,
There is a problem that characters, symbols, and the like that are originally printed at skewed positions (upper and lower) are output very unnaturally.

An object of the present invention is to arrange and output character images,
An object of the present invention is to provide a character image alignment method in which a specific character (including a symbol) is not output at an unnatural position.

[Means for solving the problem]

To achieve the above object, according to the present invention, when reading a character image on a document and aligning and outputting each character image,
The output position of each character image is basically determined based on the size of the character image and the position on the document.

Claim 1 determines the output position based on the character center line, and outputs the character image such that the smaller the character image, the larger the offset from the reference center line to the position direction in the original image. It is characterized by determining the position.

Further, claim (2) uses the upper and lower reference lines of the line image, and outputs a character image such that the size of the character image is larger than a certain value so that its lowermost coordinate is in contact with the lower reference line; For other character images, output from the position on the document so that the upper end of the character image is in contact with the upper reference line.
It is characterized in that either output is performed so that the center of the character image matches the center of the upper and lower reference lines, or output is performed so that the lower end of the character image is in contact with the lower reference line.

(Operation)

Generally, small characters are skewed (up, down), such as “拗”, ““ ”,“ tsu ”,“. ”,“, ”In punctuation, and“ ”,“ ”in square brackets. Often printed on
For other characters, it is preferable to arrange the disturbances of the original image as much as possible and output the characters without skew or vertical fluctuation of the character image. Thus, in the present invention, when determining the output position, the size of the character image is taken into consideration. This makes it possible to prevent characters that are originally output at skewed positions, such as “.” And “′”, from being output at unnatural positions, while aligning disturbances in the original image as much as possible.

〔Example〕

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 shows a configuration diagram of an embodiment of the present invention. The scanner 11 scans a document and reads a character image as an electric signal. The line cutout unit 12 cuts out the character image read by the scanner 11 one line at a time, and determines coordinates serving as a reference at the time of alignment. The character extracting unit 13 sequentially extracts one character image from one line of character image.
The character image size detection unit 14 detects the size of the cut-out character image. The character image arrangement determining unit 15 determines the output position of the cut-out character image based on the size of the character image and the position on the document. The aligned image output unit 16
The normalized character image is arranged and output at the position determined by the character image array unit 15. The processing in the character image size detection unit 14, the character image arrangement determination unit 15, and the aligned image output unit 16,
Each line of the original document extracted by the line extracting unit 12 and the character extracting unit 13 is repeated by the number of characters in each line. Hereinafter, each embodiment corresponding to claims (1) and (2) will be described in detail with specific examples.

Example 1 A document shown in FIG. 2 is assumed. In the present embodiment, in order to determine the character output position based on the character center line, the line cutout unit 12 cuts out one line of the image, and the character cutout unit 13 with the center of the upper end and the lower end of the cut out line image as the reference line. Send to In general, projection in the X direction is used for line extraction, but any method may be used as an extraction method. Further, as the reference center line, another appropriate method may be used, for example, using the center coordinates of the highest character in the line. FIG. 3 shows a character image string extracted from the second line of the original shown in FIG. Third
In the figure, 30 is the height of the character "1", -13 is the offset of the period "." From the reference line (center line), and +11 is the same offset of the symbol "".

The character extracting unit 13 extracts a character image for one character from the character image string extracted from the line, and outputs a character image size detecting unit 14.
Send to As a method of extracting characters, a method of extracting characters using projection in the Y-axis direction in a line is generally used, but any other method may be used.

The character image size detection unit 14 calculates the height H of the extracted character image, and sends the character image height H to the character image arrangement determination unit 15 as the character size S. In the present embodiment, the character image height is used as the character size. However, other appropriate parameters such as the area of the circumscribed rectangle of the character image may be used.

When the offset from the reference line (center line) in the document is f, the character image arrangement determining unit 15 obtains the offset F of the output position from the reference line by the following formula, for example.

F = (f · α) / S (1) The above equation (1) is an example of a method for obtaining F, and an appropriate conversion equation may be used such that F becomes larger as the character becomes smaller. Equation (1) for the second line of the original in FIG.
Table 1 shows the output position of each character image in the case of using. In Table 1, "Method" in the latter part is omitted.

The aligned image output unit 16 arranges and outputs the character images at the positions determined by the character image arrangement determining unit 15. 2 in FIG.
FIG. 4 shows the arrangement of the cut-out characters on the line according to the offset F in Table 1.

Embodiment 2 Here, as shown in FIG. 2, the manuscript is assumed to be a manuscript, and only portions different from those in Embodiment 1 will be described. In the present embodiment, since the upper and lower reference lines are used for each line, the line cutout unit 12 cuts out an image of one line and sets the coordinates of the upper end and the lower end of the cut line image as the upper and lower reference lines, respectively.
Send to 13. The upper and lower reference lines may be set using other suitable methods, such as using the upper and lower coordinates of the highest character in the line.

The character image size detection unit 14 obtains the height H of the cut-out character image, and sends the character image height H to the character image arrangement determination unit 15 as a character size. In this embodiment, the character image height is used as the character size, but other appropriate parameters such as the area of the circumscribed rectangle of the character image may be used.

In the character image arrangement determining unit 15, by the following processing,
Determine the output position of the character image.

If the height of the character image is larger than the threshold value T, the output is performed so that the lower end of the character image is aligned with the lower reference line.

If the height of the character image is smaller than the threshold value T and the offset of the character image is larger than + t, the upper end of the character image is aligned with the upper reference line.

If the height of the character image is smaller than the threshold value T and the offset of the character image is smaller than -t, the lower end of the character image is aligned with the lower reference line.

In other cases, the center of the character image is aligned with the center of the upper reference line.

Table 2 shows an example of determining the output position of each character image when the present embodiment is used for the second line of the original shown in FIG.

The aligned image output unit 16 arranges and outputs the character images at any of the positions to determined by the character image arrangement determining unit 15. If each cut-out character on the second line in FIG. 2 is aligned according to Table 2, the result is as shown in FIG.

〔The invention's effect〕

As is apparent from the above description, according to the present invention, a character image on a document is read by a scanner, line cutout,
When performing character segmentation and arranging each character image for output, it is possible to prevent specific characters that are originally skewed from being output to unnatural positions while aligning the disorder of the original image. it can.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment for achieving the method of the present invention, and FIG.
FIG. 3 shows an example of a document, FIG. 3 shows a cut-out line image, FIGS. 4 and 5 show examples of an output image according to the present invention, and FIG. 6 shows an example of a conventional output image. FIG. 11 ... scanner, 12 ... line cutout section, 13 ... character cutout section, 14 ... character image size detection section, 15 ... character image arrangement determination section, 16 ... character image output section.

Claims (2)

(57) [Claims] An apparatus for reading a character image on a document and arranging and outputting each character image determines an output position of each character image based on a size of the character image and a position on the document. A character image alignment method that determines an output position based on a character center line, and outputs the character image such that a smaller character image has a greater offset from the reference center line in the position direction in the original image. A character image alignment method characterized by determining a position. 2. An apparatus for reading a character image on a document and arranging and outputting each character image, wherein an output position of each character image is determined based on a size of the character image and a position on the document. Character image alignment method, using the upper and lower reference lines of the line image, if the size of the character image is larger than a certain value, output the character image so that the bottom coordinate is in contact with the lower reference line, For other character images, output from the position on the document so that the upper end of the character image is in contact with the upper reference line, output so that the center of the character image matches the center of the upper and lower reference lines, or output the lower end of the character image. Is output so as to be in contact with the lower reference line.