JPH03254275A - Picture magnification and reduction method - Google Patents

Abstract

PURPOSE:To form a magnified/reduced picture including an optimum character font in response to the magnification rate by synthesizing a character font selected among prepared plural kinds of character fonts and a magnified/reduced image area so as to form the picture. CONSTITUTION:A character font storage means 5 stores optional size of character fonts in advance depending on each character code. Upon the receipt of character information and a magnification rate alpha, a character magnification reduction means 5 selects a character font in response to a new character size calculated by the character size Zi and the magnification rate alpha and outputs a position PiXalpha as position information to be arranged on the magnification reduction picture to a synthesis means 7 together with the selected character font. The synthesis means 7 expands the character font from the character magnification/reduction means 6 to the position PiXalpha as the picture data to synthesize the image and the character font and outputs the result as the magnification/reduction picture. Thus, the magnification/reduction picture including the optimum character font is formed in response to the magnification rate.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for enlarging and reducing an image, and more particularly to a method for enlarging and reducing an image including a character font.

[Prior Art] In recent years, with the rapid spread of image information devices such as facsimile machines and image communication systems, image data is being exchanged between image information devices having various paper sizes and various image resolutions. There is. Intercommunication between devices of different sizes and resolutions requires a scaling method involving some pixel density conversion technique.

Furthermore, when obtaining duplicate image outputs that are scaled at various magnifications or scales using, for example, a printer or a copying machine, from one original image data, a scaling method that involves pixel density conversion technology is required.

Pixel density conversion techniques for images that include multivalued images such as images, binary images such as characters, etc., include a method of converting pixel density by simply performing duplication or thinning processing,
A method using an interpolation function, an optical method, etc. are common, and image enlargement/reduction processing is performed using such pixel density conversion technology.

[Problems to be Solved by the Invention] As images containing a mixture of text and images become more common, it is necessary to apply conventional image scaling methods to images containing text fonts used in word processors, etc. Opportunities for its use are also increasing.

However, since the structure of the binary data of a character font is determined according to the number of pixels, for example, the 8 x 8 point character font shown in Figure 6 (0) included in an image can be processed using conventional methods. When the image is enlarged three times, it is converted to an image equivalent to 24×24 points as shown in FIG. 6(b), and an unnaturally large step-like outline occurs in the diagonally shaded portion of the enlarged character. Conversely, if the 24X, 24 point character font included in the image, for example shown in Figure 7(b), is reduced to 1/3 by the conventional method, it will be as shown in Figure 7(a). When the image is converted into an image equivalent to 8x8 points such as 8x8 points, blurring and missing parts occur, resulting in noticeable unnaturalness and characters becoming indistinguishable.

An object of the present invention is to provide an image enlargement/reduction method that can maintain characters represented by a character font in a natural and easy-to-read state even when the image is enlarged or reduced.

[Means for Solving the Problems] In order to achieve the above-mentioned object, the method of enlarging/reducing an image including character fonts according to the present invention recognizes character fonts in an image, and enlarges/reduces characters in the image excluding the recognized character fonts. The image area is enlarged or reduced according to the set magnification, the recognized character font is selected from multiple types of character fonts prepared in advance according to the magnification, and the selected character font and the scaled image area are displayed. The feature is that a scaled image is formed by combining the two images.

[Operation] In the image scaling method of the present invention, first, character fonts in an image containing character fonts are recognized. Next, for the image area in the image excluding the recognized character fonts, a method of converting the pixel density by simply performing overlap or thinning processing similar to conventional scaling technology, or a method using an interpolation function. The image is enlarged or reduced according to a set magnification using an optical method or the like. On the other hand, for the recognized character font, a character font corresponding to the magnification is selected from among a plurality of types of character fonts prepared in advance. Finally, the selected character font and the scaled image area are combined to form a scaled image, so no matter what the scaling factor, the optimal image size is determined according to this scaling factor. A scaled image including character fonts can be created.

These effects of the present invention will become clearer from the following examples of the present invention, and other effects of the present invention will become clearer.

[Example] Embodiments of the present invention will be described based on the drawings.

As an embodiment of the image enlarging/reducing method of the present invention, an image processing apparatus 1 is shown in a block diagram in FIG.

In the figure, image data including an image composed of multivalued image data and a character font is input to an image processing apparatus 1. The character font is an N×N (N22) character font composed of binary data of “1” and “0”. Further, an enlargement/reduction magnification α is set to the image processing device 1 from the outside.

Character recognition means 2 recognizes characters included in input image data using character pattern recognition technology.

For example, when the input image data is an original image, the character recognition means 2 is a character pattern recognition device such as an OCR (optical character reader) that optically reads and inputs the characters in the input image data and recognizes them as a character pattern. Contains.

Note that as the character pattern recognition technology here, a stroke matching method, a stroke analysis method, etc. that analyzes the structure of the strokes and outlines of characters may be used, or a stroke analysis method that analyzes the distribution of strokes and outlines of characters may be used. method may be used. Also, other character feature extraction methods may be used. Among these methods, it is preferable to select the most suitable method depending on the type of characters to be recognized (alphabet, katakana, hiragana, kanji, etc.). In any case, if the type of character font to be recognized is registered in the character recognition means 2 in advance, such character recognition can be performed more easily and quickly.

The character recognition means 2 further includes an area separation device for separating areas in the input image data in which character fonts exist.

In one example of a region separation device, a character region and an image region are separated by utilizing a difference in spatial frequency distribution regarding density between these regions. That is, as shown in the graph of FIG. 2(b), a graph as shown in FIG. Interval density changes in the original image 100 are determined, and areas where steep density changes are concentrated are determined as character area candidates 1G2b. Further, in FIG. 2(b), it is assumed that the area 101b excluding the character area candidate 102b is included in the image area. In FIG. 2(b), the horizontal axis is the position on the scanning line, and the vertical axis is the density (gradation) of the original image 100 at that position. Subsequently, simple binarization processing (conversion into binary data of "0" and "1") is performed on the character area candidate 102b. At this time, the character size may be determined based on the distribution where the density change is steep.

Next, character pattern recognition is performed on the simple binarized image data using the aforementioned character pattern recognition device. In this way, only the portion where character pattern recognition was actually possible is set as a character region, and the image portion other than that portion is separated as an image region.

Note that without providing such an area separation device in the character recognition means 2, an image area in which a character font exists is specified to the character recognition means 2 from the outside, and the character pattern recognition device extracts characters from the specified image area. The character recognition means 2 may be configured to recognize.

Further, the character recognition means 2 may be configured using a character pattern recognition device other than OCR that performs character pattern recognition on input of an image converted into a digital signal instead of an original image.

After character recognition is completed as described above, the character recognition means 2 deletes the recognized character font image part from the input image data and converts the image part into the image enlargement/reduction part 3.
Output to. Further, the character recognition means 2 outputs the character code Ci1, character size Zi1, and position Pi in the image data to the character information storage means 4 for each recognized character.

When the image area image portion excluding the character font and the magnification α are input, the image scaling means 3 converts the pixel density by simply performing overlapping or thinning processing in the same way as conventional scaling methods; The image part is magnified by α using a method using an interpolation function or an optical method.
Enlarge or reduce.

The character information storage means 4 stores the character code C11 character size Zi1 inputted from the character recognition means 2 and the position Pi in the image data as character information (Ci, Zi).

Pi).

The character font storage means 5 stores character fonts of arbitrary sizes for each character code in advance.

For example, the character font storage means 5 stores the 8×
8 point character font, 24 as shown in Figure 7(b)
It stores character fonts of many sizes, including character fonts of x24 points.

The character enlarging/reducing means 6 converts character information (Ci, Zi.

Pi) and magnification α are input, select a character font corresponding to the new character size calculated from the character size Zi and magnification α, which is the character code Ci stored in the character font storage means 5. do. The character enlarging/reducing means 6 also outputs the position PiXα as positional information where the character font should be placed in the enlarged/reduced image to the synthesizing means 7 together with the selected character font.

The synthesizing means 7 develops the character font from the character enlarging/reducing means 6 as image data on the enlarged/reduced image part from the image enlarging/reducing means 3 at position PiXα, thereby synthesizing the image and the character font. and output the result as a scaled image.

The flow of image processing in this embodiment will be explained based on FIG. 3.

In FIG. 3, input image data 11 is first processed by character recognition means 2 into image data 1 containing only an image portion.
It is separated into two parts. Regarding the character font recognized at this time, character information (Ci, Zi.

Pi) is stored in the character information storage means 4.

The image data 12 is converted to a magnification α by the image enlarging/reducing means 3.
The image data 13 is obtained by enlarging or reducing the size according to the size.

On the other hand, in the character enlarging/reducing means 6, a character font having the optimum size when the character font is enlarged/reduced by the scaling factor α is selected from the character font storage means 5, and the character font after scaling is selected from the character font storage means 5. A position Pixα is calculated. That is, character information (
Ci, Zi, Pi) is character information (CiSZixα, Pi
Xα).

Finally, the image data 13 is converted into character information (
One composite image data 14 is obtained by combining with character fonts based on fonts (Ci, Zixα, PiXα).

For example, if we take the X and Y coordinates with the upper left corner of the image data 11 as the origin O as shown in FIG. 4, the character font at the position Pi=(XSY) in the image data 11 is
The enlarged/reduced image data 14 is synthesized at a position Pi=(XXα, Y×α) with respect to the origin Oα.

Therefore, for example, if the image data 11 including the 8x8 point character font shown in FIG.
Enlarged image data 14 including the 24×24 point character font shown in FIG. 3(b) at the position Pi=(Xx3, Y×3) is finally obtained.

Conversely, if the image data including the character font shown in Figure 5(b) is reduced to 1/3, the image data shown in Figure 5(a)
Reduced image data including the character font shown in is obtained.

As described above, according to this embodiment, it is possible to form a scaled image by compositing images so as to include a character font most suitable for the pixel density and image size after scaling.

[Effects of the Invention] According to the present invention, after selecting a character font according to the scaling factor for the character font recognized in an image from among a plurality of pre-prepared character fonts, the selected character font is Since a scaled image is created by compositing the scaled image area with the scaled image area, it is possible to create a scaled image containing the optimal character font for any scale ratio. can.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an image processing device according to an embodiment of the present invention, FIG. 2 is an explanatory diagram of separation processing in an example of the region separation device according to the embodiment of FIG. 1, and FIG. FIG. 4 is an explanatory diagram of the flow of image processing. FIG. 4 is an explanatory diagram of the positional relationship of character fonts before and after enlarging processing in this embodiment. FIG. 5 is an explanatory diagram of enlarging/reducing processing in this embodiment. The figure is an explanatory diagram of enlargement/reduction processing by a conventional method. 1... Image processing device, 2... Character recognition means, 3... Image scaling means, 4...
... Character information storage means, 5 ... Character font storage means, 6 ... Character enlargement/reduction means, 7...
...Synthesis means, 11.12.13.14.100...
····image data.

Claims (1)

[Claims] A method for scaling an image containing character fonts, the method comprising:
A character font in an image is recognized, an image area in the image excluding the recognized character font is enlarged or reduced by a set magnification, and a character font according to the magnification is prepared in advance for the recognized character font. An image scaling method comprising: selecting a character font from among a plurality of types of character fonts, and forming a scaled image by composing the selected character font and the scaled image area.