JPH0969156A - Image processor and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To quickly change the magnification of a multivalue image while holding its high quality. SOLUTION: In the case of changing the magnification of a multi-valued image, a binary image preparing part 31 prepares a binary image set up by using each density level as a threshold from the multi-valued image. The contour of the prepared binary image becomes an equal density line in the density gradient of the original multi-valued image. Therefore, an outline variable magnification part 32 extracts the outline vectors of each binary image and applied magnification change/smoothing to the vectors. Since the obtained outlines of respective binary images express equal density lines in the density gradient of the variable magnification multi-valued image, picture elements having the threshold used at the time of generating each binary image as density are written in the multi-valved image area corresponding to the black picture elements of each binary image. The writing processing is applied to all corresponding binary images in order from a binary image with low density up to a binary image with high density and a multi-valued image reproducing part 33 generates a multi- valued image. Processing for smoothing the density gradient of the obtained multi- valued image is executed and a variable magnification multi-valued image is obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and a method thereof, and more particularly to an image processing apparatus and a method thereof for performing scaling processing of a digital multi-valued image.

[0002]

2. Description of the Related Art Japanese Unexamined Patent Publication (Kokai) No. 5-174140 has already been published as an apparatus of this type for digital binary images.

This proposal proposes to scale a binary image.
Rather than scaling the binary image itself, the contour information of the character / line drawing of the binary image is extracted and a scaled image is generated based on the extracted contour information.
Its purpose was to obtain a value image.

Specifically, Japanese Unexamined Patent Publication No. 5-174140 extracts outline vector data from a binary image and performs a smoothing process on the extracted outline vector data. Then, the smoothed outline vector data is scaled at a desired scale factor (arbitrary), and thereby contour drawing and internal painting are performed to regenerate a binary image. In this way, it is intended to obtain a high-quality digital binary image scaled at a desired magnification (arbitrary).

The main parts will be outlined below. FIG. 56 is a diagram best showing the features disclosed in Japanese Patent Laid-Open No. 5-174140. In the figure, a binary image acquisition unit 101 acquires a digital binary image to be subjected to a scaling process,
A binary image in raster scan format is output. The outline extraction unit 102 extracts a rough contour vector (outline vector before smoothing / magnification processing) from a binary image in raster scanning format. The outline smoothing / scaling unit 103 smoothes and scales the rough contour vector data in the form of vector data. The binary image reproducing unit 104 reproduces the raster scanning type binary image data from the outline vector data. The binary image output unit 105 is a printer, a display, or the like that displays binary image data in a raster scanning format, makes a hard copy, or outputs the data to a communication path or the like.

The binary image input unit 101 reads, for example, a document image as a binary image and outputs the read image as a binary image.
The value data is output in a raster scanning format as a raster scanning format, and is composed of a known binary image input device.

The outline extracting unit 102 is composed of, for example, the device described in Japanese Patent Laid-Open No. 4-157578 previously proposed by the applicant of the present application. FIG. 11 shows the binary image acquisition unit 1.
01 shows the scanning form of the binary image data in the raster format output from 01 and also shows the scanning form of the binary image data in the raster scanning format input by the outline extraction unit 102. In this format, the binary image acquisition unit 101
The outline extraction unit 102 inputs the binary image data of the raster scanning format output by the above. In FIG. 57, a pixel 111 indicates a pixel having a binary image during raster scanning, and an area 112 indicates a 9-pixel area including 8 pixels in the vicinity of this pixel 111. The outline extraction unit described in JP-A-4-157578 mentioned above is
The target pixel is moved in the raster scanning order, and the target pixel and the target pixel are selected according to the state (white pixel or black pixel) of each pixel in the 3 × 3 9 pixel region shown in the region 112 for each target pixel. The contour side vector (horizontal vector or vertical vector) existing between the neighboring pixels of is detected. When there is a contour edge vector, the data of the starting point coordinates and the direction of the edge vector are extracted, and the rough contour vector is extracted while updating the connection relationship between the edge vectors.

FIG. 58 shows an example of a contour side vector extraction state between a target pixel and neighboring pixels of the target pixel. In the figure, the mark Δ represents the start point of the vertical vector (or the end point of the horizontal vector), and the mark ◯ represents the start point of the horizontal vector (or the end point of the vertical vector).

FIG. 59 shows an example of the rough contour vector loop extracted by the outline extracting section described above. Here, each grid separated by a grid shows the pixel position of the input image, and the blank grid is filled with white pixels and dot patterns.
The mark means a black pixel. Similar to FIG. 58, the Δ mark represents the starting point of the vertical vector, and the ◯ mark represents the starting point of the horizontal vector.

As can be seen from the example of FIG. 59, the outline extracting unit 102 always alternates the horizontal vector and the vertical vector in the area where the pixels are connected, although the horizontal vector and the vertical vector may have different lengths. And is extracted as a continuous rough contour vector loop. However, the direction in which the extraction process proceeds here is such that the right side is the black pixel region with respect to the direction in which the extraction process proceeds. The starting point of each rough contour vector is extracted as an intermediate position between pixels of the input image. That is, when the existence position of each pixel is represented by an integer (x, y), the starting point of the extracted vector is a value obtained by adding 0.5 to each coordinate value or a value obtained by subtracting 0.5. More specifically, one pixel in the original image is determined as having a significant area (as a "shape") and extracted as a rough contour loop.

The rough contour vector loop group thus extracted is output from the outline extracting section 102 of FIG. 56 in a data format as shown in FIG. That is, it is composed of the total rough contour loop number a extracted from the image and each rough contour loop data group from the first contour loop to the a-th loop. Each rough contour loop data includes the total number of starting points of the contour side vectors existing in the rough contour loop (which can also be considered as the total number of contour side vectors) and the starting point coordinates of each contour side vector in the order of forming the loop. It is composed of a column of (x, y coordinate values) values (the starting points of the horizontal vector and the starting points of the vertical vector are alternately arranged).

Next, in the outline smoothing / magnifying unit 103 shown in FIG. 56, the rough contour vector data (see FIG. 60) output from the outline extracting unit 102 is input, and the smoothing vector data is smoothed and a desired magnification is obtained. It is carried out in the form of outline vector data (coordinate values).

FIG. 61 shows a more detailed structure of the outline smoothing / magnifying section. In FIG. 61, the first smoothing / scaling unit 152 smoothes and scales the input rough contour data at the scaling factor set by the scaling factor setting unit 151. The processing result is further smoothed by the second smoothing unit 153 to obtain a final output. The magnification setting unit 151 may pass a value preset by a dip switch, a dial switch, or the like to the first smoothing / magnifying unit 152, or may be provided from the outside via an I / F (interface). The value to be passed may be passed to the first smoothing / changing unit. The magnification setting unit 151 is means for giving information on how many times to multiply the image size given as an input independently in the main scanning (horizontal) direction and the sub-scanning (vertical) direction.

The first smoothing / magnifying section 152 obtains magnification information from the magnification setting section 151 and performs smoothing / magnifying processing.

FIG. 62 shows a hardware configuration example for realizing the outline smoothing / magnifying processing. In the figure, R
The OM 164 stores the operation processing procedure of the CPU 161.

The output of the outline extraction unit 102 of FIG. 56 is stored as a file (coarse contour vector data) in the disk device 162 in the data format shown in FIG.

The CPU 161 operates according to the procedure given in FIG. 63, and executes outline smoothing / magnifying processing.

In FIG. 63, first, step S170.
Via disk I / O 163
2 reads the rough contour data stored in 2 and stores it in the RAM 16
6 into a working memory area (not shown). Next, in step S171, the first smoothing and scaling processing is performed.

The first smoothing process is performed in open / close loop units of rough contour data. Each contour edge (horizontal vector or vertical vector) vector of each rough contour data is sequentially set as a target edge, and up to three vectors before and after each target contour edge vector (that is, before the target edge 3 Book,
The pattern is divided by the combination of the length and the direction of the side of interest itself, and the side vectors of up to a total of seven side vectors after the side of interest). For each pattern, the coordinate value of the contour point after the first smoothing with respect to the target side and the additional information (hereinafter referred to as corner point information) indicating whether or not the contour point is a corner point are output.
The corner point here means a point located at a meaningful corner, that is, a point that is not changed by the smoothing process. Then, it is determined that the points other than the corner points are points due to the jagged portions due to noise or other factors, notches, etc., and are to be smoothed. Now, the contour points after the first smoothing determined to be corner points are treated as points that are not smoothed by the subsequent second smoothing, that is, their positions are fixed points. In other words, the contour points after the first smoothing that have not been determined to be corner points (hereinafter referred to as non-corner points) are further smoothed by the second smoothing later.

FIG. 64 shows this state, that is, the target rough contour side vector Di and the three side vectors Di-1, Di-2, Di-3 before the target rough contour side vector and the target rough contour side vector. The state of the following three vectors Di + 1, Di + 2, Di + 3 and the state of the contour points after the first smoothing defined for the target side Di are shown. That is, a vector that connects the contour points redefined in this way (diagonal vectors are allowed) is constructed.

The contents of the first smoothing process have been described above.
The data after the first smoothing is sequentially constructed on a predetermined area of the RAM 166. As a result, with respect to the rough contour vector data, the vector data after the first smoothing also allows a vector in an oblique direction, so that it is possible to form an image without jaggies. Thus, the processing of step S171 of FIG. 63 is completed, and the CPU 161 executes step S171.
The second smoothing process of 172 is performed.

In the second smoothing, the data after the first smoothing is input and processed. That is, the number of closed loops, the number of contour points for each closed loop, the coordinate value data string of the first smoothed contour points for each closed loop, and the additional information data string of the first smoothed contour points for each closed loop are calculated. Input and output contour point data after the second smoothing.

As shown in FIG. 65, the contour data after the second smoothing is composed of a closed loop number, a contour point number table for each closed loop, and a coordinate value data string of the second smoothed contour point for each closed loop. To be done.

The outline of the second smoothing process will be described below with reference to FIG. Similar to the first smoothing, the second smoothing is processed for each contour loop, and the processing is advanced for each contour point in each contour loop.

For each contour point, when the contour point of interest is a corner point, the input contour point coordinate value itself is used as the second smoothed contour point coordinate data for that contour point of interest. That is, nothing is changed.

If the contour point of interest is a non-corner point, the coordinate value obtained by weighted averaging the coordinate values of the front and rear contour points and the coordinate value of the contour point of interest is the contour point of interest. To the second smoothed contour point coordinate value. That is,
Let the input contour point of interest, which is a non-corner point, be Pi (xi, yi), and Pi
, The contour point immediately before in the input contour loop of is Pi-1 (xi-1, yi-1) is defined as Pi + 1 (xi + 1, yi + 1), and the focused input contour point Pi Let Qi (x'i, y'i) be the second smoothed contour point for x'i = ki-1.xi-1 + ki.xi + ki + 1.xi + 1 y'i = ki-1・ Yi-1 + ki ・ yi + ki + 1 ・ yi + 1 (1) is calculated. Where ki-1 = ki + 1 = 1/4, k
i = 1/2.

In FIG. 66, points P0, P1, P2, P3,
P4 is a part of the first smoothed continuous contour point sequence as an input, P0 and P4 are corner points, and P1, P2, and P3 are non-corner points. The processing results at this time are points Q0,
It is indicated by Q1, Q2, Q3 and Q4. Since P0 and P4 are corner points, their coordinate values are Q
The coordinates are 0 and Q4. Further, the point Q1 has a value calculated from P0, P1 and P2 according to the above-mentioned formula as a coordinate value.
Similarly, Q2 is from P1, P2, and P3, and Q3 is from P2, P3, and P4.
The value calculated according to the above equation is used as the coordinate value.

In this way, the CPU 161 controls the RAM 1
A second smoothing process is performed on the first smoothed contour data in a predetermined area on 66. This process proceeds from the first loop to the second loop and the third loop in sequence for each loop, and the process of all the loops ends, whereby the second smoothing process ends. In the processing of each loop, the processing proceeds from the first point to the second point and the third point in sequence, and when the processing shown in the equation (1) is completed for all contour points in the corresponding loop, the processing of the loop is performed. Ends and the process proceeds to the next loop.

When there are L contour points in the loop, the point before the first point is the Lth point. In other words, the point after the L-th point is the first point. In the second smoothing, the same number of contour points as that of the first smoothed contour data to be input is generated, and the number of contour points on each loop remains unchanged. CPU 161
Outputs the above result to another area of the RAM 166 or the disk device 162 in the form shown in FIG. 65, and ends the process of the second smoothing process (step 172).

Next, the CPU 161 proceeds to step 173 and outputs the data obtained as a result of the second smoothing to the I / O 163.
The image is transferred to the binary image reproducing unit 104 via, and the series of processing shown in FIG. 56 is completed.

The binary image reproducing unit 104 can be constituted by, for example, the device described in Japanese Patent Application Laid-Open No. 5-20467 proposed by the applicant of the present application. According to the apparatus, a binary image generated by filling the area surrounded by the vector graphic represented by the contour data based on the second smoothed contour data transferred via the I / O. Can be output in a raster scan format. Also, the same proposal is visualized by using a value image output unit such as a video printer as described in the description. Now, Japanese Patent Laid-Open No. 6-1249
The proposal of No. 0 is based on the above-mentioned Japanese Patent Laid-Open No. 5-174140.
This is a further improvement of the No. 1 problem so that a low-magnification variable-magnification image does not become fat. That is, in the outline extraction unit of Japanese Patent Laid-Open No. 5-174140, a black pixel between a white pixel and a black pixel of an original image is extracted (a black pixel region is narrower than a white pixel region), and It is modified to perform outline smoothing in accordance with this.

According to the above-mentioned Japanese Patent Laid-Open No. 5-174140, it is possible to obtain a good scaled image even in a binary image, especially in a character, a line drawing, a table and a figure.

On the other hand, as a scaling process for a multi-valued digital image, conventionally, it has been general to perform a scaling process by performing an interpolation process using pixel values of the multi-valued image.
The interpolation processing is also a known scaling method such as a method of interpolating the pixel values between the sampled original image pixels with a bilinear linear function or a method of approximating the sampling function with a cubic expression. , "Introduction to Computer Image Processing", Tamura,
It has been introduced in Soken Publishing.

[0034]

However, regarding the scaling processing for the multi-valued digital image described above, in the technique of focusing only on the pixel value of the original image and interpolating the pixel value of the image to perform scaling, In some cases, deterioration of image quality such as jaggies and generation of lattice distortion was caused.

The present invention has been made in view of the above conventional example, and an object of the present invention is to provide an image processing apparatus and method for preventing deterioration of image quality such as jaggies and generation of lattice distortion.

Also, since only a certain density level of equal density is drawn in the binary image from the equal density line vector data and the multivalued image is reproduced by using this, it is drawn in the binary image. An object of the present invention is to provide an image processing apparatus and method capable of reproducing a multi-valued image at a higher speed than a case where a multi-valued image is reproduced after all the density lines are once filled.

In addition, 2 for each density level from the input multi-valued image
When a value image is created, by performing the process of not creating a binary image in a low density area or a high density area where pixels having that density value do not exist, high-speed processing is possible without degrading image quality. It is an object of the present invention to provide an image processing apparatus and a method capable of reducing the cost and working memory.

[0038]

SUMMARY OF THE INVENTION It is an object of the present invention to provide an image processing apparatus and method which prevent jaggies and lattice distortion from occurring when scaling a multi-valued image.

In order to achieve the above object, the image processing apparatus of the present invention has the following configuration.

An image processing apparatus for executing a scaling process on a multi-valued image, including iso-density line vector extraction means for extracting an outline vector for each density level of the multi-valued image, and each extracted density level Magnification / smoothing means for scaling and smoothing the outline vector for each image, and a multivalued image generating multivalued image based on the outline vector of each density level obtained by the processing of the scaling / smoothing means. And an image generating means.

The image processing method of the present invention has the following configuration.

An image processing method for executing scaling processing on a multi-valued image, which comprises an iso-density vector extraction step of extracting an outline vector for each density level of the multi-valued image, and each extracted density level Magnification / smoothing step of scaling and smoothing the outline vector for each, and multi-value generating a multi-valued image based on the outline vector of each density level obtained in the processing of the scaling / smoothing step. An image generation step.

The computer-readable memory of the present invention has the following configuration.

A computer-readable memory for storing a program for executing a scaling process on a multi-valued image, and a code for an iso-density vector extraction process for extracting an outline vector for each density level of the multi-valued image, Based on the code of the scaling / smoothing process for scaling and smoothing the extracted outline vector for each concentration level, and the outline vector of each concentration level obtained in the processing of the scaling / smoothing process. And a code of a multi-valued image generation step for generating a multi-valued image.

With the above arrangement, it is possible to scale a multi-valued image without jaggies or distortions while maintaining high quality.

[0046]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings.

[First Embodiment] FIG. 1 is a block diagram showing the arrangement of an image processing apparatus according to the first embodiment. In the figure, the multi-valued image acquisition unit 10 acquires a digital multi-valued image to be subjected to a scaling process, and outputs a multi-valued image in a raster scanning format. The multi-valued image outline smoothing / magnifying section 11 includes magnification changing magnification instruction data 18 set by the magnification setting section 14.
The raster-type multi-valued image data 15 output from the multi-valued image acquisition unit 10 is input, and scaling processing is performed on the contour shape of the multi-valued image data 15. The density smoothing unit 12 receives the scaling instruction data 18 set by the scaling setting unit 14 and the multivalued image data 16 scaled by the multilevel image outline smoothing / scaling unit 11 and scales the data. Multi-valued image data 1
Pixel value interpolation is performed on 6 to smooth the density. Then, the multi-valued image output unit 13 is a printer, a display, or the like that displays the obtained scaled image, makes a hard copy, or outputs to a communication path.

FIG. 2 is a conceptual diagram showing the hardware configuration of the image processing apparatus according to the present invention. Multi-valued image input device 2
Reference numeral 1 corresponds to the multi-valued image acquisition unit 10 described above, which reads an image with an image reader and outputs it in a raster scanning format. However, it may be a device that captures data from a storage medium that stores a multi-valued image, a device that inputs an image captured by a still camera, or a device that inputs a video signal.

The multi-valued image output device 22 is a printer, a display, etc., and corresponds to the multi-valued image output unit 13. The storage device 23 stores a program for executing control of the entire image processing apparatus, image processing, image data, and the like. The display device 24 is a display or the like and displays the operation content. The operation input device 25 is a keyboard, a mouse, or the like. The central processing unit 26 controls the entire image processing apparatus and executes image processing. Image input device 2
The multi-valued image input from 1 or stored in the external storage device 27 is stored in the storage device 23. On the operation content display device 24, an instruction for scaling processing of the image input by the operation input device 25 is displayed, and in response to this, the central processing unit 26 executes the designated processing while accessing the storage device 23, It is output to the image output device 22 or the processing result is stored in the external storage device 27. In the latter case, the external storage device 27 is used as the multivalued image output unit 1.
It will also serve as 3.

Each block in FIG. 1 can also be called a program module executed by the central processing unit 26. By executing these program modules in the system of FIG. 2, the device of this embodiment is realized.
Each module is stored in the external storage device 27 and supplied to the system.

Next, FIG. 3 shows a detailed configuration diagram of the multi-valued image outline smoothing / magnifying unit 11 in FIG.

In the figure, the binary image creating section 31 is based on the multi-valued image data 15 output from the multi-valued image acquisition section 10 in FIG. To create. The binary image outline scaling unit 32 includes a raster scanning format binary image data group 34 created by the binary image creation unit 31 and the magnification setting unit 1 in FIG.
The scaling factor 18 set in 4 is input, the outline vector is extracted, and smoothing and scaling processing is performed in the extracted outline vector data form.
Represent the data from scaled outline data 2
The value image data group 35 is output. Multivalued image reproduction unit 33
Is a multi-valued image reproducing unit for reproducing the scaled multi-valued image 16 from the binary scaled image group 35 that has been scaled by the binary outline scaling unit 32.

FIG. 4 shows the flow of processing of the binary image creating section 31. For the digital multi-valued image data 15 input in the raster scanning format, a binary image for each density level is created by the processing flow of FIG. The input multi-valued image has m + 1 pixels in the main scanning direction (that is, 0 to m) and n + in the sub scanning direction.
One pixel (that is, 0 to n), and the pixel value at the main scanning direction coordinate x and the sub scanning direction coordinate y is described as f (x, y). Further, the number of gradations of the multi-valued image is set to the variable Nlev.
Represented by el. Further, the binary image at the density level indicated by the variable Ilevel is set as blevel (x, y).

First, in step S41, "1" is assigned as an initial value to the variable Ilevel for expressing the current gradation.

Next, in step S42, the input image f (x, y) is examined, and blev is determined according to whether it is greater than or less than the value held in the Ilevel variable.
Determine el (x, y).

Specifically, when f (x, y) ≧ Ilevel, blevel
When (x, y) ← “1” f (x, y) <Ilevel, blevel
(X, y) ← “0” This process is performed for all x, y (x = 0, 1, ..., M, y =
0, 1, ..., N) to obtain a binary image blevel indicated by Ilevel for the input image f.

When the generation of the binary image blevel indicated by Ilevel is completed in this way, the process proceeds to step S4.
3, the binary image blevel is output to the binary image outline scaling unit 32 which is the next processing unit in FIG.

When the process proceeds to step S44, the variable Ilevel is incremented by "1". Then, in step S45, the value of the variable Ilevel is 2
It is determined whether or not the value is equal to 56, and if it is determined that the value is still less than 256, the processing from step S42 is repeated.

The above processing will be described with reference to FIG. still,
Here, the original image has 4 gradations, that is, 0 as the density value,
The description will be given assuming that there are four types of density of 1, 2, and 3.

It is assumed that an original image as indicated by reference numeral 220 in the figure (numbers 0, 1, 2, 3 in the image respectively indicate density values) is input.

At this time, when the variable Ilevel = 1, the illustrated binary image 221 is generated and output to the binary image outline scaling section 32. When Ilevel = 2, the binary image 222 is I
When level = 3, the binary image 223 is generated.

In the above process, when the generation of the binary image for a certain density level is completed, it is passed to the lower process and the process for the next density level is executed. May be passed to a lower-level process after being generated. However, in the latter case, since a binary image for all density levels is generated at once, the memory capacity increases, but when the processing of the entire apparatus is performed by one CPU, switching of each task is performed. It is efficient because it reduces the overhead.

Further, as described above, when the generation of the binary image for a certain density level is passed to the lower processing, there is no problem if the CPU has a sufficient function as multitask. , Each task is a separate C
This is a particularly suitable process when it is executed by the PU.

Next, regarding the binary image outline scaling section 32 in FIG. 3, for example, the above-mentioned Japanese Patent Laid-Open No. 5-1741.
It can be configured by the device disclosed in the patent application filed by the applicant of the present application as No. 40. As shown in FIG. 6, the binary image outline scaling unit 32 includes an outline extraction unit 51, an outline smoothing / scaling unit 52, and a binary image reproduction unit 53. Specifically, the outline extraction unit 51 sequentially inputs the binary image group 34 created by the binary image creation unit 31, extracts the outline vector (coarse contour vector) for each density level, and outline smooths the outline vector. In the scaling section 52, an outline vector smoothly scaled at a desired scaling factor (arbitrary) designated by the scaling factor signal 18 is created in the form of the extracted outline vector expression. Then, the binary image reproducing unit 53 generates a raster-scanned binary scaled image group 35 indicated by each Ilevel from the outline vector for each smoothly scaled density level.

Next, the processing flow of the multi-valued image reproducing section 33 in FIG. 3 will be described with reference to FIG. Here, a binary image scaled image group B (main scanning direction size N, sub-scanning direction size M) scaled by the binary image outline scaling unit 32 is used to generate a multi-value scaled image F (main scaled image). The size N in the scanning direction and the size M in the sub-scanning direction are reproduced.

The binary image of each density level generated by smoothing and scaling processing is represented as Blevel,
The pixel value of the coordinates x and y is Blevel (x, y) (1 pixel 1 bit). In addition, the target final image is the image F
Then, the pixel value of the coordinates x and y is represented as F (x, y) (1 pixel 8 bits).

First, in step S61, the variable Il
Substitute “1” as the initial value for the level and all output image data F (x, y) (x = 0, 1, ..., M, y =
0, 1, ..., N) are cleared to “0”.

Next, in step S62, the vector is smoothed, and the variable Ile generated by the smoothing is performed.
A binary image Blevel indicated by vel is input.

In this step S64, the input Ble
When vel (x, y) is “1”, the output image F (x,
The value of the variable Ilevel at that time is substituted into y). When Blevel (x, y) is “0”,
Do nothing, that is, do not change the value of F (x, y). This process is performed for all x and y.

When the process proceeds to step S65, the variable Il
The level is incremented by "1" to prepare for the input of image data of the next density level.

Then, in step S66, the variable Ilev
The processes in and after step S62 are repeated until it is determined that the value of el has become “256”.

As a result of the above processing, each pixel in the output image F is stored with a value of 0 to 255 according to the binary image of each density level, and as a result, the halftone after the scaling is changed. The image will be stored.

The multi-valued image outline smoothing / magnifying section 11 in FIG. 1 has been described above.

Next, FIG. 7 shows the structure of the density smoothing unit 12 shown in FIG.

The density smoothing unit 12 comprises a uniform weight filter processing unit 71 and a filter size determining unit 72. The filter size determination unit 72 is the magnification setting unit 14 in FIG.
Enter the magnification 18 obtained from the filter size 73
Is output. The uniform weight filter processing unit 71 outputs the output 16 of the multi-valued image outline smoothing / magnifying unit 11 in FIG.
Then, the filter size 73, which is the output from the filter size determination unit 72, is input, and finally the multivalued image 17 subjected to the pixel value interpolation processing is output.

The uniform weight filter processing unit 71 performs the filter processing normally used for noise removal, etc., by a known method introduced in the above-mentioned "Introduction to Computer Image Processing" Tamura, Soken Shuppan, etc. . FIG. 9 shows an outline of this processing. In FIG. 9, the pixel 81
Indicates a pixel in a multi-valued image during raster scanning, and an area 82 represents a 25 pixel area (also referred to as a 5 × 5 area) including 24 pixels in the vicinity of this pixel 81.

In the filtering process, the arithmetic mean value is calculated by multiplying the value (density) of the pixel of interest 81 and the density of the pixel values in the vicinity thereof by a certain weight, and the calculated value is taken as the new pixel value of the pixel of interest. Processing. Further, in the uniform weight filter processing unit 71 in the present embodiment, the weight values are all processed as "1" as shown in FIG.

Although the image data F is obtained by the processing by the multi-valued image outline smoothing / magnifying unit 11, the pixel value of interest in the input image is also set to F (x, y) and the pixel of interest after the filtering process is performed. Expressing the value as D (x, y),
D (x, y) is represented by the following equation. D (x, y) = ΣF (i, j) / K (2) where x−2 ≦ i ≦ x + 2 y−2 ≦ j ≦ y + 2 K = sum of weight values in the filter.

The filter size determination unit 72 determines the size of the filter used by the uniform weight filter processing unit 71, which is determined from the scaling factor 18 set by the scaling factor setting unit 14 in FIG.

Specifically, the magnification in the main scanning direction is set to G, the magnification in the sub scanning direction is set to H, and the size of the filter (the area near the pixel of interest) is set in the main scanning direction. g,
When g and h are rectangular in the sub-scanning direction h, g = min {[G], even ([G + 1])} h = min {[H], even ([H + 1])}. However, even (x) is a function that returns an odd number that is not less than x, [...] is a Gaussian symbol, min {}
Represents a function that returns the minimum value in ｛｝.

Next, FIG. 10 shows the principle by which the uniform weight filter processing unit 71 can perform pixel value interpolation. In FIG. 10, a one-dimensional digital signal is typically used for description. Considering that the input digital signal 91 is magnified five times in the coordinate direction (G = 5 in equation (2)), the input digital signal 91 is transformed by the multilevel outline smoothing / magnifying unit 11 as indicated by reference numeral 92. Doubled. Next, the result of performing the uniform weight filtering process in the density smoothing unit 12 (g = 5 in the filter size determination unit) is the density distribution indicated by reference numeral 93, and the amplitude value ( It can be confirmed that pixel values are interpolated in the case of images.

As described above, the density smoothing unit 12 processes the scaling result 16 in which the contour shape has been smoothly scaled by the multi-valued outline smoothing / scaling process 11 in FIG. It is shown that the pixel value can be interpolated.

As described above, the multi-valued image scaling unit 11 and the density smoothing unit 1 are used as the multi-valued image scaling process.
By performing the processing according to 2, it is possible to obtain a multi-value scaled image in which jaggies, lattice-like distortion, and the like do not occur in each part of the image ring during multi-scale image scaling.

<Modification of the First Embodiment> In the above configuration, the uniform weight filter processing unit 71 of FIG. 8 determines by the filter size determination unit 72 according to the scaling factor 18 set by the scaling factor setting unit 14. Although it is configured to use the filter having the size described above, it is not always necessary to configure in this way. That is, the same effect as that of the filter of the size obtained by the equation (3) can be obtained by repeatedly performing the filtering process on the basic small filter size a plurality of times.

For example, when the magnification of variable magnification is set to 5 times in the main scanning direction and the sub scanning direction, 5 × 5 filter processing is performed as 1.
The result obtained by performing the filtering process and the case where the filtering process is performed twice by setting the size of the basic filter to 3 × 3 are approximately the same. As a result, the working memory capacity can be reduced and the filter processing time can be shortened.

Further, the shape of the filter used in the uniform weight filter processing section of FIG. 8 is a square (rectangular) shape, and the weighting coefficient of the filter is all 1 and is uniform, but it is not always so. No need to configure. That is, as shown in FIG. 11, a non-square (non-rectangular) shape may be used, and the weighting coefficient for the pixel of interest 211 is the largest, and the weighting coefficient may become smaller as the distance from the pixel increases. . As a result, pixel value interpolation that preserves the edge components in the multi-valued image becomes possible.

In this case, the above equation (2) is given by: D (x, y) = ΣkijF (i, j) / K where kij is the weighting coefficient at the position (i, j) in the filter,
K = Σkij.

Further, in the above apparatus, as shown in the flow of processing in FIG. 4, a binary image is created for every density level,
Although the processing is performed, it is not always necessary to perform the processing for a multi-valued image having a small number of gradations such as a multi-valued document image instead of a photographic image. That is, instead of incrementing the variable Ilevel by “1” in step S44 of FIG. 4, the processing may be performed for each c gradation with a value c larger than that. Along with this, step S65 in FIG. 7 is also changed to Ilevel = Ilevel + c. As a result, it is possible to shorten the processing time of the multi-valued image outline scaling processing.

Specifically, when the apparatus of the embodiment is applied to, for example, a copying machine, a key for selecting a text image or a photographic image as the type of document image is provided, and the gradation interval c is set by the key. Set. Alternatively, the gradation interval c may be freely set to suit the user's preference.

Of course, it can be easily conceived from the above description that the present invention can be applied not only to the copying machine but also to any apparatus.

In the present embodiment, for example, the portions corresponding to the respective processing portions in FIG. 1 have been described as being realized by the central processing unit executing the program, but each processing portion is constituted by an independent circuit. You may.

FIG. 12 is a flow chart of a program executed by the central processing unit 26 of FIG. 2 in order to realize the configuration shown in FIG.

In the multivalued image input step S121, the multivalued image is read from the input device 21, and in the binary image creation step S122, the binary image corresponding to each density level of the multivalued image is processed by the procedure shown in FIG. To create.

In step S123, the method disclosed in
According to the procedure disclosed in Japanese Patent No. 174140, the binary image at each density level is scaled and smoothed.

In step S124, step S
Based on the binary image smoothed and scaled in 1234, the multi-valued image is reproduced by the procedure shown in FIG.

In step S125, the filter shown in FIG. 9 is used to perform the filtering process according to the equation (2) to smooth the density.

Finally, in step S126, the output device 2
The image reproduced from 2 is output.

The image processing of this embodiment can be realized by executing the program as described above.

As described above, the image processing apparatus according to the present embodiment performs appropriate scaling processing on the contour shape of a multivalued input image to scale the image, and the scaled multivalued image. By performing pixel value interpolation on an image, it is possible to obtain a scaled image with good image quality. That is, a multivalued image outline scaling unit that performs scaling processing on the contour shape of a multivalued image, and a density smoothing unit that performs pixel value interpolation on the multivalued image scaled by the multivalued image outline scaling unit. By providing, it is possible to obtain an effect that it is possible to obtain a good multi-value scaled image in which image quality deterioration such as jaggies and lattice distortion does not occur.

[Second Embodiment] A second preferred embodiment of the present invention will be described below with reference to the drawings. FIG. 13 is a diagram showing a flow of outline smoothing / magnifying processing of a multivalued image according to the present embodiment. The multi-valued image acquisition unit 120 acquires a digital multi-valued image to be subjected to smoothing / magnification processing and outputs it via multi-valued image data 125 in a raster scanning format. The binary image creating unit 121 creates binary image data in the raster scanning format for each gradation included in the multivalued image data 125 in the raster scanning format, and outputs a binary image data group 126 that is a set thereof. The binary image outline smoothing / scaling unit 122 extracts an outline vector from the binary image data group 126 in the raster scanning format, performs smoothing and scaling processing in the extracted outline vector data form, and performs smoothing / scaling processing. From the outline data, a binary image represented by the data is reproduced as binary image data in a raster format to obtain a scaled image. Then, a binary scaled image group 127 that is a set of scaled images corresponding to each density level is output. The multi-valued image reproducing unit 123 reproduces the multi-valued image 128 that has been scaled from the binary scaled image group 127. The multi-valued image output unit 124 displays the obtained scaled image, forms a visible image, or transmits the image to an external device via a communication path.

FIG. 2 shows an example of the hardware configuration of the outline smoothing / magnifying apparatus for multivalued images.

The multi-valued image input from the image input device 21 or stored in the external storage device 27 is read into the storage device 23 (work memory) by the central processing unit 26 as needed. The operation content display device 24 displays an instruction (for example, a start instruction for smoothing / magnifying processing) input by the operation input device 25. The central processing unit 26
The instruction input through the operation input device 25 is recognized, the specified process is executed while appropriately accessing the storage device 23 according to the instruction, and the result is output to the image output device 22 or the external storage device 27. The processing result is stored in.

The blocks shown in FIG. 13 can be said to be program modules executed by the system shown in FIG. Each module is stored in the external storage device 27, supplied to the system, and executed by the central processing unit 26.

FIG. 14 is a flow chart showing the flow of processing of the binary image creating section 121. Binary image creation unit 121
Generates a binary image for each density level from the digital multi-valued image data 125 input in the raster scanning format and outputs a binary image data group 126.

First, in step S131, the multi-valued image data 125 is input. Here, the size of the multi-valued image of the multi-valued image data 125 is the size n in the main scanning direction and the size m in the sub-scanning direction, and the pixel values at the coordinates x in the main scanning direction and the coordinates y in the sub-scanning direction are f (x, y). Define as. Further, the number of gradations of the multi-valued image is L (for example, in the case of a multi-valued image of 8 bits for each pixel, the number of gradations L = 256).

In step S132, the variable Ilevel indicating the gradation to be processed is initialized to "1". In step S133, f (x, y) ≧ Ileave
1 (black pixel) if f, and f (x, y) <Ile
Binary image b, which is 0 (white pixel) if vel
(X, y, level) (main scanning direction size n, sub scanning direction size m) is created. In step S133, Ile
The value of vel is determined, and if Ilevel is (L-1) or less, the process proceeds to step S135, 1 is added to Ilevel, and the process returns to step S133. On the other hand, if Ilevel is larger than (L-1), it is determined that the creation of the binary image corresponding to all the density levels has been completed, and step S136 is performed.
And outputs the binary image data group 126. FIG. 15 is a block diagram showing the flow of processing of the binary image outline smoothing / magnifying unit 122. The binary image outline smoothing / magnifying unit 122 is, for example, the above-mentioned Japanese Patent Laid-Open No. 1741
This can be realized by applying the technology of No. 40, and the outline extraction unit 141 and the outline smoothing / magnifying unit 1
42 and a binary image reproducing unit 143. Specifically, the outline extraction unit 141 uses the binary image creation unit 12
Binary image data group 126 (b (x,
An outline vector (coarse contour vector) is extracted from each binary image data of (y, Ilevel). The outline smoothing / magnifying unit 142 creates an outline vector smoothly scaled by a desired magnification in the state of the extracted outline vector expression, and the binary image reproducing unit 14
At 3, a binary image is regenerated from the smoothly scaled outline vector. As a result, a high-quality binary image that is scaled by a desired magnification is generated. By this processing, b created by the binary image creation unit 121
A binary variable value image group 127 obtained by smoothing and scaling each binary image of the binary image group 126 given by (x, y, Ilevel) is output. In addition, b (x, y, Ilevel) after smoothing and scaling is B
(X, Y, Ilevel) (main scanning direction size N, sub-scanning direction size M).

FIG. 16 is a flow chart showing the flow of processing of the multi-valued image reproducing section 124. In this processing, a binary scaled image group B (X, Y, Ilevel) is used to generate a multi-scale scaled image F.
(Size N in the main scanning direction, size M in the sub scanning direction) is reproduced.

First, in step S151, B
Binary scaled image group 127 given by (X, Y, Ilevel)
Read in. In the next step S52, the multi-value scaled image F (x, y) is set to "0", and the variable Ilev representing the density level is set.
Initialize el to "1". Next, in step S53, it is determined whether or not all B (X, Y, Ilevel) are white pixels, that is, there is no pixel having a density of Ilevel, and if all are white pixels, the process proceeds to step S155. On the other hand, when there are black pixels, the density is Ilev.
Since there is a pixel of el, step S154
Proceed to. In step S154, B (X, Y, Ilevel)
If = 1 (black pixel), the value of Ilevel is set to the pixel value F (X, Y) of the multi-value scaled image. On the other hand, B (X, Y,
If Ilevel) = 0 (white pixel), do nothing (pixel value F
(The values of (X, Y) are unchanged). In step S155, the value of Ilevel is determined, and when the value of Ilevel is (L-1) or less, that is, when processing is not performed for all gradations, the process proceeds to step S156, and Ile
1 is added to vel and the process returns to step S153 to repeat the process. On the other hand, if the processing has been completed for all the density levels, the process advances to step S157 to output the scaled multi-valued image data 128 given by F (X, Y).

In the present embodiment, when creating a binary image from a multivalued image and when reproducing a multivalued image from the binary image, a variable (Ileve) corresponding to each density level is generated.
Although l) is used as a standard, other standards may be used.

As described above, the multi-valued image is decomposed into binary images for each density level, the outline vector is extracted from each decomposed binary image, smoothed and scaled, and then processed again. By reproducing the multi-valued image from the subsequent binary image, it is possible to obtain a multi-valued scaled image in which jaggies and lattice distortion in the contour portion of the image at the time of scaling the multi-valued image are suppressed. .

Further, in the present embodiment, it is checked for each density level whether or not there is a pixel of that level, and if there is no such density level, the density level is skipped. Therefore, the process of reproducing a multi-valued image is quick. become.

[Modification of the Second Embodiment] In the second embodiment, binary images corresponding to all the density levels are created, and outline vector extraction, smoothing / magnifying, and multivalue image reproduction processing are performed. However, this is not always necessary for a multi-valued image having a small number of gradations such as a multi-valued document image. That is, "Ilevel = Ile" in step S135 (binary image creation process shown in FIG. 14).
vel + 1 ”becomes“ Ilevel = Ilevel + c (c
May be performed for every c gradation. In this case, also in step S156 (multivalue image reproduction processing shown in FIG. 16), “Ilevel = Ilevel +”.
Needless to say, change to "c".

If the gradations of the input multi-valued image are concentrated in a certain range of gradations, only that range may be processed. That is, even for an input image having 256 gradations (0 to 255 levels), if the gradations of the pixels forming the input image are concentrated in 50 to 100 levels, only the 50 to 100 levels are processed. May be.

As described above, when the input image is a multi-valued image with a small number of gradations, or when deterioration of the image quality can be tolerated, the time required for smoothing / magnifying processing is reduced by thinning the gradations to be processed. Can be shortened.

In the present embodiment, the binary image creating unit 121 creates binary images corresponding to all density levels, and
The configuration has been shown in which the value image outline smoothing / magnifying unit 122 performs binary outline scaling processing on all binary images, and the multivalued image reproducing unit 123 reproduces the binary outline scaling processing to obtain a multivalued scaling image. However, in this case, since it is necessary to temporarily store all the data in the storage device 23 or the external storage device 27 for each process, a huge storage capacity is required.

In this modification, the above processing procedure is modified as follows in order to solve this problem. That is, for each density level, binary image creation, outline smoothing / magnification, and multivalued image reproduction processing are executed, and the density level of a part of the multivalued image is reproduced each time. Then, this series of processing is repeatedly executed for all the density levels (for example, after the multi-value image reproduction processing, it is determined whether or not the processing has been completed for all the density levels, and when there is an unprocessed density level, Repeats a series of processes from the binary image creating process) to reproduce a complete multi-valued image. As a result, the capacity of the storage device required for processing can be reduced to the capacity corresponding to a binary image for one gradation.

The apparatus of the second embodiment (FIG. 13) can be realized by executing the program of the procedure of FIG. 12 in the first embodiment by the central processing unit 26 in the configuration of FIG. However, the present embodiment is different from the first embodiment in that the process of the procedure of FIG. 14 is performed as step S122, and the process of the procedure of FIG. 16 is performed as step S124. [Third Embodiment] In this embodiment, an input multi-valued image is separated into an image subjected to outline smoothing / magnification processing and an image subjected to magnification / magnification processing by pixel value interpolation, and the respective magnification / magnification processing is performed. After that, the scaled images are combined and generated.

FIG. 17 is a diagram showing a flow of outline smoothing / magnifying processing of a multivalued image according to this embodiment. FIG.
In the multivalued image acquisition unit 178, the multivalued image output unit 17
Reference numeral 9 denotes a multi-valued image acquisition unit 12 in the second embodiment, respectively.
0, has the same function as the multi-valued image output unit 124. The multi-valued image outline scaling unit 172 has the same function as the multi-valued image smoothing / scaling unit 1200 in the second embodiment. The pixel value interpolation unit 171 scales while interpolating the pixel value. The multi-valued image separation unit 170 separates the input image, which is a scaling object, into multi-valued images 174 and 176, and outputs the multi-valued images 174 and 176. The multi-valued image integration unit 173 uses the pixel value interpolation unit 17
The multi-valued scaled image 175 interpolated by 1 and scaled, and the multi-scaled scaled image 177 scaled by the multi-level image smoothing / scaling unit 172.
To obtain the final multi-value scaled image.

FIG. 18 is a block diagram showing an example of the structure of the multi-valued image separation section 170. Multivalued image data holding unit 18
1 temporarily holds the multi-valued image data supplied from the multi-valued image acquisition unit 178. Smoothing / magnifying image creating unit 18
2 is a multi-valued image supplied to the multi-valued image smoothing / magnifying unit 172 that executes outline smoothing / magnifying processing of the multi-valued image,
It is created with reference to the data stored in the multi-valued image data holding unit 181, and stored in the smoothing / magnifying image data holding unit 184. The proportional distribution setting unit 185 sets conditions for separating multi-valued images. Pixel value interpolation image creating unit 183
Creates a pixel value interpolation image to be supplied to the pixel value interpolation unit 171 by referring to the data stored in the multivalued image data holding unit 181 and the smoothing / magnification image data holding unit 184.

In the smoothing / magnifying image creating section 182, the input image data f (x, y) (main scanning direction size n, sub scanning direction size m) held in the multi-valued image data holding section 181 is The variable s indicating the proportion of proportional distribution set by the proportional distribution setting unit 185 is referred to, and the smoothing / magnifying image f1 (x, y) is created according to Expression (4).

F1 (x, y) = f (x, y) >> s Equation (4) (0 ≦ x ≦ n−1, 0 ≦ y ≦ m−1) where >> s is s Indicates a logical operation that shifts to the right by one bit.

When the input multi-valued image is composed of S bits for one pixel, the proportional distribution setting unit 185 sets s that does not exceed S. The smooth / variable-magnification image data created by the smooth / variable-magnification image creating unit 182 is temporarily held in the smooth / variable-magnification image data holding unit 184.

The image f1 is an image in which the number of gradations of the image f is represented by 1 / s to the power of 2. For example, S = 8, s =
If 4, the image f has 256 gradations. An image represented by 16 gradations is f1 by dividing this into 1/4 of 2, that is, 1/16.

In the pixel value interpolation image forming section 183, the input image data f (x, y) held in the multivalued image data holding section 181 and the smoothing / magnifying image data holding section 184 are held.
Smoothing / magnifying image data f1 (x,
y), the pixel value interpolation multivalued image data f2 is created according to the equation (5).

F2 (x, y) = f (x, y) −f1 (x, y) Equation (5) (0 ≦ x ≦ n−1, 0 ≦ y ≦ m−1) smoothing / variation The double image data f1 (x, y) and the pixel value interpolation image data f2 (x, y) are used as input images of the multivalued image smoothing / magnifying unit 172 and the pixel value interpolating unit 171 respectively.

It should be noted that the separation of the input image is not limited to the above method, and it is obvious that the separation may be performed by performing another calculation or the like.

The pixel value interpolation unit 171 is a multivalued image separation unit 1
70 pixel value interpolation image data 174 is input, and a desired magnification (the same magnification as the multivalued image smoothing / magnifying unit 172)
Then, the scaling processing is performed and the pixel value interpolation image 175 is output. Pixel value interpolation is basically a method of interpolating pixel values between pixels of a sampled input multi-valued image using pixel values in the vicinity thereof. For example, “Introduction to Computer Image Processing” Tamura, Soken There is a method of interpolating with a bilinear linear function introduced in Publication, etc., and a method of approximating a sampling function with a cubic expression to interpolate.

The multi-valued image smoothing / magnifying unit 172 multiplies the multi-valued image smoothing / magnifying image data 176 supplied from the multi-valued image separating unit 170 with a desired magnification (the same magnification as the pixel value interpolating unit 171). Scale with. Multi-valued image smoothing / magnifying unit 1
Reference numeral 72 denotes the binary image creating unit 121 and the binary image outline smoothing / magnifying unit 1 in the second embodiment, as described above.
22 and a multi-valued image reproduction unit 123, which performs a predetermined process on the multi-valued image smoothing / magnifying image data 176 and outputs a smoothing / magnifying image 177.

That is, in the multi-valued image smoothing / magnifying unit 172, the smoothing / magnifying image in which the number of gradations of the original image f is reduced to 1 / s of 2 is the same as in the second embodiment. Apply processing.

Further, the image value interpolating unit 171 performs density interpolation on the image f2 obtained by removing the value used as the smoothing / magnifying image from the original image f. The image f2 is (f-f1)
Since it is given by, the characteristics of the original image are well retained.
Therefore, the interpolated image is also faithful to the original image, although the density of the high-density portion of the original image is reduced by the amount of the image f1.

Note that each unit in the block diagrams of FIGS. 16 and 17 executes the program stored in the storage device 23 or the external storage device 27 by the central processing unit 26 of FIG. 2 as in the first embodiment. Will be realized in. In this case, the multi-valued image acquisition unit, the pixel value interpolation unit, the multi-valued image smoothing /
The variable power unit and the multi-valued image output unit are the same as those in the first embodiment.

The blocks shown in FIGS. 16 and 17 are the same as those shown in FIG.
It can also be said to be a program module executed by the central processing unit 26 of the above. In that case, each module is stored in the external storage device 27 and supplied to the system.

The multivalued image integration unit 173 adds the pixel value interpolation image 175 and the smoothed / scaled image 177 to reproduce the scaled image. Since f = f1 + f2 from the equation (5),
The reproduced image is an image faithful to the input image 180.

As described above, the input multi-valued image is separately supplied to the pixel value interpolating unit 171 and the smoothing / magnifying / scaling unit 172, scaled by different methods, and then integrated to obtain a high magnification. Even when the image is enlarged by, the contour portion is not overemphasized, and it is possible to obtain a scaled image that suppresses the occurrence of jaggies and grid-like distortion that become noticeable when the scaling process by pixel value interpolation is performed. it can. The output result is an intermediate scaled image between the case where only the outline smoothing / magnification processing according to the second embodiment is applied and the case where only the pixel value interpolation is applied.

Furthermore, since the number of gradations of the image to be subjected to the smoothing / magnifying processing is 1 / s to the sth power of 2, the amount of the binary image outline smoothing / magnifying processing is also the first and second embodiments. 2 compared to
It is 1 / S.

Therefore, by changing the distribution ratio, the processing time can be shortened as compared with the second embodiment.

19 and 20 are flow charts of a program executed by the central processing unit 26 in the system shown in FIG. 2 to realize the configuration of the third embodiment.

When the multi-valued image is input in step S191, it is divided into two as described in step S192. Steps S193 to S195 are the same as steps S122 to S124 of FIG. 12 in the second embodiment. However, the image to be processed is shown in FIG.
It is the image obtained in steps S201 and S202 of the procedure shown in FIG.

Step S198 is the same process as step S125 of FIG. 12 in the second embodiment. However, the image to be processed is the step S2 in FIG.
03, the image obtained in S204.

Note that once the images have been separated, the processing of steps S193 to S195 and the processing of step S198 can be performed regardless of their order. Therefore, step S198 may be between steps S192 and S193 or between steps S195 and S196. Also, both may be executed in parallel.

Corresponding pixels of the two images thus obtained are added in step S197.

By doing so, the apparatus of FIG. 13 can be realized.

According to the present embodiment, it is possible to obtain a good multi-value scaled image in which deterioration of image quality such as jaggies, moire or lattice distortion is suppressed.

[Fourth Embodiment] FIG. 21 best shows the structure of the present embodiment. Multi-valued image acquisition unit 211
0 acquires a digital multi-valued image to which the scaling process is applied, and outputs a raster-scanning-type multi-valued image. The uniform density line smoothing / magnifying unit 2111 inputs the magnification 2120 of the magnification set by the magnification setting unit 2115 and the raster scanning multi-valued image data 2116 output from the multi-valued image acquisition unit 2110. , Outline vector data of the contour line shape of the multi-valued image data 2116 (hereinafter referred to as the contour line vector)
The smoothing and scaling processing is performed on. Multi-valued image reproduction unit 211
2 is the same density line vector data 2117 that has been smoothed and scaled by the constant density line smoothing / magnifying unit 2111, and a raster scanning multi-value scaled image 2118 is reproduced at high speed from the form of the constant density line vector data. To do. The density smoothing unit 2113 inputs the raster scanning multi-valued image data 2118 reproduced by the multi-valued image reproducing unit 2112 and the scaling factor 2120 set by the scaling setting unit 2115, and scales the multi-valued image. Density smoothing is performed on the image data 2118. Then, the multi-valued image output unit 2114 displays the obtained multi-valued image, makes a hard copy, or outputs it to a communication path or the like.

FIG. 2 is a conceptual diagram showing the hardware configuration of the image processing system according to the embodiment of the present invention. Each part is as described in the first embodiment. <Smoothing / Scaling of Isodensity Line> Next, FIG. 22 shows a detailed configuration diagram of the smoothing / smoothing unit 2111 of the isodensity line in FIG. The binary image creation unit 2201 has the multi-valued image data 211 output from the multi-valued image acquisition unit 2110 in FIG.
A binary image data group 2204 in raster scanning format is created from 6 for each density level. Binary image outline extraction unit 2
In all of the raster scan type binary image data groups 2204 created by the binary image creating unit 2201, 202 extracts vector data of the contour shape and creates a binary image outline vector data group 2205. The outline smoothing / magnifying unit 2203 includes a binary image outline vector data group 2205 extracted by the binary image outline extracting unit 2202 and a magnification setting unit 2115 in FIG.
The input scaling factor 2120 set in step 2 is used to perform smoothing / scaling processing on the binary image outline vector data group 2205 in the vector data form, and output a smoothing / scaling binary image outline vector data group 2206. This iso-concentration line smoothing / scaling unit 2111 creates a binary image for each gradation, extracts the contour shape thereof, and smoothes / scales the extracted outline vector. This is equivalent to extracting vector data of isoconcentration lines and performing smoothing and scaling.

FIG. 23 shows the processing flow of the binary image creating section 2201. With respect to the digital multi-valued image data 2116 input in the raster scan format, a binary image for each density level is created by the processing flow of FIG. The input multi-valued image is defined as f (main scanning direction size n, sub-scanning direction size m), and the pixel value at the main scanning direction coordinate x and the sub-scanning direction coordinate y is described as f (x, y). To do. Further, the number of gradations of the multi-valued image is L (if the pixel is an 8-bit multi-valued image, the number of gradations L = 256).

In step S231, the variable Ilevel for expressing the current density level is initialized to 1. In step 42, a binary image blevel (a size in the main scanning direction) is set to 1 (black pixel) if f (x, y) ≧ Ilevel and 0 (white pixel) if f (x, y) <Ilevel. n, sub-scanning direction size m) is created.
If the variable Ilevel is (L-1) or less in step S233, the process proceeds to step S234, 1 is added to the variable Ilevel, and the process proceeds to step S232. If the variable Ilevel is larger than (L-1), this process ends. By this processing flow, the binary image group 2204 of each density level in FIG. 22 is output.

Next, the binary image outline extracting section 2202 and outline smoothing / magnifying section 2203 in FIG.
However, for example, it can be constituted by the device disclosed in the patent application filed by the present applicant as Japanese Patent Laid-Open No. 5-174140. Specifically, in the binary image outline extraction unit 2202, the binary image group 2204 created by the binary image creation unit 2201 is input, and the binary image outline vector data group (coarse contour vector), that is, the multivalued image. Is extracted in the outline smoothing / magnification unit 2203, and the extracted isoconcentration line vector data 2205 in the outline smoothing / magnification unit 2203 is expressed in outline vector data. ) To create an outline vector that has been smoothly scaled by
206 is output. <Multi-Valued Image Reproduction> Next, the multi-valued image reproduction unit 2112 in FIG. 21 will be described. This can be configured, for example, by using the device described in Japanese Patent Application Laid-Open No. 5-204467 previously filed by the applicant of the present application. The flow of this processing is shown in FIG.

First, in step S241, the variable Ilevel representing the gradation is set to 1, and the binary image buffer B and the multivalued image buffer G (main scanning size X, sub-scanning size Y) are initialized. In step S243, the binary image buffer B
Draw an iso-concentration line whose density level is Ilevel. In step S244, the drawn isodensity line is used to reproduce the multivalued image. That is, the binary image buffer B is raster-scanned, and when the number of intersections with the contour lines drawn there is an odd number, the multivalued image buffer G is changed from the pixel having the same coordinates as the intersections. The filling is started with the gradation given by Ilevel, and the multi-valued image buffer G is painted in the same order as the scanning of the binary image buffer B. When the number of intersections is an even number, the coordinates of the intersections are set to the end of the multi-valued image buffer. Finish filling G. This process is performed for L 2
As for the value image, the original multivalued image is reproduced in the multivalued image buffer G.

Further, when raster-scanning the binary image buffer B, the pixels (black pixels) intersecting the isodensity line are changed to white pixels. This means that the binary image buffer is initialized while raster scanning, and it is not necessary to initialize the binary image buffer B each time a certain density level is processed.

This principle diagram is shown in FIG. First, the isodensity line 253 is drawn in the binary image buffer B. Then 2
When the raster scanning of the value image buffer B is started and the first (odd) crossing of the raster scanning and the contour (pixel 254)
And a multi-valued buffer G with the same coordinates as the binary image buffer B
In, the filling is started, and the filling is finished at an even number of times, that is, at the next intersection (pixel 256).

As described above, since only the iso-concentration lines of a certain gradation are drawn on the binary image and the multi-valued image is reproduced by using this, it is drawn on the binary image. It is possible to reproduce the multi-valued image at a higher speed than when the multi-valued image is reproduced after the entire density line is completely filled. <Smoothing of Density> Next, the density smoothing unit 21 in FIG.
FIG. 26 shows the configuration of No. 13. The density smoothing unit 2113 includes a uniform weight filter processing unit 261 and a filter size determination unit 2
It is composed of 62. The filter size determining means is shown in FIG.
Magnification ratio 2 obtained from the magnification setting unit 2115 in
120 is input and the filter size 263 is output.
The uniform weight filter processing unit 261 inputs the output 2118 of the multi-valued image reproduction unit 2112 and the filter size 263 which is the output from the filter size determination unit 262 in FIG. 21, and finally multi-values the density smoothing process. Image 211
9 is output. The uniform weight filter processing unit 261 is a known method introduced in, for example, "Introduction to Computer Image Processing" (Tamura, Soken Publishing Co., Ltd.), and is usually used for noise removal. This processing is the same as in the first embodiment, and is performed using the filter shown in FIG.

Next, the filter size determination unit 262 can determine the size of the filter used by the uniform weight filter processing unit 261 from the scaling factor 2120 set by the scaling factor setting unit 2115 in FIG. Now, let G be the scaling factor in the main scanning direction and H be the scaling factor in the sub-scanning direction.
When the size of the filter (area near the pixel of interest) is rectangular (square) in the main scanning direction g and the sub-scanning direction h, g = min {[G], even ([G + 1]) } (5) h = min {[H], even ([H + 1])} is given. However, even (x) is a function that returns an odd number that is not less than x, [•] is a Gaussian symbol, and min {}
Represents a function that returns the minimum value in {}.

Next, the principle that the uniform weight processing unit 261 can perform density smoothing is as shown in FIG. 10, as in the first embodiment.

As described above, the multi-valued image reproducing section 21 in FIG.
It was shown that the density change can be smoothly smoothed by applying the density smoothing unit 2113 to the reproduced multi-valued image 2118 in which the equal density lines are smoothed / scaled and reproduced in 12.

As described above, as the multi-valued image scaling process, the iso-density line smoothing / magnifying process (operation of the iso-density line smoothing / magnifying unit 2111 in FIG. 21) and the density smoothing process (density in FIG. 21) are performed. By incorporating the smoothing unit 2113), it is possible to obtain a multivalued image magnified image in which no jaggies or grid-like distortions occur at the edge portion in the image during multivalued image scaling. Further, in the multi-valued image reproducing means, only the equi-density of a certain density level is drawn from the iso-density line vector data on the binary image, and the multi-valued image is reproduced by using this, and
The multi-valued image can be reproduced at a higher speed than the case where the multi-valued image is reproduced after all the constant density lines drawn in the binary image have been filled.

In order to realize the configuration of FIG. 21 by a program, in the flowchart of FIG.
The processing of the procedure of FIG. 23 is executed as step S122, and the processing of FIG. 24 is executed as step S124. This program is supplied to the system using the external storage device 27 as a medium.

In the present embodiment, the binary image forming unit 2201 in FIG. 22, that is, the processing flow shown in FIG. 23, is configured to create a binary image for every gradation and perform the processing. The multi-valued image having a small number of gradations such as a multi-valued document image does not necessarily have to be configured in this way. That is,
Ilevel = Ilevel in step S234 in FIG.
Ilevel = Ilevel + c and c instead of l + 1
You may perform a process for every gradation. As a result of this, FIG.
It is possible to shorten the processing time of the iso-concentration line smoothing / magnifying unit 2111 and the multi-valued image reproducing unit 2112 in 1.

Further, in this embodiment, the binary image creating section 2201 of FIG. 22 creates binary images of all density levels,
The binary image outline extraction unit 2202 extracts the outline vector from the binary image of all the density levels, and the outline smoothing / magnifying unit 2203 is configured to smooth / magnify the outline vector of all the density levels. There is no need to make such a flow. That is, the storage device 23 in FIG. 2 is used to execute binary image creation, binary image outline extraction, outline smoothing / magnification, gradation by gradation, and reproduce into a multivalued image (multivalued image reproduction in FIG. 21. Part 2
The series of processes 112) may be repeated for the number of gradations.
By this processing, the access time to the storage device 23 increases and the overall processing time also increases.
The binary image memory used in the outline vector extraction means is required for all the gradations, but only one surface (one gradation) is required, so that the memory capacity can be considerably reduced.

[Modification of Fourth Embodiment] In all of the present embodiments, as shown in FIGS. 21 and 22, a multi-valued image acquisition unit, a binary image creation unit, and a binary image outline extraction unit are included. However, it is not always possible to configure without them. That is, it can be implemented by providing means for inputting isodensity line vector data created for each gradation of a multi-valued image from the outside. Figure 2 when doing so
27 shows the configuration corresponding to No. 1, FIG. 28 is a conceptual diagram of the hardware configuration corresponding to FIG. 2, and FIG. 29 is the configuration corresponding to FIG.

As shown in the hardware conceptual diagram of FIG. 28, instead of the multi-valued image input device 21, the iso-concentration line vector data extracted by the iso-concentration line extraction means in the external multi-valued image is transmitted via the communication path. By providing the input device 281 for inputting the data, it is possible to perform the scaling process on the isoconcentration line vector data from the outside. Further, it is also possible to implement so as to input the isoconcentration line vector data of the multivalued image stored in the storage device 23. In this case as well, after the input isoconcentration line vector data smoothing / magnifying processing is executed in the same manner as in FIG. 27, the same processing flow as that of FIGS. 21 and 22 in the fourth embodiment can be configured. is there.

As described above, the image processing apparatus and method according to the present embodiment make it possible to obtain a scaled image of good image quality, and image quality deterioration such as jaggies and lattice distortion does not occur. This has an effect that a large multi-value scaled image can be obtained.

Further, since the multi-valued image is reproduced by using the iso-dense lines of a certain gradation, the multi-valued image is reproduced after all the iso-dense lines symbolized by the binary image are completely filled. The multi-valued image can be reproduced at a higher speed than the case.

[Fifth Embodiment] FIG. 30 is a diagram best showing the structure of the present invention. Reference numeral 3010 denotes a multi-valued image acquisition unit that acquires a digital multi-valued image to be subjected to a scaling process and outputs the raster-scanned multi-valued image to a unit that executes the next process. Reference numeral 3011 sets the magnification setting unit 3015. Magnification scaling P15, binary image creation condition data P16 set by the multi-valued image state determination unit 3016, and raster-scanned multi-valued image data P by the multi-valued image acquisition unit 3010.
11 is a smoothing / concentration scaling unit that performs scaling processing on the contour shape of the image of each density level of the multivalued image data P11. Reference numeral 3012 denotes a smooth density varying unit 301
This is a multi-valued image reproducing unit for inputting the constant density line vector data P12 scaled by 1 and reproducing a multi-value scaled image P13 in the raster scanning format from the form of the constant density line vector data P12. A multi-valued image P13 set by the multi-valued image reproducing unit 3012 and a scaling factor P15 set by the scaling factor setting unit 3015 are input to the unit 3013, and density smoothing is performed on the scaled multi-valued image data P13. This is a density smoothing section for performing.
Reference numeral 3014 denotes a multi-valued image output unit that displays the obtained scaled image, makes a hard copy, or outputs to a communication path or the like. Reference numeral 3015 denotes a magnification setting unit that sets the magnification. A multivalued image state determination unit 3016 inputs the multivalued image data P11 output from the multivalued image unit 3010 and creates the binary image creation condition data P16.

The hardware configuration of the image processing apparatus of this embodiment is the same as that of the first embodiment shown in FIG.

Next, the isoconcentration line smoothing / magnifying unit 3 in FIG.
A detailed configuration diagram of 011 is shown in FIG. 3016 is shown in FIG.
The multi-valued image data P11 output from the multi-valued image unit 3010 at 0 is input, and the binary image creation condition data P16 is input.
Is a multi-valued image state determination unit. Based on the binary image creation condition data P16 determined by the multi-valued image state determination unit 3016, reference numeral 312 indicates the multi-valued image unit 3010 in FIG.
The multi-valued image data P11 output from the above is created into a binary image data group P32 in the raster scanning format for each density level 2
It is a value image creation unit. Reference numeral 313 denotes a binary image creation unit 312.
Raster scan format binary image data group P32 created in
Is a binary image outline extraction unit for inputting the input vector and extracting the outline vector P33. 314 receives the outline vector data P33 extracted by the binary image outline extraction unit 313 and the scaling factor P15 of the scaling factor setting unit 3015 in FIG. 30 for smoothing and scaling processing, and smoothing / scaling outline data. It is an outline smoothing scaling unit that outputs P12.

FIG. 32 shows the flow of processing of the multi-valued image state judging section 3016. Data P16 that gives conditions for creating a binary image is created in the process flow of FIG. 32 for digital multi-valued image data input in the raster scanning format. The input multi-valued image is represented by f (size n in the main scanning direction and size m in the sub-scanning direction), and pixel values at coordinates x in the main scanning direction and coordinates y in the sub-scanning direction are described as f (x, y). I do. Further, the number of tones of the multi-valued image is set to L (in the case of a multi-valued image of 8 bits for each pixel, the number of tones L = 256). In step S321, a variable max = 1 for representing the maximum pixel value of the input image is set, and a variable min = L for expressing the minimum pixel value of the input image is initialized,
Also, the value of the counter y that holds the coordinate in the sub-scanning direction during scanning is initialized to 0. In step S322, the counter x holding the coordinates in the main scanning direction during scanning is initialized to zero. In step S323, the pixel value f (x, y) at the position being scanned is read from the memory area holding the input moving image. In step S324, the pixel value f (x, y) is compared with max,
If f (x, y)> max, the value of f (x, y) is substituted for max in step S325. Step S
At 326, the pixel value f (x, y) is compared with min, and f
If (x, y) <min, the value of f (x, y) is substituted for min in step S327. Step S3
In step 28, x and the size m in the sub-scanning direction are compared and compared, and x <m
If so, the value of x is incremented by 1 and the process returns to step S323. In step S329, y is compared with the size n in the main scanning direction, and if y <n, the value of y is incremented by 1 and step S is performed.
Return to 322. If y ≧ n, the scanning ends, so the processing ends.

When the above steps are performed, the maximum pixel value of the input image is substituted for max, and the minimum pixel value of the input image is substituted for min. Data P that gives these conditions the conditions for creating a binary image created by the binary image creating unit 312
Output as 16.

Next, FIG. 33 shows the flow of processing in the binary image creating section 312. With respect to the digital multi-valued image data input in the raster scanning format, a binary image group P32 for each density level is created by the flow of the processing in FIG. The input multi-valued image is represented by f (size n in the main scanning direction and size m in the sub-scanning direction), and pixel values at coordinates x in the main scanning direction and coordinates y in the sub-scanning direction are described as f (x, y). I do. From the binary image creation condition data P16 output from the multi-valued image state determination unit 3016, the pixel value f (x, y) of the input image
Is set to min, and the maximum value of the pixel value f (x, y) of the input image is set to max. In step S331, a variable Ilev for expressing the current density level.
Initialize to l = min. In step S332, f (x,
If y) ≧ Ilevel, 1 (black pixel), f (x,
If y) <Ilevel, a binary image blevel (main scanning direction size n, sub-scanning direction size m) is set to 0 (white pixel). Also in step S333, if Ilevel is equal to or less than max, the process proceeds to step S334, 1 is added to Ilevel, and step S333 is performed.
The process moves to 332. If Ilevel is greater than max, this process ends. By this processing flow, the binary image group P32 for each gradation is output.

As described above, the binary image creating unit 31
In FIG. 2, by performing the processing shown in the flow of FIG. 33, it is not necessary to create a binary image in a low gradation area or a high gradation area where no pixel exists, and (max
It is possible to output the required binary image group only by scanning for (-min) times, and it is possible to speed up the process and save the work memory.

Next, the binary image outline extracting section 313 and the outline smooth scaling section 314 in FIG.
However, for example, it can be configured by the device disclosed in the patent application filed by the present applicant as Japanese Patent Laid-Open No. 5-174140. Specifically, the binary image outline extraction unit 313 creates the binary image created by the binary image creation unit 312.
The value image group P32 is input and the outline vector (bare contour vector) P3 is extracted. The outline smoothing / scaling unit 314 creates an outline vector smoothly scaled by a desired scaling factor (arbitrary) indicated by the scaling factor signal P15 in the form of the extracted outline vector representation, and this smooth scaling is performed. A raster scanning type binary scaled image group P12 is generated from the outline vector thus generated.

The isodensity curve smoothing / magnifying unit 3011 in FIG. 30 has been described above.

Next, the multi-valued image reproducing section 30 in FIG.
The process flow of 12 will be described. This is, for example,
JP-A-5-2046 previously proposed by the applicant
It can be configured by using the device described in No. 7. The flow of this processing is shown in FIG. First, in step S341, a binary image buffer B and a multivalued image buffer G (main scanning size X, sub-scanning size Y) are initialized to 0 in a variable Ilevel = min representing a density level. Step S3
At 42, the density level is set to Ileve in the binary image buffer B.
Draw a contour line of l. In step S343, the multi-valued image is reproduced using the drawn isodensity lines. That is, the binary image buffer B is raster-scanned,
When the number of intersections with the equal density line is odd, the multi-valued image buffer G starts to fill the pixel with the density level Ilevel at the same coordinates as the binary image buffer B, and when the number of intersections is even, it is binary. The filling is completed at the same coordinates as the image buffer B. Further, when the binary image buffer B is raster-scanned, the pixels (black pixels) intersecting the isodensity line are changed to white pixels. As a result, the binary image buffer is rasterized and initialized again to proceed with the process. And
A multi-valued image is reproduced by drawing only iso-concentration lines of a certain density level on a binary image and using it to sequentially add the processing of each density level portion.

Next, the density smoothing unit 3013 in FIG.
However, for example, it can be configured by the device in the fourth embodiment. Specifically, the multivalued image P13 reproduced by the multivalued image reproduction unit 3012 and the magnification setting unit 3 in FIG.
The scaling factor P15 set by 015 is input, the filter size is determined from the scaling factor P15, and the multi-valued image P in which the density is smoothly smoothed by the uniform weight filtering process is input.
Get 14.

As described above, when a binary image for each density level is created from an input multi-valued image, a binary image is generated in a low density area where no pixel having the density value exists or in a high density area. By not performing the process, it is possible to speed up the process and save the work memory without degrading the image quality.

[Sixth Embodiment] FIG. 35 is a histogram showing the number of pixels having the density value at each density level. The horizontal axis represents the density level and the vertical axis represents the density level. Represents the number of pixels that have a density value. Smooth density variable / magnifying unit 3011 in FIG. 30 of the fifth embodiment.
, As shown in FIG. 35, min to max
The process of creating a binary image is performed in the density level area 71 up to. However, as shown in FIG.
Even within the density level range 71 from min to max, the density level 7 in which no pixel having that density value exists
With respect to the input multi-valued image having 2, the processing is performed even on the density level area 72 in which the pixel having the density value does not exist.

Therefore, as shown in FIG. 35, the number of pixels having the density value at each density level is counted, and a binary image is created for the density level at which the number of black pixels is 0. By performing the process not to
It is possible to further speed up the processing and save the working memory as compared with the above embodiment. Specifically, FIG. 36 shows the flow of processing of the multivalued image state determination unit 3016 in FIG. 32 of the fifth embodiment.
As shown in FIG. 33, the processing flow of the binary image creation unit 312 in FIG. 33 can be replaced with that shown in FIG. 37, and the processing of the multivalued image reproduction unit 3012 shown in FIG. 34 can be replaced with that shown in FIG. Becomes The other parts are the same as in the fourth embodiment.

The processing of FIG. 36 will be described. Each concentration level I
A data area H (n) (n = 0 to L-1) for storing the number of pixels having the density value in the level is reserved in the storage device 23. In step S361, all of these values are initialized to 0, and the value of y, which is a counter that holds the coordinate in the sub-scanning direction during scanning, is initialized to 0. In step S362, the value of x, which is a counter that holds the coordinate in the main scanning direction during scanning, is initialized to 0. In step S363, the pixel value f (x, y) at the position being scanned
Is read from the memory area that holds the input moving image. In step S364, the pixel value f (x, y) of the input multi-valued image is incremented by 1 for the value H (f (x, y)). In step S365, x is compared with the size m in the sub-scanning direction, and if x <m, the value of x is set to 1
Increase and return to step S363. Step S366
Then, y is compared with the size n in the sub-scanning direction, and if y <n, the value of y is incremented by 1 and the process returns to step S362. y ≧
If it is n, the scanning ends, so the process ends. If all pixels on the input image are processed, H (0) to H (L-1)
Will store the number of black pixels at each density level. These values are converted to the binary image creation unit 312.
The data P16 is output as data P16 that gives the conditions for creating the binary image.

The processing of FIG. 37 will be described. Step S37
1, the variable Ilevel representing the current concentration level
= 1. Based on the binary image creation condition data P16, the number of pixels having the density value at each density level is H (0) to H (L-1). If H (Ilevel) = 0 in step S372, the binary image creation process is not performed, and the process proceeds to step S375. H (Ile
vel)> 0, in step S373, 2
Create a value image blevel. In step S374, if the variable Ilevel is (L-1) or less,
The process proceeds to step S375, 1 is added to Ilevel, and the process proceeds to step S372. If Ilevel is (L-
If it is larger than 1), this process ends.

By performing the above-described processing, it is possible to create a binary image only in the density level where the pixels having the density value exist over the entire density range.

The processing of FIG. 38 showing the processing flow of the multi-valued image reproducing section 3012 will be described. First, in step S381, the variable Ilevel = 1 is set to represent the density level, and the binary image buffer B and the multivalued image buffer G (main scanning size X, sub-scanning size Y) are initialized. Step S382
In step S3, if the value of the binary image creation condition data H (Ilevel) input from the multi-valued image state determination unit 3016 is 0, the contour line drawing process is not performed and step S3 is performed.
Go to 85. If the value of H (Ilevel) is not 0, the process proceeds to step S383. In step S383,
An isodensity line having a density level of Ilevel is drawn in the binary image buffer B. In step S384, the drawn isodensity line is used to reproduce the multivalued image. That is, 2
The value image buffer B is raster-scanned, and when the number of intersections with isodensity lines is an odd number, filling of pixels with density Ilevel in the multivalued image buffer G is started at the same coordinates as the binary image buffer B. Then, when the number of intersections is an even number, the filling is completed at the same coordinates as the binary image buffer B. Also, when raster-scanning the binary image buffer B,
Pixels (black pixels) that intersect the isodensity line are changed to white pixels. As a result, the binary image buffer is rasterized and initialized again to proceed with the process. In step S385, if Ilevel> L−1, the process is ended, and if Ilevel ≦ L−1, the process proceeds to step S386, the value of Ilevel is incremented by 1, and the process proceeds to step S382. With the above processing, it is possible to draw only iso-concentration lines of a certain density level on a binary image, and by using it, the processing of each density value part can be sequentially added to reproduce a multi-valued image. Becomes

In the sixth embodiment, the number of black pixels existing is counted for each density level, but it is not always necessary to count the number of black pixels, and the black pixels are not counted for each density level. It suffices to set a memory area for storing whether or not it exists. Specifically, the processing content of step S364 of FIG. 36 in the sixth embodiment is set to H (f
You may change to the process of substituting 1 for (x, y)). As a result, the value of H (n) is set to 0 at the density level where the pixel having the density value does not exist, and the value of H (n) is set to 1 at the density level where the pixel having the density value exists. Therefore, if the value of H (n) is output as the data P16 that gives the condition for creating the binary image, the binary image creating unit 3012 follows the processing flow of FIG. It is possible to perform a process of creating a binary image only at a density level at which a pixel having that density value exists. By performing the above-described processing, it is possible to perform a faster calculation than in the case of the second embodiment.

In the fifth and sixth embodiments, the binary image creating section 312 in FIG. 31 creates binary images of all necessary density levels, and the binary image outline extracting section 3
An outline is extracted for all necessary binary images in 13, and an outline smoothing / magnifying unit 314 performs smoothing / magnifying processing. However, it is not always necessary to follow this flow. That is, a process of creating a binary image of a certain density level, performing outline smoothing / magnifying processing on the binary image, and reproducing the multi-valued magnified image according to the process of step S343 of FIG. Should be repeated for the number of gradations.

For example, the flow of processing of the fifth embodiment is shown in FIG.
You may replace like the flow of the process shown in 9. This flowchart is a procedure of a program executed by the central processing unit 26 of FIG. 39 will be described. In step S391, the multi-valued image acquisition unit 3010 in FIG.
A multi-valued image is input by the same process as in step S392, and in step S392, binary image creation condition data min, max from the state of the input multi-valued image is processed by the same process as the multi-valued image state determination unit 3016 of FIG. To set. In step S393,
Initialize to a variable Ilevel = min that represents the current concentration level. In step S394, the density level is Ilev.
l binary image is created, outline smoothing / magnifying processing is performed on the binary image in step S395, and in step S396, a multi-valued image stored in a predetermined area according to the processing in step S343 of FIG. Processing such as reproducing the variable-magnification image is performed. I in step S397
If level ≦ max, the process moves to step S398 and I
Add 1 to level and proceed to step S394, where Il
If ever> max, the process is terminated.

The processing flow in the sixth embodiment may be replaced with the processing flow shown in FIG. This processing is also executed by the central processing unit 26. 40 will be described. In step S401, a multi-valued image is input by the same process as the multi-valued image part 3010 in FIG. 30, and in step S402, the input multi-valued image is processed by the same process as the multi-valued image state determination part 3016 in FIG. Binary image creation condition data H (n) is set from the state. Step S403
Is initialized to a variable Ilevel = 1 representing the current density level. In step S404, if H (Ile
vel) = 0, the process moves to step S408,
If not, the processes in and after step S405 are performed. In step S405, a binary image of a region whose density level is Ilevel or higher is created, and in step S406, the binary image is generated.
Outline smoothing / magnifying processing is performed on the value image, and in step S407, processing is performed such that the multi-value magnified image stored in a predetermined area is reproduced according to the processing of step S384 in FIG. Step S408
If Ilevel ≦ L−1, the process proceeds to step S409, 1 is added to Ilevel and the process proceeds to step S404, and the process is ended if Ilevel> L−1.

By performing the above processing, a series of processing is performed for each density level, so that it is possible to considerably reduce the necessary storage area.

As described above, the image processing apparatuses and methods according to the fifth and sixth embodiments can obtain a scaled image with good image quality, and can prevent image quality deterioration such as jaggies and lattice distortion. Can be suppressed.

Further, since the distribution state of pixels for each density level area of the multi-valued input image is grasped and the binary image is not created at the density level where the black pixel does not exist, it is necessary to create the binary image group. The required work memory can be saved, and the processing is not performed at the density level that is not required, so that high-speed processing can be performed.

[Seventh Embodiment] <Device Configuration> FIG. 41 is a view best showing the structure of the present embodiment. Reference numeral 4111 denotes a multi-valued image acquisition unit that acquires a digital multi-valued image to which scaling processing is performed and outputs a raster-scanning-type multi-valued image P20. Reference numeral 4112 determines that it is better to reverse the input multi-valued image P20. If so, the input multivalued image is inverted and output, otherwise, the input image is output as it is, and the inversion flag P26 indicating whether or not the input image P20 is inverted is also output, and the multivalued image inversion unit. Is. 4113 inputs the multi-valued image data P21 from the multi-valued image inversion unit 4112 and the scaling factor P27 set by the magnification setting unit 4115, and smoothes the contour vector data of the uniform density line shape of the multi-valued image data P21. -It is a unit for smoothing and scaling the iso-density line that performs scaling processing. 41
14 inputs the constant density outline vector data P22 smoothed and scaled by the constant density smoothing / scaling unit 4113 and reproduces a raster scanning multi-value scaled image P23 from the form of the constant density vector data. It is a multi-valued image reproducing means. Reference numeral 4116 inputs the scaling factor P27 set by the multi-valued image reproducing unit 4115, and scales the multi-valued image data P.
23 is a density smoothing unit that performs density smoothing for 23. 411
7 determines whether or not to invert the multi-valued image P24 output from the density smoothing unit 4116 based on the inversion flag P26 output from the multi-valued image inversion unit 4112, and when the inversion flag is ON. Is a multivalued image inverting unit that inverts the multivalued scaled image P24, and outputs the multivalued scaled image as it is when it is OFF. Reference numeral 4118 denotes a multivalued image output unit that displays the obtained multivalued variable-magnification image, makes a hard copy, or outputs the image to a communication path or the like.

Similar to the first embodiment, FIG. 2 is a conceptual diagram showing the hardware configuration of an image processing apparatus which realizes the configuration of FIG. The multi-valued image input from the image input device 21 or stored in the external storage device 27 such as a hard disk, a floppy disk, or a CDROM is stored in the storage device 23. An instruction for image scaling processing input by the operation input device 25 is displayed on the operation content display device 24. In response to this, the central processing unit 23 executes the specified process while accessing the storage device 26, Output to the output device 22 or the external storage device 2
The processing result is stored in 7. Thus, the central processing unit 26
By controlling the entire apparatus of FIG. 2, the configuration of FIG. 41 is realized.

Here, the processing executed by the central processing unit 26 includes the procedures represented by the flow charts described later, and these procedures are stored in the external storage device 27, read out to the storage device 23 and read out by the central processing unit. 26. <Inversion of Image> Next, the multivalued image inversion unit 4 in FIG.
The process flow of 112 will be described with reference to FIG.

It is often convenient to treat an image having a high average density as a white (relatively low density) foreground drawn on a black (relatively high density) background. The multi-valued image inversion unit 4112 inverts such an image once, and black (of the high density part) against a white (low density part) background.
The image is the foreground. As a result, an image with a white (low density) foreground drawn on a black (high density) background becomes an image with a black (high density) foreground drawn on a white (low density) background. By the scaling / smoothing processing, it is possible to avoid the problem that a white (low density) area that should be continuous is divided.

For example, as shown in FIG. 42 (a), the black pixels connected in an oblique direction are treated as a part of a continuous shape in which the black pixels are connected, not as an array of independent points, and scaling / smoothing is performed. Was going on. Therefore, for example, in FIG.
A black line having a width of one pixel in the diagonal direction as shown in (b) can retain the shape as a line in the original image by scaling and smoothing, but as shown in FIG. If there is a line, the white pixels are divided because they are treated as continuous black pixels, and each white pixel is regarded as an independent white point, and the shape of the white line should be maintained after scaling and smoothing. I couldn't.

As described above, the pixels are arranged so that the pixels having the diagonal positional relationship have the same color and the adjacent pixels have different colors, that is, the parts in which the pixels are arranged in a staggered pattern. Then, by processing as if the black pixels are continuous, as a result of scaling and smoothing, the white pixels are divided even if the white pixels are originally continuous, The original image may be lost due to scaling and smoothing. The apparatus of this embodiment can solve such a problem.

In FIG. 43, in step S431,
The inversion flag INV is turned off and initialized. In step S432, the average density value S of the input image is calculated. For example, if the input image is f (main scanning size n, sub scanning size m,
The maximum density value L), and the density value at coordinates (x, y) is f
When expressed by (x, y), it can be obtained by the following formula.

S = {ΣxΣyf (x, y)} / (n × m) (6) (where x = 0, 1, ..., N-1, y = 0, 1, ...
m-1) In step S433, the average density value S obtained in step S432 is compared with a certain threshold value T. If the average density value S is greater than or equal to the threshold value T, the process proceeds to step S434 and the input multi-valued image P20. Invert. For example, when the inverted multi-valued image is represented by f ′, the inverted image is created as follows: f ′ (x, y) = L−f (x, y) (7) And step S
At 435, the inversion flag INV is turned on. If the average density value S is smaller than the threshold value T in step S433, the input image P20 is output as it is. <Equal Density Line Outline Magnification / Smoothing> Next, for the equal density line smoothing / magnification unit 4113 in FIG. 41, the multi-valued image P21 output from the multi-valued image inverting unit 4112 is input and each density is changed. A raster type binary image data group is created for each level, and vector data of the contour shape is extracted for each binary image data. That is an isodensity line that is the contour of a region of equal density. Since this isodensity line is vector data, it is not just a contour line, but it is necessary to determine which side of the isodensity line is the black area, that is, which side has a higher density than the density gradation indicated by the isodensity line. You can Smooth and scale the contour vectors. In this way, the equal density line vector data P22 that is smoothly smoothed and scaled is output.

Next, FIG. 44 shows a detailed configuration diagram of the isodensity curve smoothing / magnifying unit 4113 in FIG. 4401
Is a binary image creation unit that creates a binary image data group P61 in the raster scanning format for each density from the multivalued image data P21 output from the multivalued image inversion unit 4112 in FIG. A binary image 4402 is a binary image for creating the binary image outline vector data group P62 by extracting the vector data of the contour shape of all the raster scanning binary image data groups P61 created by the binary image creating unit 4401. It is an outline extraction unit. Reference numeral 4403 denotes the binary image outline vector data group P62 extracted by the binary image outline extraction unit 4402 and the scaling factor P27 set by the scaling factor setting unit 4115 in FIG. Is smoothed and scaled in vector data format, and smoothed and scaled 2
An outline smoothing / magnifying unit for outputting a value image outline vector data group P22. This isobaric line smoothing
The scaling unit 4113 creates a binary image for each density level, extracts the contour shape of the binary image, and smoothes and scales the extracted outline vector. Is equivalent to extracting and smoothing and scaling.

Next, FIG. 45 shows the flow of processing of the binary image creating section 4401. With respect to the digital multi-valued image data P21 input in the raster scan format, a binary image for each gradation is created by the processing flow of FIG. The input multi-valued image is defined as f (main scanning direction size n and sub scanning direction size m), and the pixel value in the main scanning direction coordinate x and the sub scanning direction y is described as f (x, y). . Further, the number of gradations of the multi-valued image is L. In the case of a multi-valued image having 8 bits for each pixel, the number of gradations L = 256.

In step S451, a variable Ilevel = 1 for expressing the current gradation is initialized. In step S452, 1 (black pixel) if f (x, y) ≧ Ilevel, and 0 if f (x, y) <Ilevel.
A binary image blevel (main scanning direction size n, sub-scanning direction size m) that is (white pixels) is created. If Ilevel is (L-1) or less in step S453, the process proceeds to step S454, 1 is added to Ilevel, and the process proceeds to step S452. If Ile
If vel is larger than (L-1), this process ends. By the flow of this processing, the binary image group P61 for each density level in FIG. 44 is output.

Next, the binary image outline extracting section 4402 and the outline smoothing / magnifying section 4403 in FIG.
However, it can be configured by the technique disclosed in Japanese Patent Laid-Open No. 5-174140. Specifically, in the binary image outline extraction unit 4402, the binary image creation unit 4401.
By inputting the binary image group P61 created in, the binary image outline vector data group (coarse contour vector), that is,
The iso-density vector data P62 of the multi-valued image is extracted, and the iso-density vector data P62 extracted by the outline smoothing / magnifying unit 4403 is expressed in outline vector expression in the form of an outline vector expression. An (outright) smoothly scaled outline vector is created, and smoothed / scaled equal density line outline vector data P52 is output.

Next, the multi-valued image reproducing section 41 in FIG.
14 will be described. This can be configured by using, for example, the device described in Japanese Patent Application Laid-Open No. 5-20467 previously proposed by the present applicant. The flow of this processing is shown in FIG. First, in step S461, the variable Ilevel representing the gradation is set to Ilevel = 1, and the binary image buffer B and the multivalued image buffer G (main scanning direction size X, sub-scanning direction size Y) are initialized. Step S4
At 62, the density level is set to level in the binary image buffer B.
Is drawn. In step S463, the drawn isodensity line is used to reproduce the multivalued image. That is, the binary image buffer B is raster-scanned, and when the number of intersections with the isodensity line is an odd number, the multivalued image buffer G
Then, the filling of the pixel having the density level is started at the same coordinates as the binary image buffer B, and the filling is finished at the same coordinates as the binary image buffer B when the number of intersections is an even number. Further, when the binary image buffer B is raster-scanned, the pixels (black pixels) intersecting the isodensity line are changed to white pixels. This means that the binary image buffer is initialized while raster scanning, and it is not necessary to initialize the binary image buffer B each time a certain density level is processed.

As described above, since the equal density lines of a certain gradation are drawn in the binary image and the multi-valued image is reproduced by using the drawn lines, the equal density drawn in the binary image is reproduced. It is possible to reproduce the multi-valued image at a higher speed than in the case of reproducing the multi-valued image after completely filling the inside of the line. <Density Smoothing> Next, the density smoothing unit 411 in FIG.
The configuration of No. 6 is shown in FIG. The density smoothing unit 4116 includes a uniform weight filter processing unit 4700 and a filter size determining unit S91. The filter size determining means is shown in FIG.
Magnification P2 obtained from the magnification setting unit 4115 in
7 is input and the filter size 91 is output. The uniform weight filter processing unit 4700 receives the output image P23 of the multi-valued image reproducing unit 4114 in FIG. 5 and the filter size P91 which is the output from the filter size determining unit 4701, and finally performs the density smoothing processing. Value image P24
Is output. The uniform weight filter unit 4700 is, for example, the above-mentioned “Introduction to Computer Image Processing” (Tamura, Soken Publishing).
It is a well-known method introduced in, for example, and is usually used for noise removal. The outline of this process is shown in FIG. 9 as in the first embodiment.

The method of determining the filter size is the same as in the other embodiments.

Specifically, the density is smoothed by performing a filtering process using a uniform filter as shown in FIG. 9 determined according to the formula (3) of the first embodiment according to the magnification. FIG. 10 is a diagram schematically showing this density smoothing. <Re-inversion of image> Next, FIG. 48 shows a flow of processing of the multi-valued image re-inversion unit 4117 in FIG. The multi-value image re-inversion unit 4117 is the multi-value image inversion unit 4112 in FIG.
The inversion flag INV set in step 4 and the multivalued image P24 whose density is smoothed by the density smoothing unit 4116 are input. FIG. 48 is a flowchart showing a processing procedure by the multi-valued image re-inversion unit 4117.

In step S481, it is determined whether the input inversion flag INV is ON or OFF. If the input inversion flag INV is ON, that is, if the input image is inverted before processing, the process proceeds to step S482 to smooth the density. Then, the multivalued image P24 is inverted again and output. That is, the density smoothed image P
25 is F (main scanning size N, sub-scanning size M, maximum density value L), and the density value at coordinates (x, y) is F (x, y).
Expressed as, the inverted image F ′ can be obtained by the following formula.

F ′ (x, y) = L−F (x, y) (8) (where x = 0, 1, ... N−1, y = 0, 1, ... M
-1) If OFF, density-smoothed multivalued image P25
Is output as is.

In this way, the input multi-valued image can be scaled and smoothed and output.

As described above, as the multi-valued image scaling process for smoothing and scaling the isodensity curve shape and additionally smoothing the density, the property of the input image, that is, whether the average density is higher than a predetermined threshold value, is determined. By reversing the image once depending on whether it is low, even if the image is formed by connecting white pixels in an oblique direction, it is possible to scale and smooth as drawn in the original image. It is possible to obtain a high-quality multi-valued image smoothed as expected in any multi-valued image.

In this embodiment, 2 corresponding to each gradation is used.
In order to smooth the equal density lines by regarding the black pixels of the value image as continuous, the image with the average density equal to or higher than the predetermined threshold was inverted, but in the binary image corresponding to each gradation,
When the white pixels are regarded as continuous and the contour vector extraction and smoothing are configured, that is, when the black pixels arranged in the diagonal direction are divided by the white region, the determination when inverting is performed. By reversing the reference and reversing the density of the multi-valued image input when the reference value is equal to or less than the predetermined threshold value, the same effect as that of the above-described embodiment can be obtained.

<Modification of Seventh Embodiment> In the seventh embodiment, the multi-valued image inverting unit 4112 of FIG. 41, that is, FIG.
When calculating the average density S of the input multi-valued image P20 in step S432 in the flow of the process shown in FIG. 3, the average density S was calculated using the density values of all the pixels of the input multi-valued image (Equation 6 ), And does not necessarily have to be configured in this way. That is, the input image may be thinned out and the average density (S ′) thereof may be used. That is, if the thinning rate is c (c ≧ 1), then s ′ = c ^ 2 {ΣxΣyf (x, y)} / (n × m) (9) (where x = 0, c, 2c ... , c × [n / c] -1; y = 0, c, ..., c
× [m / c] -1; where a ^ b represents a to the bth power, and [α] represents truncation of α) As a result, compared with the case where the average density is calculated using all pixels The time can be shortened by thinning out the average concentration.

[Second Modification of Seventh Embodiment] In the seventh embodiment, the multi-valued image inverting section 4112 of FIG. 41 uses the average density value as a reference value for whether or not to invert, but it is not always necessary. It need not be so configured. For example,
In the case of an image showing a density distribution (vertical axis: frequency, horizontal axis: density value) as shown in FIG. 49, the density average is a value indicated by a dotted line 490. The average density of the input image densities is used as the reference value S in step S433 in FIG. 43, and the threshold value T = L /
2 (dotted line 491 in FIG. 49, L is the maximum density value),
Since S <T, the input image is processed without being inverted. However, the maximum frequency density (dotted line 492 in FIG. 49) is used as the image density S which is the criterion, and the threshold T
= L / 2, S> T, and the inversion process is executed for this input image.

As a result, not only the average value but also the state of the density distribution can be reflected as the judgment criterion for the inversion process, and the property of the input image can be judged more accurately, so that a better scaling result can be obtained. It becomes possible to obtain.

In addition to the maximum frequency, a statistical value such as a median value (median) can be used as a criterion for the multivalued image inverting unit 4112.

FIG. 54 is a flow chart of a program executed by the central processing unit 26 of FIG. 2 in order to realize the seventh embodiment.

In step S541, a multi-valued image is input from the input device 21, and in step S542 it is inverted as necessary. This procedure is as shown in FIG. In step S543, a binary image for each density level is generated from the multivalued image. This procedure is as shown in FIG. In step S544, the outline vector of each binary image, that is, the contour line vector is extracted. In step S545, the extracted isodensity curve vector is scaled and smoothed. Step S544
S545 uses the technique disclosed in Japanese Patent Laid-Open No. 5-174140.

In step S546, the multivalued image is reproduced according to the procedure shown in FIG. After density-smoothing it, in step S548, the image is inverted in the procedure of FIG.
Finally, in step S549, the image is output.

[Eighth Embodiment] In the seventh embodiment,
The multi-valued image inversion unit 4112 inverts the input multi-valued image using some reference value, or outputs it as it is,
Although it is compatible with any multi-valued image, it does not necessarily have to be configured in this way. A configuration diagram in that case is shown in FIG. That is, the multivalued image inversion determination unit 5001 only determines whether or not to invert based on the nature of the input image, and the isodensity curve smoothing / magnifying unit 5002 determines when a binary image is created (see FIG. 45).
Step S452) of the
The same result as when the input multi-valued image is inverted can be obtained.

FIG. 51 shows the processing procedure of the multi-valued image inversion determination unit 5001 in that case.
FIG. 52 shows a flowchart of the processing operation for creating a binary image according to No. 2.

In FIG. 51, step S434 of FIG. 43 is omitted, and if the average density S of the input image is larger than T = L / 2, the inversion flag INV is turned on.

FIG. 52 shows a process of creating a binary image while inverting it according to the flag INV when creating a binary image corresponding to each density level. That is, the flag I
If NV is off, pixels with a density equal to or higher than the target density level are set to 1 in step S452, but if the inversion flag INV is on, step S47.
In 1, the pixels having a density lower than the target density level are set to 1. By doing so, a reversed image can be generated when it is necessary to reverse the image while creating a binary image.

In accordance with this, in the multivalued image reproducing section 5003 which reproduces the multivalued image from the constant density line vector, the reversed image is restored while being reproduced. A flowchart of this processing operation is shown in FIG.

In FIG. 53, the contour line of the binary image of the density level of interest is drawn in the binary image buffer B (step S
462), while scanning it, the pixel of the density represented by the variable Ilevel in the multi-valued image buffer G is moved to the position of the buffer G corresponding to the buffer B being scanned every time the contour line and the scanning line intersect. Start / end writing. The start / end condition is that when the inversion flag INV is off,
As in the case of FIG. 46, the writing is not started at the start of scanning, the writing is started at the odd-numbered intersections, and the writing is ended at the even-numbered intersections (step S463).

On the other hand, when the inversion flag INV is on, writing is started from the start of scanning, the writing is ended at the odd-numbered intersections, and the writing is started at the even-numbered intersections (step S531). By doing so, at the same time as the reproduction of the multi-valued image, the image can be inverted if necessary.

With these processes, it is possible to speed up the process and reduce the memory capacity.

The apparatus of this embodiment can also be applied to a video printer which converts data from a television into multi-valued image data and outputs an image based on the converted multi-valued image data.

FIG. 55 is a flow chart of a program executed by the central processing unit 26 of FIG. 2 to realize the device of the eighth embodiment. The same processes as those in FIG. 54 are designated by the same reference numerals, and description thereof will be omitted.

In step S551, it is determined by the procedure of FIG. 51 whether the input image should be inverted.

At step S552, step S551.
A binary image corresponding to each density level is generated without being inverted or while being inverted in accordance with the result of the determination in (1). This is shown in FIG.

In step S553, the image is reproduced according to the determination result in step S551 according to the procedure shown in FIG. That is, a multi-valued image is generated so that the inverted binary image is re-inverted.

The programs shown in the flow charts of the first to eighth embodiments are executed by the external storage device 2 of FIG.
7, can be supplied to the system of FIG. 2 by a removable medium such as an FD, an optical disc, a magneto-optical disc, or the like. The program is executed by the CPU, and the devices of the first to eighth embodiments can be realized.

As described above, the image processing apparatus and method according to the present invention can obtain a multi-value scaled image that is scaled / smoothed faithfully to the original image.

[0232]

[Other Embodiments] Even if the present invention is applied to a system including a plurality of devices (for example, a host computer, an interface device, a reader, a printer, etc.), an apparatus (for example, a copying machine) Machine, facsimile machine, etc.).

Another object of the present invention is to supply a storage medium storing a program code of software for realizing the functions of the above-described embodiment to a system or apparatus, and to supply a computer (or CPU) of the system or apparatus.
Or MPU) reads and executes the program code stored in the storage medium.

In this case, the program code itself read from the storage medium realizes the novel function of the present invention, and the storage medium storing the program code constitutes the present invention.

A storage medium for supplying the program code is, for example, a floppy disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD.
-R, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used.

Further, not only the functions of the above-described embodiments are realized by executing the program code read by the computer, but also the OS (operating system) running on the computer based on the instructions of the program code. ) And the like perform some or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

Further, after the program code read from the storage medium is written in the memory provided in the function expansion board inserted into the computer or the function expansion unit connected to the computer, based on the instruction of the program code, This also includes a case where a CPU or the like included in the function expansion board or the function expansion unit performs some or all of the actual processing and the processing realizes the functions of the above-described embodiments.

When the present invention is applied to the above-mentioned storage medium, the storage medium stores the program code corresponding to the above-mentioned flow chart. Briefly, in the memory map example of FIG. Each module shown will be stored in the storage medium. That is, at least the program code of each module of the isodensity curve vector extraction step, the scaling / smoothing step module, and the multivalued image generation step module may be stored in the storage medium.

[0239]

As described above, the image processing apparatus and method according to the present invention have an effect of preventing deterioration of image quality such as generation of jaggies and lattice distortion.

Also, for each density level, it is checked whether or not there is a pixel of that level, and if there is no such density level, the density level is skipped, so that the process of reproducing a multi-valued image becomes quick.

Further, the input multi-valued image is subjected to pixel value interpolation and smoothing /
By supplying separately for scaling, scaling with different methods and then integrating, even when enlarging at high magnification, the contour part is not emphasized too much, and scaling processing by pixel value interpolation is performed. It is possible to obtain a variable-magnification image in which the occurrence of jaggies and grid-like distortion that become noticeable when the above is applied is suppressed.

Furthermore, since the number of gradations of the image to be subjected to the smoothing / magnifying processing is 1 / s to the power of 2, the amount of the binary image outline smoothing / magnifying processing is also 1 / S to the power of 2. become. Therefore, the processing time can be shortened as compared with the second embodiment by changing the distribution ratio.

In addition, since only a certain density level of the uniform density line vector data is drawn in the binary image and the multivalued image is reproduced by utilizing that, the binary image is drawn. It is possible to reproduce the multi-valued image at a higher speed than when the multi-valued image is reproduced after the entire density line is completely filled.

Further, from the input multi-valued image, 2 for each density level is obtained.
When a value image is created, by performing the process of not creating a binary image in a low density area or a high density area where pixels having that density value do not exist, high-speed processing is possible without degrading image quality. And work memory can be saved.

Further, as a multi-valued image scaling process for smoothing / scaling the iso-density line shape and additionally smoothing the density, depending on the nature of the input image, that is, whether the average density is higher or lower than a predetermined threshold value. By reversing the image once, even if the image is formed by connecting white pixels in a diagonal direction, it is possible to scale and smooth the image as drawn in the original image. Also in, it is possible to obtain a high-quality multi-valued image smoothed as expected.

[0246]

[Brief description of drawings]

FIG. 1 is an overall block configuration diagram of an apparatus according to a first embodiment.

FIG. 2 is a diagram showing a specific configuration of all embodiments.

FIG. 3 is a block configuration diagram of a multi-valued image outline scaling unit of the first embodiment.

FIG. 4 is a flowchart showing a processing operation of a binary image creating unit of the first embodiment.

FIG. 5 is a diagram showing a principle of generating a gradation level binary image from a multi-valued image.

FIG. 6 is a block diagram of a binary image outline scaling unit according to the first embodiment.

FIG. 7 is a flowchart showing a processing operation of a multi-valued image reproducing unit of the first embodiment.

FIG. 8 is a block configuration diagram of a pixel value interpolation unit of the first embodiment.

FIG. 9 is a conceptual diagram showing a uniform weight filter processing operation of the first embodiment.

FIG. 10 is a conceptual diagram showing a processing principle of a pixel value interpolation unit of the first embodiment.

FIG. 11 is a diagram showing a filter used in a third embodiment.

FIG. 12 is a flow chart of a program for realizing the devices of the first and second embodiments.

FIG. 13 is a diagram showing a flow of outline smoothing / magnifying processing of a value image in the second embodiment.

FIG. 14 is a binary image creating unit 121 according to the second embodiment.
3 is a flowchart showing the flow of the processing of FIG.

FIG. 15 is a diagram showing a processing flow of a binary image outline smoothing / magnifying unit 122 in the second embodiment.

FIG. 16 is a multi-valued image reproducing unit 124 in the second embodiment.
3 is a flowchart showing the flow of the processing of FIG.

FIG. 17 is a diagram showing a flow of outline smoothing / magnifying processing of a multivalued image according to the third embodiment.

FIG. 18 is a multi-valued image separation unit 170 according to the third embodiment.
FIG. 3 is a block diagram showing an example of the configuration of FIG.

FIG. 19 is a flowchart of a program for realizing the device of the third embodiment.

20 is a flowchart of an image separation step in the flowchart of FIG.

FIG. 21 is a block diagram showing the configuration of the fourth embodiment.

FIG. 22 is a block configuration diagram of a contour line smoothing / magnifying unit in the fourth embodiment.

FIG. 23 is a flowchart showing the processing operation of the binary image creating means in the fourth embodiment.

FIG. 24 is a flow chart showing the processing operation of the multi-valued image reproducing means in the fourth embodiment.

FIG. 25 is a conceptual diagram showing the processing principle of the multi-valued image reproducing means in the fourth embodiment.

FIG. 26 is a block diagram of a density smoothing unit according to the fourth embodiment.

FIG. 27 is a block diagram of a modification of the fourth embodiment.

FIG. 28 is a hardware configuration diagram according to a modification of the fourth embodiment.

FIG. 29 is a block configuration diagram of a constant density line smoothing / magnifying unit in a modification of the fourth embodiment.

FIG. 30 is a block diagram showing the configuration of the fifth embodiment.

FIG. 31 is a block configuration diagram of a uniform density line smoothing / magnifying unit in fifth and sixth embodiments.

FIG. 32 is a flow chart showing the processing operation of the multi-valued image state determination means in the fifth embodiment.

FIG. 33 is a flowchart showing the processing operation of the binary image creating means in the fifth embodiment.

FIG. 34 is a flowchart showing the processing operation of the multi-valued image reproducing means in the fifth embodiment.

FIG. 35 is a diagram for explaining the sixth embodiment.

FIG. 36 is a flowchart showing a processing operation of a multi-valued image state judging means in the sixth example.

FIG. 37 is a flowchart showing the processing operation of the binary image creating means in the sixth embodiment.

FIG. 38 is a flowchart showing the processing operation of the multi-valued image reproducing means in the sixth embodiment.

FIG. 39 is a flowchart for realizing the fifth embodiment.

FIG. 40 is a flowchart for realizing the sixth embodiment.

FIG. 41 is a block diagram showing the configuration of the seventh embodiment.

FIG. 42 is a diagram for explaining a problem.

FIG. 43 is a flowchart showing the processing operation of the multi-valued image inverting means in the seventh embodiment.

FIG. 44 is a block diagram of a contour line smoothing / magnifying unit in the seventh embodiment.

FIG. 45 is a flowchart of binary image generation according to the seventh embodiment.

FIG. 46 is a flowchart of multi-valued image generation in the seventh embodiment.

FIG. 47 is a block diagram of a density smoothing unit in the seventh embodiment.

FIG. 48 is a flowchart showing the processing operation of the multi-valued image re-inversion means in the seventh embodiment.

FIG. 49 is a conceptual diagram for explaining a modification of the seventh embodiment.

FIG. 50 is a block diagram of an eighth example.

FIG. 51 is a flowchart showing a multi-valued image inversion determination process in the eighth embodiment.

FIG. 52 is a flowchart of a binary image creating operation process according to the eighth embodiment.

FIG. 53 is a flowchart of multi-valued image reproduction processing according to the eighth embodiment.

FIG. 54 is a flow chart of a program for realizing the seventh embodiment.

FIG. 55 is a flow chart of a program for realizing the eighth embodiment.

FIG. 56 is a diagram showing an overall configuration in the technique previously proposed by the applicant.

FIG. 57 is a diagram showing a scanning direction when an outline vector is extracted from a binary image in the technique previously proposed by the applicant.

[Fig. 58] Fig. 58 is a diagram for explaining the previously proposed principle of extracting a contour vector.

FIG. 59 is a diagram showing a relationship between a proposed binary image and a contour vector.

FIG. 60 is a diagram showing a format of proposed and extracted contour vector data.

FIG. 61 is a diagram showing a configuration of a proposed smoothing scaling unit.

FIG. 62 is a diagram showing a specific configuration of a proposed device.

FIG. 63 is a flowchart showing the flow of a proposed process.

[Fig. 64] Fig. 64 is a diagram illustrating an outline of a proposed first smoothing process.

[Fig. 65] Fig. 65 is a diagram illustrating the format of vector data after the proposed smoothing.

FIG. 66 is a diagram showing an outline of a proposed second smoothing process.

FIG. 67 is a memory map of a program that realizes the image processing apparatus of the present invention.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Takahiro Oshino 53 Imaiuecho, Nakahara-ku, Kawasaki-shi, Kanagawa Canon Inc. Kosugi Plant

Claims (61)

[Claims] 1. An image processing apparatus for executing a scaling process on a multi-valued image, comprising iso-density line vector extraction means for extracting an outline vector for each density level of the multi-valued image, and each extracted A scaling / smoothing unit that scales and smoothes the outline vector for each density level, and a multivalued image is generated based on the outline vector of each density level obtained by the processing of the scaling / smoothing unit. The image processing apparatus according to claim 1, further comprising a multivalued image generation unit. 2. The iso-concentration line vector extraction means uses a predetermined density level as a threshold value to binarize the multi-valued image, and a binary image generation means and two binary image generation means. The image processing apparatus according to claim 1, further comprising an outline extraction unit that extracts an outline vector of the value image. 3. The image processing apparatus according to claim 2, wherein the predetermined density level is all possible density levels of each pixel of the multi-valued image. 4. The image processing apparatus according to claim 2, wherein the predetermined density level is a predetermined density level for every predetermined number of density levels. 5. The outline extracting means extracts a boundary line between a white pixel and a black pixel of a binary image as a vector having a direction according to a positional relationship between the white pixel and the black pixel, and the scaling ratio. The image processing apparatus according to claim 2, wherein the smoothing unit scales the vector and moves the endpoints of the vector according to the shape formed by the vector to smooth the vector. 6. The multi-valued image generation means generates binary image data corresponding to a density level to be written based on the outline vector for each density level scaled and smoothed by the scaling / smoothing means. 2. The multi-valued image is generated based on the binary image corresponding to each density level reproduced by the binary image reproducing means, having a binary image reproducing means for generating the binary image. Image processing device. 7. The multi-valued image generating means moves a target pixel in the binary image of each density level reproduced by the binary image reproducing means, and if the target pixel is a black pixel, multi-valued image is generated. The image processing apparatus according to claim 6, wherein a pixel having a density level corresponding to the binary image is written in a position corresponding to a pixel of interest in the image. 8. The multi-valued image generating means determines whether all the binary images reproduced by the binary image reproducing means are white and ignores the binary image if all the white pixels are white pixels. The image processing apparatus according to claim 7. 9. The image processing apparatus according to claim 1, further comprising a density smoothing unit that smoothes a density gradient of the multivalued image generated by the multivalued image generating unit. 10. The weighted average calculation means for calculating a weighted average value according to the densities of a pixel of interest and a pixel group in the vicinity of the pixel of interest in the multivalued image generated by the multivalued image generation means. The image processing apparatus according to claim 9, further comprising: a unit and an updating unit that updates the density of the pixel of interest with the calculated weighted average value. 11. The image processing apparatus according to claim 10, wherein the weighted average calculation means calculates the weighted average using a number of neighboring pixels according to the scaling factor. 12. The image processing apparatus according to claim 10, wherein the weighted average value calculation means calculates a weighted average by weighting a pixel closer to the pixel of interest with a greater weight. 13. The image processing apparatus according to claim 10, wherein the weighted average value calculation means calculates a weighted average by uniformly weighting a pixel of interest and a pixel group in the vicinity thereof. 14. A first method in which the number of gradations of the multi-valued image is reduced.
And a means for generating a multi-valued image, and generating a second multi-valued image from which the components of the first multi-valued image have been removed,
Interpolating means for interpolating pixels for the second multi-valued image, wherein the scaling / smoothing means performs scaling / smoothing on the first multi-valued image, and the multi-valued image generating means. The image processing apparatus according to claim 1, wherein the first multivalued image that has been scaled and smoothed is combined with the second multivalued image that has been interpolated by the interpolating means. 15. The multi-valued image generation means is configured to perform the scaling.
There is a contour reproducing means for generating a contour image corresponding to each density level based on the outline vector for each density level scaled and smoothed by the smoothing means, and each density reproduced by the contour reproducing means. The image processing apparatus according to claim 1, wherein a multivalued image is generated based on a contour image corresponding to a level. 16. The multi-valued image generating means scans the contour image and a multi-valued image area region in which the multi-valued image is to be reproduced in synchronization with each other in raster scan order, and intersects with the contour line on the contour image. The image processing apparatus according to claim 15, wherein writing of pixels having a density corresponding to the contour image in the multi-valued image area is started or stopped for each time. 17. A determination means for determining a density level included in the multivalued image is further provided, and the isoconcentration line extraction means extracts an outline vector for a density level not included in the multivalued image. The image processing apparatus according to claim 1, wherein the multi-valued image generation means ignores density levels not included in the multi-valued image. 18. The multi-valued image, wherein the determination means checks the highest density level and the lowest density level included in the multi-valued image, and determines that a density level between them is included in the multi-valued image. 18. The image processing apparatus according to claim 17, wherein the density level included in is determined. 19. The image processing apparatus according to claim 17, wherein the discriminating means examines whether all the density levels that can be included in the multi-valued image are included in the multi-valued image. 20. Further comprising a vector input means for inputting an outline vector for each density level of the multi-valued image,
18. The image processing apparatus according to claim 17, wherein the scaling / smoothing unit scales / smooths the outline vector input by the vector input unit. 21. When the multi-valued image has a predetermined density characteristic, the iso-concentration line vector extraction means extracts an outline vector for each density level of a reverse image of the multi-valued image, and the multi-valued image is extracted. The image processing apparatus according to claim 1, wherein the image generating unit generates a multivalued image which is inverted again when the outline vector for the inverted image is extracted by the isoconcentration line vector extracting unit. 22. An image inverting means for deciding that the multi-valued image has a predetermined density characteristic, and inverting the density of the multi-valued image according to the result, and the multi-valued image being inverted by the image inverting means. In this case, it further comprises a re-inversion means for inverting a multi-valued image, wherein the iso-concentration vector extraction means extracts an equi-density vector from the multi-valued image inverted by the image inversion means, and the re-inversion means comprises: The image processing apparatus according to claim 21, wherein the multi-valued image generated by the multi-valued image generating means is inverted. 23. A density determination means for determining whether or not the multi-valued image has a predetermined density characteristic is further provided, and the isoconcentration line vector extraction means sets each density level as a threshold value in accordance with a result of the density determination means. The outline vector of the inverted binary image is extracted, and the multi-valued image generation means generates the inverted multi-valued image based on the outline vector corresponding to each density level according to the determination result by the determination means. The image processing device according to claim 21, wherein the image processing device is generated. 24. The image processing apparatus according to claim 21, wherein the predetermined density characteristic is that an average density of pixels included in a multi-valued image is higher than a predetermined threshold value. 25. The image processing apparatus according to claim 24, wherein the average density is an average density of all pixels included in the multi-valued image. 26. The average density is calculated from the multivalued image,
The image processing apparatus according to claim 24, wherein the image density is an average density of pixels thinned out at a predetermined ratio. 27. The image processing apparatus according to claim 21, wherein the predetermined density characteristic is that the density included most in a multi-valued image is higher than a predetermined threshold value. 28. The image processing apparatus according to claim 1, further comprising magnification setting means for setting a magnification for the scaling / smoothing means. 29. The image processing apparatus according to claim 1, further comprising a multi-valued image input means for inputting a multi-valued image to the uniform density vector extraction means. 30. The image processing apparatus according to claim 1, further comprising a multi-valued image output means for outputting the multi-valued image reproduced by the multi-valued image reproduction means. 31. An image processing method for executing scaling processing on a multi-valued image, comprising an iso-density vector extraction step of extracting an outline vector for each density level of the multi-valued image, and each extracted A scaling / smoothing step of scaling and smoothing the outline vector for each density level, and a multivalued image is generated based on the outline vector of each density level obtained by the processing of the scaling / smoothing step. An image processing method comprising: a multi-valued image generation step. 32. The isoconcentration line vector extraction step includes a binary image generation step of binarizing the multi-valued image using a predetermined density level as a threshold, and two binary image generation steps of the binary image generation step. 32. An image processing method according to claim 31, further comprising an outline extracting step of extracting an outline vector of the value image. 33. The image processing method according to claim 32, wherein the predetermined density level is all possible density levels of each pixel of the multi-valued image. 34. The image processing method according to claim 32, wherein the predetermined density level is a predetermined density level for every predetermined number of density levels. 35. The outline extracting step extracts a boundary line between a white pixel and a black pixel of a binary image as a vector having a direction according to a positional relationship between the white pixel and the black pixel,
33. The image processing method according to claim 32, wherein in the scaling / smoothing step, the vector is scaled, and an end point of the vector is moved in accordance with a shape formed by the vector to perform smoothing. . 36. The multi-valued image generation step comprises:
There is a binary image reproduction step of generating binary image data corresponding to each density level based on the outline vector for each density level that has been scaled and smoothed in the smoothing step, and the binary image reproduction step is provided. 32. The image processing method according to claim 31, wherein a multi-valued image is generated based on the binary image corresponding to each density level reproduced by. 37. In the multi-valued image generating step, in the binary image of each density level reproduced in the binary image reproducing step, while moving the target pixel, if the target pixel is a black pixel, the multi-valued image is generated. At the position corresponding to the pixel of interest in the image,
The image processing method according to claim 36, wherein pixels of a density level corresponding to the binary image are written. 38. The multi-valued image generating step determines whether or not the binary images reproduced in the binary image reproducing step are all white, and if the pixels are all white pixels, the binary image is ignored. The image processing method according to claim 37. 39. The image processing method according to claim 31, further comprising a density smoothing step of smoothing a density gradient of the multi-valued image generated by the multi-valued image generating step. 40. The weighted average calculation for calculating the weighted average value by the densities of the pixel of interest and the pixel group in the vicinity of the pixel of interest in the multivalued image generated in the multivalued image generation step in the density smoothing step. 40. The image processing method according to claim 39, further comprising a step and an updating step of updating the density of the pixel of interest with the calculated weighted average value. 41. The image processing method according to claim 40, wherein the weighted average calculation step calculates the weighted average using a number of neighboring pixels according to a scaling factor. 42. The image processing method according to claim 40, wherein in the weighted average value calculation step, a pixel closer to the pixel of interest is weighted more to calculate the weighted average. 43. The image processing method according to claim 40, wherein in the weighted average value calculating step, the weighted average is calculated by uniformly weighting the pixel of interest and a pixel group in the vicinity thereof. 44. A first method in which the number of gradations of the multi-valued image is reduced.
Generating a multi-valued image, and generating a second multi-valued image by removing the components of the first multi-valued image from the multi-valued image,
An interpolation step of interpolating pixels for the second multi-valued image, wherein the scaling / smoothing step performs scaling / smoothing on the first multi-valued image, and the multi-valued image generation step. 42. The image processing method according to claim 41, wherein is combined with the first multi-valued image that has been scaled and smoothed and the second multi-valued image that has been interpolated by the interpolation step. 45. The multi-valued image generating step comprises:
There is a contour reproducing step for generating a contour image corresponding to each density level based on the outline vector for each density level that has been scaled and smoothed in the smoothing step, and each density reproduced by the contour reproducing step. The image processing method according to claim 41, wherein a multivalued image is generated based on a contour image corresponding to a level. 46. The multi-valued image generating step scans the contour image and a multi-valued image area region in which the multi-valued image is to be reproduced, in synchronization with each other in raster scan order, and intersects with the contour line on the contour image. 46. The image processing method according to claim 45, wherein writing of pixels having a density corresponding to the contour image in the multi-valued image area is started or stopped every time. 47. The method further comprises a determination step of determining a density level included in the multi-valued image, wherein the isoconcentration line extraction step extracts an outline vector for a density level not included in the multi-valued image. 42. The image processing method according to claim 41, wherein the multi-valued image generation step ignores density levels not included in the multi-valued image. 48. The multi-valued image, wherein the determining step examines the highest density level and the lowest density level included in the multi-valued image and determines that a density level between them is included in the multi-valued image. 48. The image processing method according to claim 47, wherein the density level included in is determined. 49. The image processing method according to claim 47, wherein the determining step checks whether or not all the density levels that can be included in the multi-valued image are included in the multi-valued image. 50. A vector input step of inputting an outline vector for each density level of a multi-valued image,
32. The image processing method according to claim 31, wherein the scaling / smoothing step scales / smooths the outline vector input in the vector input step. 51. The isoconcentration line vector extracting step extracts an outline vector for each density level of an inverted image of the multivalued image when the multivalued image has a predetermined density characteristic, 32. The image processing method according to claim 31, wherein the image generating step generates a multivalued image which is inverted again when the outline vector for the inverted image is extracted in the isoconcentration line vector extracting step. 52. An image inverting step of deciding that the multi-valued image has a predetermined density characteristic and inverting the density of the multi-valued image according to the result, and the multi-valued image being inverted by the image inverting step. In this case, the method further comprises a re-inversion step of inverting a multi-valued image, wherein the iso-concentration vector extraction step extracts an equi-density vector from the multi-valued image inverted by the image inversion step, and the re-inversion step is the The image processing method according to claim 51, wherein the multi-valued image generated in the multi-valued image generating step is inverted. 53. A density determination step of determining whether or not the multi-valued image has a predetermined density characteristic is further provided, and the isoconcentration line vector extraction step sets each density level as a threshold value according to a result of the density determination step. The outline vector of the inverted binary image is extracted, and the multi-valued image generation step generates the inverted multi-valued image based on the outline vector corresponding to each density level according to the determination result of the determination step. The image processing method according to claim 51, wherein the image processing method is generated. 54. The image processing method according to claim 51, wherein the predetermined density characteristic is that an average density of pixels included in a multi-valued image is higher than a predetermined threshold value. 55. The image processing method according to claim 54, wherein the average density is an average density of all pixels included in the multi-valued image. 56. The average density is calculated from the multivalued image,
The image processing method according to claim 54, wherein the average density of pixels thinned out at a predetermined ratio. 57. The image processing method according to claim 51, wherein the predetermined density characteristic is that the density most included in the multi-valued image is higher than a predetermined threshold value. 58. The image processing method according to claim 31, further comprising a magnification setting step of setting a magnification for the scaling / smoothing step. 59. The image processing method according to claim 31, further comprising a multi-valued image input step of inputting a multi-valued image to the iso-density vector extraction step. 60. The image processing method according to claim 31, further comprising a multi-valued image output step of outputting a multi-valued image reproduced by the multi-valued image reproduction step. 61. A computer-readable memory storing a program for executing a scaling process on a multi-valued image, the code of an iso-density vector extraction step for extracting an outline vector for each density level of the multi-valued image. Then, the extracted outline vector for each density level is scaled and smoothed, and the code of the scaling / smoothing process and the outline vector of each density level obtained by the processing of the scaling / smoothing process are added. And a code of a multi-valued image generating step for generating a multi-valued image based on the computer-readable memory.