JP2000069278A - Image processing unit, image processing method and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing unit that obtains a resolution- converted image of high image quality by a simple unit configuration without special preprocessing. SOLUTION: A magnification input section 1 inputs magnification and an image input section 2 inputs an image to a 1st interpolation processing section 5 and which applies interpolation processing to the inputted image in a main scanning direction, by using a 1st interpolation method and stores it in an intermediate buffer 7. A 2nd interpolation processing section 9 applies interpolation processing to the image stored in the intermediate buffer 7 in the sub scanning direction, by using a 2nd interpolation method and outputs the processed image to an image output section 11.

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, an image processing method, and a recording medium suitable for converting the resolution of a digital image.

[0002]

2. Description of the Related Art Image resolution conversion processing is a technique required for obtaining an output image of the same size as an input original image when the resolution of an image handled between an input device and an output device is different. Alternatively, this technique is also used in image enlargement / reduction processing. As a method of actually converting the resolution of a multi-valued digital image, a zero-order hold method, a nearest neighbor interpolation method, a bilinear interpolation method, a tertiary convolution interpolation method, and the like are generally known. FIG. 14 and FIG.
FIG. 5 schematically shows the outline of these resolution conversion methods. In these figures, the upper part shows the relationship between the pixels of the original image (referred to as original pixels) and the pixels of the image (referred to as converted pixels) whose resolution has been converted by interpolation processing of the original image, and the lower part shows the conversion. An arithmetic expression for obtaining a pixel value P representing the density or luminance of a pixel is presented.

First, as shown in FIG. 14A, the zero-order hold method is an algorithm in which original pixel data at the upper left of a point (pixel) to be interpolated is used as interpolation data. This is an algorithm in which original pixel data closest to a point to be interpolated is used as interpolation data. In both of these image resolution conversion methods, the image quality after the resolution conversion is not so good, but the algorithm is simple, so that high-speed image resolution conversion processing is possible.

[0004] Next, as shown in FIG.
The next interpolation method is an algorithm for obtaining interpolated data from one-dimensional operation (operation for weighting by a linear function in consideration of the distance to four neighboring pixels to calculate a pixel value) from original pixels at four points around a point to be interpolated. It is. In this method, the image quality after resolution conversion is standard, and the processing speed is neither fast nor slow, and is about the standard.

As shown in FIG. 15, the cubic convolution interpolation method uses a polynomial to approximate a cubic operation (sinc function based on sinx / x) from 16 original pixels around the point to be interpolated. This is an algorithm for calculating interpolation data by using the obtained nonlinear function to calculate the pixel values of the 16 neighboring pixels in the interpolation data. In this method, the image quality after resolution conversion is relatively good, but a considerable processing time is required, and the load on hardware is large.

[0006]

In recent years, printers capable of outputting high-quality multi-valued image data created by a personal computer or a workstation have been developed and supplied to users. Generally, the resolution of a display of a personal computer or a workstation is about 75 to 100 dpi (the number of dots per inch), while the resolution of a printer is about 300 to 1200 dpi. In addition, the resolution of the printer tends to increase year by year in combination with the demand for high quality output. For example, a process of outputting an image displayed on a display with a resolution of 100 dpi by a printer of 400 dpi and converting the resolution to four times in order to make the output image the same size as the image displayed on the display is performed. Need to be applied.

However, when the resolution conversion rate becomes large as described above, if the resolution conversion processing is performed using the above-described conventional zero-order hold method or nearest neighbor interpolation method, a block structure (mosaic-like structure) by interpolation is output image. Appears in Also,
The bilinear interpolation method has a disadvantage that an edge portion where the pixel value of the image data changes sharply in the original image is blurred. Furthermore, in the third-order convolution interpolation method, the degree of blur at the edge portion is slightly improved as compared with the first-order interpolation method.
There are inconveniences such as generation of a pseudo contour in a gradation part where the density (shade) of the image changes gradually.

According to the technique disclosed in Japanese Patent Application Laid-Open No. 7-50752, as a means for improving such a conventional interpolation method, a block of an original image consisting of a pixel to be interpolated and its surrounding pixels is extracted. If an edge exists in the block and the binarized pattern in the block is compared with the pixel position to be interpolated by pattern matching, and there is a matching pattern, the output of the nearest neighbor interpolation method is selected. Otherwise, there has been proposed a method of selecting an output of the bilinear interpolation method.

However, in this method, various processes are required to switch the interpolation method, and it is necessary to obtain interpolation pixels by a plurality of methods in order to obtain one output pixel. Therefore, the processing algorithm becomes complicated, and if the processing is to be performed at a high speed, it is necessary to prepare hardware having appropriate specifications, and the cost increases.

Therefore, the present invention has been made in view of such a conventional problem, and does not require the special pre-processing as described above, and obtains a high-quality resolution-converted image with a simple apparatus configuration. It is an object of the present invention to provide an image processing apparatus, an image processing method, and a recording medium that can perform the processing.

[0011]

In order to achieve the above object, an image processing apparatus according to a first aspect of the present invention is an image processing apparatus for performing an interpolation process on input image data and outputting the result. First interpolation processing means for interpolating along the first scanning direction by an interpolation method and outputting first interpolated image data, and converting the first interpolated image data into the second scanning direction by a second interpolation method And a second interpolation processing means for outputting the second interpolated image data by interpolating along.

According to a second aspect of the present invention, there is provided an image processing apparatus for performing an interpolation process on input image data and outputting the result.
When the number of pixels in the main scanning direction of the input image data is I and the number of pixels in the sub-scanning direction is J, an arrangement pattern of pixels in I row and J column is assumed. When each pixel of one line is interpolated to each interpolation position during the (i + 1) th line, the interpolation of each pixel is performed by the first interpolation method with reference to the peripheral pixels in the main scanning direction at each interpolation position. When interpolating pixels between the j-th column and the (j + 1) -th column of the one-line pixel interpolated by the first interpolation processing means and the first interpolation processing means, And a second interpolation processing means for interpolating pixels by a second interpolation method with reference to existing pixels belonging to one line of pixels.

According to a third aspect of the present invention, in the image processing apparatus of the second aspect, each pixel of one line is interpolated between an arbitrary i-th row and an (i + 1) -th row of the pixel arrangement pattern. When interpolating to the position, third interpolation processing means for performing interpolation of each pixel by a third interpolation method with reference to peripheral pixels in the main scanning direction for each interpolation position, and interpolation by the first interpolation processing means When interpolating a pixel between the j-th column and the (j + 1) -th column of the one line pixel, the fourth pixel is referred to pixels belonging to the one line pixel existing around the interpolated pixel. Fourth interpolation processing means for interpolating pixels by an interpolation method, and the first interpolation processing means for interpolating one line of pixels between the k-th row and the (k + 1) -th row of the pixel arrangement pattern And the second interpolation processing means And further interpolating one line of pixels on another line between the k-th row and the (k + 1) -th row by operating the third interpolation processing means and the fourth interpolation processing means. And means for causing it to operate.

An image processing apparatus according to a fourth aspect of the present invention is the image processing apparatus according to the first to third aspects.
In the image processing apparatus according to the invention, one of the first interpolation method and the second interpolation method performs interpolation of image data by an interpolation operation using a linear function using a pixel value of a referenced pixel. The other is characterized in that interpolation of image data is performed by interpolation using a non-linear function using the pixel value of a referenced pixel.

According to a fifth aspect of the present invention, in the image processing apparatus of the third aspect, one of the first interpolation method and the second interpolation method determines a pixel value of a referenced pixel. Is used to interpolate image data by an interpolation operation using a linear function, and the other is to interpolate image data by an interpolation operation using a non-linear function using the pixel value of a referenced pixel. One of the interpolation method and the fourth interpolation method is for performing interpolation of image data by an interpolation operation using a linear function using the pixel value of the referenced pixel, and the other is for performing interpolation of the referenced pixel. It is characterized in that image data is interpolated by an interpolation operation using a non-linear function using pixel values.

According to a sixth aspect of the present invention, there is provided an image processing apparatus for performing an interpolation process on input image data and outputting the result.
By referring to the pixel values of the surrounding m pixels in the first scanning direction with respect to the interpolation position in the input image data, and further referring to the pixel values of the surrounding n pixels in the second scanning direction with respect to the interpolation position. And means for calculating a pixel value of a pixel at the interpolation position.

An image processing method according to a seventh aspect of the present invention is an image processing method for performing interpolation processing on input image data and outputting the result.
A step of interpolating the input image data along a first scanning direction by a first interpolation method and outputting first interpolated image data; and a step of interpolating the first interpolated image data by a second interpolation method. And outputting the second interpolated image data by interpolating along the scanning direction.

An image processing method according to an eighth aspect of the present invention is an image processing method for performing interpolation processing on input image data and outputting the result.
When the number of pixels in the main scanning direction of the input image data is I and the number of pixels in the sub-scanning direction is J, an arrangement pattern of pixels in I row and J column is assumed. When each pixel of one line is interpolated to each interpolation position during the (i + 1) th line, the interpolation of each pixel is performed by the first interpolation method with reference to the peripheral pixels in the main scanning direction at each interpolation position. When interpolating a pixel between the j-th column and the (j + 1) -th column of the pixels of the one line interpolated by the first process and the first process, the 1 that exists around the interpolated pixel A second step of interpolating pixels by a second interpolation method with reference to pixels belonging to pixels of the line.

A ninth aspect of the present invention is the image processing method according to the eighth aspect, wherein each pixel of one line is interpolated between an arbitrary i-th row and an (i + 1) -th row of the pixel arrangement pattern. When interpolating to a position, a third step of interpolating each pixel by a third interpolation method with reference to peripheral pixels in the main scanning direction with respect to each of the interpolation positions, and the one interpolated by the third step The j-th column of pixels for the line and (j
+1) a fourth process of performing pixel interpolation by a fourth interpolation method with reference to pixels belonging to the one-line pixel existing around the interpolated pixel when interpolating pixels between the columns; In interpolating one line of pixels between the k-th row and the (k + 1) -th row of the pixel arrangement pattern, the first step and the second step are executed,
Further, a fifth step in which the third step and the fourth step are executed when one line of pixels is interpolated to another line between the k-th row and the (k + 1) -th row. It is characterized by the following.

An image processing method according to a tenth aspect of the present invention is the image processing method according to the seventh to seventh aspects.
In an image processing method according to a ninth aspect, one of the first interpolation method and the second interpolation method performs interpolation of image data by an interpolation operation using a linear function using a pixel value of a referenced pixel. The other is to perform interpolation of image data by interpolation using a non-linear function using the pixel value of a referenced pixel.

An image processing method according to an eleventh aspect of the present invention is the image processing method according to the ninth aspect, wherein one of the first interpolation method and the second interpolation method determines a pixel value of a referenced pixel. Is used to interpolate image data by an interpolation operation using a linear function, and the other is to interpolate image data by an interpolation operation using a non-linear function using the pixel value of a referenced pixel. One of the interpolation method and the fourth interpolation method is for performing interpolation of image data by an interpolation operation using a linear function using the pixel value of the referenced pixel, and the other is for performing interpolation of the referenced pixel. It is characterized in that image data is interpolated by an interpolation operation using a non-linear function using pixel values.

According to a twelfth aspect of the present invention, in the image processing method for performing an interpolation process on input image data and outputting the same, a first position is determined with respect to an interpolation position in the input image data.
Calculating the pixel value of the pixel at the interpolation position by referring to the pixel value of the surrounding m pixels in the scanning direction and further referring to the pixel value of the surrounding n pixels in the second scanning direction with respect to the interpolation position It is characterized by having.

A recording medium according to a thirteenth aspect of the present invention is the recording medium according to the seventh to first aspects.
And a program for executing the image processing method according to the second aspect of the present invention.
Various recording media such as ROM, DVD-RAM, and tape media can be used.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION Principle of Embodiment As a premise of an embodiment of the present invention, a basic principle of interpolation processing used in the embodiment will be described. In the following description, it is assumed that the position of each pixel is represented using the two-dimensional orthogonal coordinate system shown in FIG. That is, the upper left coordinate of the image is set as the origin, and the downward direction from this origin is the main scanning direction (i direction).
The right direction is the sub-scanning direction (j direction). In this coordinate system, the coordinates of the pixels of the input image (original image) are represented by (I, J), and the coordinates of the pixels of the output image (resolution-converted image) are represented by (i, j). In this coordinate system, the coordinates of the lower right pixel of the input image are represented by (I _max ,
J _max ), and the coordinates of the lower right pixel of the output image are (i _max , j
_max ). Further, the pixel value of the input image is represented by S (I,
J), and the pixel value of the output image is represented as D (i, j).

FIG. 1 shows the positional relationship between a converted pixel (interpolated pixel) P and original pixels P1, P2,.
As described above, the converted pixel P is composed of four neighboring pixels P1, P2,
P3, the -i direction to the rectangular region surrounded by P4 a _i, + _i direction b _i, are placed in the interpolation positions separated by a distance a _j, + _j direction b _j to -j direction.

In order to determine the pixel value of the converted pixel P at such an interpolation position with respect to the original pixel, as shown in FIG. 2, interpolation processing is performed in the i direction and the j direction separately. First, as shown in FIG. 7A, interpolation processing is performed only in the i direction to determine the pixel values of the converted pixels PA and PB in the i direction. That is, the pixel value of the converted pixel PA is a pixel (referred to as a reference pixel) appropriately selected from neighboring pixels on the same column (i direction) as PA, for example, the pixels P5, P1, P2, and P6.
And the distance to each pixel ((a _i +1),
a _i , b _i , (b _i +1)), and is calculated by a predetermined interpolation operation. Similarly, the pixel value of the conversion pixel PB is
It is calculated by a predetermined interpolation calculation in consideration of the pixel values of the reference pixels P7, P3, P4, and P8 and the distance to each pixel.

Next, as shown in FIG. 2B, interpolation processing is performed only in the j direction with reference to the plurality of converted pixels in the i direction, and the pixel value of the final converted pixel P is calculated. Ask. For example, if two pixels are used as reference pixels,
In consideration of the pixel values of the converted pixels PA and PB, and the converted pixels P and the distances a _j and b _j to these pixels, the converted pixels PA and PB are considered.
The pixel value of the converted pixel P is calculated by an interpolation formula (referred to as a second interpolation formula) different from the interpolation formula used for obtaining the pixel value of PB (referred to as a first interpolation formula).

It should be noted that a linear function can be used for one of the first and second interpolation formulas, and a non-linear function can be used for the other. Here, a linear function is a function f such that f (x + y) = f (x) + f (y) and f (ax) = af (x) are satisfied for variables x, y, and constant a. And a non-linear function refers to a function f in which at least one of these two equations does not hold. In the following embodiment, the value of an interpolated pixel is obtained by using an arithmetic expression by a bilinear interpolation method as a linear function, and the value of an interpolated pixel is obtained by an arithmetic expression by a cubic convolution method as a nonlinear function. However, the present invention is not limited to these, and various functions can be used for both the linear function and the nonlinear function.

Next, a preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings. 2. First embodiment (image processing apparatus A) 2.1. Configuration of Embodiment FIG. 5 shows a configuration of an image processing apparatus A of the present embodiment. The image processing apparatus A inputs a magnification input unit 1 for inputting magnifications B _j , B _j in the main scanning direction and the sub-scanning direction when converting the resolution, and inputs original image data (multi-valued digital image). Input unit 2, an interpolation position determination unit 3 for determining a position to be interpolated, and a first reference pixel selection unit 4 for selecting a reference pixel during the first interpolation processing
And a first interpolation processing unit 5 that performs a first interpolation process, a first interpolation coefficient selection unit 6 that selects an interpolation coefficient during the first interpolation process, and stores data after the first interpolation process. Buffer 7, a second reference pixel selection unit 8 for selecting a reference pixel in the second interpolation process, a second interpolation method unit 9 for performing the second interpolation process, and a second interpolation process. In this case, a second interpolation coefficient selection unit 10 for selecting an interpolation coefficient and an output unit 11 for outputting a final resolution-converted image are configured. In the image processing apparatus A, the interpolation processing in the i direction described in the above operation principle is performed by the first interpolation processing unit 5, and then the interpolation processing in the j direction is performed by the second interpolation processing unit 9.

2.2. Operation of Embodiment 2.2.1. Basic Operation Sequence FIG. 7 shows a basic operation sequence when performing resolution conversion processing in the image processing apparatus A.
First, a magnification B _i in the i direction and a magnification B _{j in the} j direction are input from the magnification input unit 1 (S10). Then, the pixel value data S (I, J) and I _{max of the} original image are input from the image input unit 2.
And the value of _Jmax are input (S11). Then, the coordinates (i _max , j _max ) of the lower right corner in the resolution converted image are calculated by the following equation (S12). i _max = B _i (I _max +1) -1 (1) j _max = B _j (J _max +1) -1 (2) Further, 0 is set as an initial value of i (S1).
3). Then, the interpolation processing is performed line by line in the i-direction, which is the main scanning direction, by the processing after the next step S14.

In this interpolation processing, first, the first interpolation processing unit 5 executes the interpolation processing in the i direction using the first interpolation method, calculates the interpolation data of the i-th row, and
, BUF (J) (J = 0, 1, 2,..., J _max ) (3) is stored (S14). Next, the second interpolation processing unit 9 receives the data stored in the intermediate buffer 7,
The interpolation processing in the j direction is performed using the interpolation method described above, and the data is output as the data on the i-th row of the output image (S15).

Then, the value of i is incremented by 1 (S16), and if the value of i is equal to or less than _imax , the processing of one line in steps S14 and S15 is repeated (S17 → S14 → S15 →.・ ・). Then, when it is determined in step S17 that the value of i is larger than i _max , the process ends. Then, i direction (main scanning direction)
The processing in step S14, which is the interpolation processing for the image, and the processing in step S15, which is the interpolation processing in the j-direction (the sub-scanning direction), will be described separately below.

2.2.2. FIG. 9 shows the processing flow (interpolation calculation process) of step S14 shown in FIG. First, it has a coordinate value I of a peripheral original pixel located in front of the conversion pixel in the i direction, a distance a _{i in} the i direction from the original pixel to the conversion pixel, and a coordinate value I + 1 from the conversion pixel in the i direction. and the distance b _i to the next surrounding original pixels is calculated by the following equation (S10
0). Note that the positional relationship between the original pixel and the converted pixel and I, a
_i, the relationship of b _i is as shown in FIG. 1 already described. _{I = (int) i / B} i ··· (4) a i = (float) (i / B i) -I ··· (5) b i = 1-a i ··· (6) , where , (Int) indicates that the fractional part is represented by a truncated integer value, and (float) indicates that the representation is represented by a floating point.

Next, when 0 is set as the initial value of J (S101), interpolation processing for one line in the i direction is started. First, when I = 0, the following equation BUF (J) = k ₁ (a _i ) × S (0, J) + k ₂ (a _i )
× S (0, J) + k ₃ (b _i ) × S (1, J) + k
The value of the interpolated pixel is calculated for each pixel by ₄ (b _i ) × S (2, J) and stored in the intermediate buffer 7 (S102 → S103). And J
Is incremented by 1 (S104), and J> J
_The processing is repeated until the value reaches _max (S103 → S1).
04 → S105 → S103 → S104 →...), And one line of interpolation data is stored in the intermediate buffer 7.

Here, the values of the interpolation coefficients k _{1 to} k _{4 in} the i direction are calculated by the first interpolation coefficient selector 6 according to the following equation. k ₁ (a _i ) = f ₁ (a _i ) k ₂ (a _i ) = f ₂ (a _i ) k ₃ (b _i ) = f ₂ (b _i ) k ₄ (b _i ) = f ₁ (b _{i) However, f 1 (x) = 4-8} (1 + x) +5 (1 + x) 2 - (1+
x) ³ f ₂ (x) = 1-2x ² + x ³ These interpolation coefficients k _{1 to} k ₄ are coefficients calculated according to the distance in the i direction from the converted pixel to the peripheral original pixel. The function f ₁ (x), f
By storing ₂ (x), the values of the interpolation coefficients k _{1 to} k ₄ can be obtained. However, since the set of (a _i , b _i ) is determined in advance by the magnification B _{i in the} i direction, a coefficient table describing the values of k ₁ to k ₄ corresponding to these (a _i , b _i ) It is desirable to ask for.

Next, when I = I _max -1, the following equation BUF (J) = k ₁ (a _i ) × S (I _max −2, J) + k ₂
(A _i ) × S (I _max− 1, J) + k ₃ (b _i ) × S (I
_max , J) + k ₄ (b _i ) × S (I _max , J) The value of the interpolation pixel is calculated pixel by pixel and stored in the intermediate buffer 7 (S102 → S106 → S107).
Then, the value of J is incremented by 1 (S10
8) Repeat processing is performed until J> _Jmax (S
107 → S108 → S109 → S107 → S108 →
..), One line of interpolation data is stored in the intermediate buffer 7.

When I = I _max , the following equation is used: BUF (J) = k ₁ (a _i ) × S (I _max −1, J) + k ₂
(A _i ) × S (I _max , J) + k ₃ (b _i ) × S (I _max ,
J) + k ₄ (b _i ) × S (I _max , J), the value of the interpolated pixel is calculated for each pixel and stored in the intermediate buffer 7 (S 102 → S 106 → S 110 → S
111). Then, by incrementing the value of J by 1 (S112), repeating the process until J> J _max is performed (S111 → S112 → S113 → S111 → S1
12 →...), The interpolation data for one line is stored in the intermediate buffer 7.

[0038] 0 <I <For I _max, the following equation _{BUF (J) = k 1 (} a i) × S (I-1, J) + k
₂ (a _i ) × S (I, J) + k ₃ (b _i ) × S (I + 1,
J) + k ₄ (b _i ) × S (I + 2, J), the value of the interpolated pixel is calculated pixel by pixel and stored in the intermediate buffer 7 (S 102 → S 106 → S 110 → S
114). Then, by incrementing the value of J by 1 (S115), repeating the process until J> J _max is performed (S114 → S115 → S116 → S114 → S1
15 →...), The interpolation data for one line is stored in the intermediate buffer 7. Thus, the interpolation processing for one line in the main scanning direction is completed, and the processing shifts to the interpolation processing in the sub-scanning direction for the data of one line stored in the intermediate buffer 7.

2.2.3. Fig. 1 Interpolation processing in the sub-scanning direction
0 indicates the processing flow (interpolation calculation process) of step S15 shown in FIG. When 0 is set as the initial value of j (S200), the process of calculating the pixel value of the interpolated pixel one by one in the j direction after step S201 is started.

First, the values of J, a _j and b _j shown in FIG.
It is calculated by the following equation (S201). J = (int) j / B _j (7) a _j = (float) (j / B _j ) -J (8) b _j = 1-a _j (9)

[0041] Then, J = in the case of _{J max, D (I, J} ) = output image pixel value by BUF (J _max) is calculated (S202 → S2
03). On the other hand, if J < _Jmax , D (I, J) = b _j × BUF (J) + a _j × BUF (J +
1) is used to calculate the pixel value of the output image (S202 → S2)
04)

Subsequently, the value of j is incremented by 1 (S205), and the above-mentioned S201 is repeated until j> _jmax.
The processing from S205 to S205 is repeatedly performed, and the interpolation processing in the j direction is completed for the data of one line stored in the intermediate buffer 7, and the creation of the i-th image data of the output image is completed.

2.3. Effects of Embodiment The following effects can be obtained by using the image processing apparatus A according to the present embodiment. (1) In the resolution conversion processing in the image processing apparatus A, processing using a cubic convolution interpolation method having an effect of sharpening an edge portion in a certain direction is performed, and smoothing is performed in a direction different from the direction. The use of bilinear interpolation, which has the effect of (1), can prevent jaggies and blurring at edges when only bilinear interpolation is used, and use only cubic convolution. It is possible to prevent a pseudo contour or the like of a gradation portion which occurs when the gradation portion is used. Therefore, a high-quality output image can be obtained as a whole. (2) The number of reference pixels is reduced by half, and the time required for the resolution conversion processing is greatly reduced as compared with the case where the cubic convolution interpolation method is applied in both the main scanning direction and the sub-scanning direction. (3) Further, since special preprocessing and setting of parameters and the like associated with the resolution conversion processing are unnecessary, the hardware circuit configuration is simplified, and high-speed processing can be performed with an inexpensive device. (4) By judging the nature of the document in advance and applying the cubic convolution interpolation method to the direction with more edges and the primary interpolation method to the other direction, a higher quality output image is obtained. Can be obtained.

3. Second Embodiment (Image Processing Apparatus B) The configuration of an image processing apparatus B of the second embodiment is the same as the configuration of the image processing apparatus A shown in FIG. Also, the basic operation sequence is the same as that shown in FIG. Further, the interpolation processing in the sub-scanning direction is executed based on the flow shown in FIG. Therefore, the interpolation processing in the main scanning direction will be mainly described in detail.

3.1. Interpolation Processing in Main Scanning Direction FIG. 11 shows a processing flow (interpolation calculation process) of step S14 shown in FIG. First, I shown in FIG. 1, a _i, the value of b _i (4) (5) (6) is calculated based on the equation (S300). Next, when 0 is set as the initial value of J (S301), interpolation processing in the main scanning direction is started.

First, when I = I _max , the value of the interpolated pixel is calculated pixel by pixel according to the following equation: BUF (J) = S (I _max , J) (10) and stored in the intermediate buffer 7. (S302 → S303). And J
Is incremented by 1 (S304), and J> J
These processes are repeatedly performed until the value becomes _max (S3
03 → S304 → S305 → S303 → S304 → ...
..), one line of interpolation data is stored in the intermediate buffer 7.

If I <I _max , the following equation BUF (J) = k ₁ (a _i ) × S (I, J) + k ₂ (a _i )
The value of the converted pixel is calculated for each pixel by × S (I + 1, J) and stored in the intermediate buffer 7 (S306). Here, k ₁ and k ₂ are
In the first interpolation coefficient section selection section 6, it is calculated by the following equation. k ₁ (a _i ) = f (a _i ) k ₂ (a _i ) = 1−f (a _i ) where f (x) = (cos (πx) +1) / 2 FIG. x is a graph representing the general form of
The edge part is sharpened by determining the interpolation coefficient using this function. These interpolation coefficients can be tabulated similarly to the interpolation coefficients k _{1 to} k ₄ described in the first embodiment, and can be stored in the first interpolation coefficient selection unit 6.

[0048] Then, the value of J is incremented by 1 (S307), J> J max become until these processes are repeated (S306 → S307 → S308 → S3
06 → S307 →...), One line of interpolation data is stored in the intermediate buffer 7.

3.2. Effects of Embodiment (1) The image processing apparatus B of the present embodiment uses an interpolation method based on a cos function having an effect of sharpening an edge portion, and has a primary function having a smoothing effect in the other direction. By using the interpolation method, it is possible to prevent jaggies and blurring of an edge portion that occur when the resolution conversion processing is performed using only the bilinear interpolation method, and a high-quality image can be obtained. (2) As in the case of the image processing apparatus A, since special pre-processing and setting of parameters and the like are not required for the resolution conversion, the circuit configuration of the hardware is simplified. (3) Higher quality image output by judging the nature of the document in advance and performing resolution conversion processing using the cos function in the direction with many edges and coprimary interpolation in the other direction Images can be obtained.

4. Third embodiment (image processing apparatus C) 4.1. Configuration of Embodiment FIG. 6 shows a configuration of an image processing apparatus C of the present embodiment. This image processing apparatus C inputs a magnification input unit 1 for inputting magnifications B _j , B _j in the main scanning direction and the sub-scanning direction when converting the resolution, and inputs original image data (multi-valued digital image). Input unit 2, an interpolation position determining unit 3 for determining a position to be interpolated, and a first reference pixel selecting unit 4 for selecting a reference pixel in the first interpolation processing
And a first interpolation processing unit 5 for performing a first interpolation process, a first interpolation coefficient selection unit 6 for selecting an interpolation coefficient at the time of the first interpolation process, and storing data during a resolution conversion process. An intermediate buffer 7, a second reference pixel selection unit 8 for selecting a second reference pixel, a second interpolation method unit 9 for performing an interpolation process using a second interpolation method, and a second A second interpolation coefficient selection unit 10 for selecting an interpolation coefficient, an output unit 11 for outputting a final resolution-converted image, and a direction selection unit 12 are provided.

In accordance with the input magnification, the pixel position after the resolution conversion is calculated by the interpolation position determining unit 3, and whether the resolution-converted pixel is on an even-numbered line or an odd-numbered line in the pixel array of the output image. It is determined whether the line is on the line, and a signal indicating the result is sent to the direction selecting unit 12. The direction selection unit 12 selects A when the resolution conversion pixel is on an even line, and selects B when the resolution conversion pixel is on an odd line. When A is selected, the interpolation processing is performed by the first interpolation processing unit 5, the interpolation processing is performed by the second interpolation processing unit 9, and the result is transferred to the output unit 11. On the other hand, when the above B is selected, after the interpolation processing is performed by the second interpolation processing unit 9, the interpolation processing is performed by the first interpolation processing unit 5,
The data is transferred to the output unit 11.

4.2. Operation of Embodiment 4.2.1. Basic Operation Sequence FIG. 8 shows a basic operation sequence of the resolution conversion process in the image processing device C. In the present embodiment, in order to obtain the pixel value of the conversion pixel shown in FIG.
It is assumed that the number of reference pixels in the i direction and the j direction and the interpolation method are changed for each line.

First, the magnification B from the magnification input unit 1 in the i direction
_{The i} and the magnification B _{j in the} j direction are input (S20). Then, the pixel value data S (I,
J) and the value of I _max and J _max are input (S21). Then, the coordinates (i _max , j) of the lower right corner in the resolution-converted image
_max ) is calculated by the equations (1) and (2) (S2
2). Further, 0 is set as the initial value of i (S
23). Then, the interpolation processing is performed line by line in the i direction, which is the main scanning direction, by the processing of the next step S24 and subsequent steps.

In this interpolation processing, when the value of i is an even number, that is, when processing an even-numbered line in the main scanning direction, first, the first interpolation method is used to perform the i-direction interpolation processing. The data of the line is acquired and stored in the intermediate buffer 7 (S25). Next, the data stored in the intermediate buffer 7 is subjected to interpolation processing in the j-direction using the second interpolation method, and is output as data on the i-th row of the output image (S26).

On the other hand, in this interpolation processing, when the value of i is an odd number, that is, when an odd-numbered line is processed in the main scanning direction, first, the interpolation processing in the i-direction is performed using the third interpolation method. , I-th row data is acquired and stored in the intermediate buffer 7 (S27). Next, the data stored in the intermediate buffer 7 is subjected to interpolation processing in the j direction using the fourth interpolation method, and is output as the data on the i-th row of the output image (S28).

When the processing of steps S25 and S26 or the processing of steps S27 and S28 is completed, the value of i is incremented by 1 (S29), and the value of i becomes i
If not more than _max , the processing of steps S25 and S26 (i
Is repeated, the processing of one line by the processing of steps S27 and S28 (when the value of i is odd) is repeatedly performed. Then, in step S30, the value of i is i
_The process ends when it is determined to be larger than _max .
In the present embodiment, the processing of steps S25 and S26 when the value of i is an even number is the same as that of the image processing apparatus A. That is, in step S25, the processes from S100 to S116 shown in FIG. 9 are executed, and in step S26, the processes in steps S200 to S116 shown in FIG.
The processing up to 206 is executed.

Therefore, when the value of i is an odd number, i
Step S27 which is an interpolation process in the direction (main scanning direction)
And the processing in step S28, which is interpolation processing in the j-direction (sub-scanning direction), will be described separately below.

4.2.2. Interpolation Processing in the Main Scanning Direction for Odd Lines The processing in step S27 is executed according to the flow in FIG. As shown in the figure, first, I, a _i, the value of b _i is calculated by (4) to (6) below (S400). Next, an initial value 0 of J is set (S401). Interpolation processing for one line in the main scanning direction is started.

[0059] First, in the case of I = I _max is the value of the interpolation pixel by (10) is calculated pixel by pixel, is stored in the intermediate buffer 7 (S403). Then, the value of J is incremented by 1 (S404), J> J max and until these processes are repeated (S403 → S40
4 → S405 → S403 → S404 →...), And one line of interpolation data is stored in the intermediate buffer 7.

[0060] I <For I _max, the following equation _{BUF (J) = b i ×} S (I, J) + a i × S (I + 1,
J), the value of the converted pixel is calculated pixel by pixel and stored in the intermediate buffer 7 (S406). Then, the value of J is incremented by 1 (S407), J> J max and until these processes are repeated (S406 → S4
07 → S408 → S406 → S407 →...), And one line of interpolation data is stored in the intermediate buffer 7.

4.2.3. Interpolation processing in the sub-scanning direction in the case of an odd line The processing in step S28 is executed according to the flow in FIG. As shown in the figure, first, as the initial value of j, 0
Is set (S500), and the interpolation process in the sub-scanning direction is started. First, the values of J, a _j and b _j are (7) to (9)
It is calculated by an equation (S501). Then, based on the calculated value of J, the pixel value of the interpolation pixel is obtained by the following arithmetic expression.

When J = 0, the following equation is obtained: D (i, j) = k ₁ (a _j ) × BUF (0) + k ₂ (a _j )
× BUF (0) + k ₃ (b _j ) × BUF (1) + k
₄ (b _j ) × BUF (2) is used to calculate the pixel value of the converted pixel (S502 → S5
03).

[0063] J = case of J _max -1, the formula D (i, j) = k 1 (a j) × BUF (J max -2) + k 2
(A _j ) × BUF (J _max −1) + k ₃ (b _j ) × BUF
The pixel value of the converted pixel is calculated by (J _max ) + k ₄ (b _j ) × BUF (J _max ) (S502 → S5)
04 → S505).

[0064] J = For J _max, the following equation D (i, j) = k 1 (a j) × BUF (J max -1) + k 2
(A _j ) × BUF (J _max ) + k ₃ (b _j ) × BUF (J
_max ) + k ₄ (b _j ) × BUF (J _max ) to calculate the pixel value of the converted pixel (S502 → S5)
04 → S506 → S507).

In other cases, D (i, j) = k ₁ (a _j ) × BUF (J−1) + k
₂ (a _j ) × BUF (J) + k ₃ (b _j ) × BUF (J +
1) The pixel value of the converted pixel is calculated by + k ₄ (b _j ) × BUF (J + 2) (S502 → S5)
04 → S506 → S508).

The above steps S503, S505, S5
07 or S508, the value of j is incremented by 1 until J> _Jmax .
The processing of S501 to S509 is repeatedly performed, and the interpolation pixel data of the i-th row is output.

4.3. Effects of Embodiment (1) In this embodiment, when the interpolated pixel corresponds to an even-numbered line of the output image, the third-order convolution interpolation method in the main scanning direction is used.
In the case of an odd-numbered line using bi-primary interpolation in the sub-scanning direction, the result of performing interpolation using bi-primary interpolation in the main scanning direction and tertiary convolution in the sub-scanning direction A biased image that is sharpened only in one direction and smoothed only in another direction is not obtained, and an image that is uniformly sharpened and smoothed in both directions is obtained. (2) Therefore, the occurrence of jaggies and the occurrence of false contours in the gradation portion can be prevented, and a high-quality image can be obtained. (3) In the resolution conversion method of the present embodiment, no special preprocessing or parameters are required, so that the hardware circuit configuration is simplified.

5. Modifications The present invention is not limited to the above embodiment, and various modifications such as the following are possible.

(1) In the first resolution conversion processing in the image processing apparatus A, conversion pixels are obtained by using cubic convolution interpolation with reference to four peripheral pixels in the main scanning direction. The final output pixel is obtained by obtaining the converted pixel in the sub-scanning direction using the bilinear interpolation method with reference to two of the converted pixels in the sub-scanning direction. And the interpolation method used in the sub-scanning direction may be interchanged. That is, first, the primary output method is used to refer to two pixels in the main scanning direction, and then the final output pixels are used to refer to four pixels in the sub-scanning direction and use tertiary convolution interpolation. You may ask. (2) Similarly, in the image processing apparatus B, the interpolation method in the main scanning direction and the interpolation method in the sub-scanning direction may be exchanged to execute the interpolation processing. (3) In the image processing apparatus C, one line Although the two types of interpolation methods in the main scanning direction and the sub-scanning direction are alternately changed every time, the interpolation method may be changed for each of a plurality of rows.
It may be changed randomly for each line. (4) In the interpolation processing of the image processing apparatus C, an interpolation coefficient 1 that refers to a total of eight pixels, i.e., four pixels in the i direction and two pixels in the j direction, and two pixels in the i direction and four in the j direction.
An interpolation coefficient 2 that refers to a total of 8 pixels is determined, and these interpolation coefficients 1 and 2 are set for each predetermined line.
The image may be output while alternately using. (5) In the image processing apparatuses A to C, the interpolation coefficients used in the well-known bilinear interpolation method and the cubic convolution interpolation method are adopted, but the method of determining the interpolation coefficients is not limited to these. (It is possible to change various interpolation coefficients according to the type of the input original image such as a line drawing, an illustration, a natural image, an abstract image, and the pattern or texture of the image). (6) In connection with this, in the image processing apparatuses A to C, when determining the interpolation coefficient in the interpolation in the main scanning direction and the interpolation in the sub-scanning direction, one of the linear functions is used and the other is a non-linear function. Is used, but a function of a non-linear function may be used as long as the effect on interpolation is different. (7) In the interpolation processing in the image processing apparatuses A to C, an intermediate buffer for one line is secured and the interpolation processing is performed line by line in the main scanning direction. It is possible to execute interpolation processing for a plurality of lines at once. Further, it is also possible to directly write the data of the interpolated pixel in an area on the memory reserved for the output image without passing through the intermediate buffer. (8) Although plane-sequential data is assumed as input image data in the image processing apparatuses A to C, not only pixel-sequential data and line-sequential data but also run-length data are converted into plane-sequential data, A resolution conversion process can be performed.

[0070]

According to the present invention, a high-quality output image can be obtained by performing resolution conversion processing with a simple apparatus configuration without requiring special preprocessing.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a positional relationship between a converted pixel and four surrounding original pixels that are referred to when calculating a pixel value of the converted pixel.

FIG. 2 is a diagram showing a pixel position when conversion is performed only in the i direction referred to in the interpolation processing in the i direction and a position in the i direction from a peripheral pixel of a converted pixel.

FIG. 3 shows a coordinate system used to represent a pixel position in the embodiment of the present invention.

FIG. 4 is a diagram schematically illustrating a graph of a function used for obtaining an interpolation coefficient in a second embodiment of the present invention.

FIG. 5 is a block diagram of an image processing apparatus A according to the first embodiment of the present invention and an image processing apparatus B according to the second embodiment.

FIG. 6 is a block diagram of an image processing apparatus C according to a third embodiment of the present invention.

FIG. 7 is a flowchart showing a basic operation of an interpolation process in the image processing apparatus A and the image processing apparatus B;

FIG. 8 is a flowchart showing a basic operation of an interpolation process in the image processing apparatus C;

FIG. 9 is a flowchart showing a flow of an i-direction interpolation process in the image processing apparatus A;

FIG. 10 is a flowchart showing a flow of an interpolation process in the j direction in the image processing apparatus A.

FIG. 11 is a flowchart showing a flow of an i-direction interpolation process in the image processing apparatus B;

FIG. 12 is a flowchart showing a flow of an i-direction interpolation process when an odd line is processed in the interpolation process of the image processing apparatus C;

FIG. 13 is a flowchart illustrating a flow of an interpolation process in the j direction when an odd line is processed in the interpolation process of the image processing apparatus C;

14A is a diagram for explaining an outline of a known zero-order hold method and a nearest neighbor interpolation method, and FIG.
FIG. 3 is a diagram for explaining an outline of a known four-point interpolation method (bilinear interpolation method).

FIG. 15 is a diagram for explaining an outline of a known 16-point interpolation method (third-order convolution interpolation method).

[Explanation of symbols]

 Reference Signs List 1 magnification input unit 2 image input unit 3 interpolation position determination unit 4 first reference pixel selection unit 5 first interpolation processing unit 6 first interpolation coefficient selection unit 7 intermediate buffer 8 second reference pixel selection unit 9 second interpolation processing unit 10 Second interpolation coefficient selection unit 11 Image output unit 12 Direction selection unit

Claims (13)

[Claims] 1. An image processing apparatus for performing interpolation processing on input image data and outputting the same, wherein the input image data is interpolated along a first scanning direction by a first interpolation method to obtain first interpolated image data. A first interpolation processing means for outputting, and a second interpolation method for converting the first interpolated image data into a second
2. Interpolation processing means for interpolating along the scanning direction and outputting second interpolated image data. 2. An image processing apparatus for performing interpolation processing on input image data and outputting the same, wherein the number of pixels in the main scanning direction of the input image data is I, and the number of pixels in the sub-scanning direction is J. Assuming an arrangement pattern of pixels in a column, when interpolating each pixel of one line to each interpolation position between an arbitrary i-th row and an (i + 1) -th row of this pixel arrangement pattern, First interpolation processing means for interpolating each of the pixels by a first interpolation method with reference to peripheral pixels in the scanning direction, and j-th column of pixels of the one line interpolated by the first interpolation processing means When the pixel is interpolated between the (j) and (j + 1) th columns, the second interpolation is performed by referring to the pixels belonging to the one-line pixel existing around the interpolated pixel and by the second interpolation method. Processing means, and An image processing device. 3. The image processing apparatus according to claim 2, wherein each pixel of one line is interpolated at each interpolation position between an arbitrary i-th row and an (i + 1) -th row of the pixel arrangement pattern. Third interpolation processing means for performing interpolation of each pixel by a third interpolation method with reference to peripheral pixels in the main scanning direction with respect to each interpolation position; and the one line interpolated by the first interpolation processing means When the pixel is interpolated between the j-th column and the (j + 1) -th column of the pixel, the pixel interpolating by the fourth interpolation method with reference to the pixels belonging to the one-line pixel existing around the interpolated pixel A first interpolation processing unit and a second interpolation unit for interpolating one line of pixels between the k-th row and the (k + 1) -th row of the pixel arrangement pattern. Activate the processing means Means for operating the third interpolation processing means and the fourth interpolation processing means when interpolating one line of pixels on another line between the k-th row and the (k + 1) -th row; An image processing apparatus comprising: 4. The image processing apparatus according to claim 1, wherein: the first interpolation method and the second interpolation method
Either one performs image data interpolation by a linear function using the pixel value of the referenced pixel, and the other performs image data interpolation by a non-linear function using the pixel value of the referenced pixel. An image processing apparatus for performing interpolation of (1). 5. The image processing apparatus according to claim 3, wherein, among the first interpolation method and the second interpolation method,
Either one performs image data interpolation by a linear function using the pixel value of the referenced pixel, and the other performs image data interpolation by a non-linear function using the pixel value of the referenced pixel. And among the third interpolation method and the fourth interpolation method,
Either one performs image data interpolation by a linear function using the pixel value of the referenced pixel, and the other performs image data interpolation by a non-linear function using the pixel value of the referenced pixel. An image processing apparatus for performing interpolation of (1). 6. An image processing apparatus for performing an interpolation process on input image data and outputting the result, referring to pixel values of peripheral m pixels in a first scanning direction with respect to an interpolation position in the input image data, further comprising: An image processing apparatus, comprising: means for calculating a pixel value of a pixel at the interpolation position by referring to a pixel value of n peripheral pixels in a second scanning direction with respect to the position. 7. An image processing method for performing interpolation processing on input image data and outputting the input image data, wherein the input image data is interpolated along a first scanning direction by a first interpolation method to obtain first interpolated image data. Outputting the first interpolated image data by a second interpolation method.
Outputting a second interpolated image data by interpolating along the scanning direction of the image processing method. 8. An image processing method for performing interpolation processing on input image data and outputting the result, wherein the number of pixels in the main scanning direction of the input image data is I and the number of pixels in the sub-scanning direction is J. Assuming an arrangement pattern of pixels in a column, when interpolating each pixel of one line to each interpolation position between an arbitrary i-th row and an (i + 1) -th row of this pixel arrangement pattern, A first step of interpolating each of the pixels by a first interpolation method with reference to peripheral pixels in the scanning direction, and a j-th column of pixels of the one line interpolated by the first step and (j + 1) A second process of interpolating pixels by a second interpolation method with reference to pixels belonging to the one line of pixels existing around the interpolated pixel when interpolating pixels between columns. An image processing method comprising: 9. The image processing method according to claim 8, wherein each pixel of one line is interpolated at each interpolation position between an arbitrary i-th row and an (i + 1) -th row of the pixel arrangement pattern. A third step of interpolating each of the pixels by a third interpolation method with reference to peripheral pixels in the main scanning direction for each of the interpolation positions, and pixels of the one line interpolated by the first interpolation processing means When the pixel is interpolated between the j-th column and the (j + 1) -th column, the pixel is interpolated by the fourth interpolation method with reference to the pixels belonging to the one-line pixel existing around the interpolated pixel. In a fourth step, when interpolating one line of pixels between the k-th row and the (k + 1) -th row of the pixel arrangement pattern, the first step and the second step are executed, Further, the k-th row and (k + 1) In interpolating one line of pixels on another line between the eyes, the image processing method characterized by having a a fifth step of the third process and the fourth process is executed. 10. The image processing method according to claim 7, wherein, among the first interpolation method and the second interpolation method,
Either one performs image data interpolation by a linear function using the pixel value of the referenced pixel, and the other performs image data interpolation by a non-linear function using the pixel value of the referenced pixel. An image processing method for performing interpolation of 11. The image processing method according to claim 9, wherein, among the first interpolation method and the second interpolation method,
Either one performs image data interpolation by a linear function using the pixel value of the referenced pixel, and the other performs image data interpolation by a non-linear function using the pixel value of the referenced pixel. And among the third interpolation method and the fourth interpolation method,
Either one performs image data interpolation by a linear function using the pixel value of the referenced pixel, and the other performs image data interpolation by a non-linear function using the pixel value of the referenced pixel. An image processing method for performing interpolation of 12. An image processing method for performing interpolation processing on input image data and outputting the result, wherein a pixel value of peripheral m pixels in a first scanning direction is referred to for an interpolation position in the input image data, and the interpolation is further performed. An image processing method comprising: calculating a pixel value of a pixel at an interpolation position by referring to a pixel value of n pixels in a vicinity in a second scanning direction with respect to the position. 13. A recording medium on which a program for executing the image processing method according to claim 7 is recorded.