JPH04330858A - Digital picture enlarging and reducing method and device - Google Patents

Abstract

PURPOSE:To solve the deterioration of a picture due to the jaggy of the oblique line of an interpolated picture at the time of reducing a digital picture, and the blur of the picture at the time of enlarging the digital picture. CONSTITUTION:At the time of reducing the digital picture, the picture is interpolated by an optimal interpolating filter prepared by using a tertiary spline interpolation for a low-pass filter designed by an FIR digital filter method, according to a reduction ratio, so that the reduced picture can be prepared. And also, at the time of enlarging the digital picture, the interpolation of a picture signal is operated by using the optimal interpolating filter prepared by using the tertiary spline interpolation for a high-pass emphasizing filter designed by the FIR digital filter method, so that the enlarged picture can be prepared.

Description

[Detailed description of the invention]

0001

[Field of Industrial Application] The present invention relates to a method and apparatus for enlarging/reducing digital images, and more particularly to a method newly designed as an interpolation filter for enlarging/reducing digital images, and an apparatus using the same. be.

0002

[Prior Art] Conventionally, image enlargement/reduction was performed as shown in FIGS. 9 and 10.
This has been done using density interpolation using coordinate transformation of a digital image, as shown in Figure 1. Generally, the coordinates of an enlarged/reduced image are not grid points in the coordinate system of the original image, so the density of each pixel in the enlarged/reduced image is determined by using the density at the surrounding grid points. need to be interpolated.

As conventional interpolation methods, the following three methods have been mainly used. One of them is the nearest neighbor method, which uses the density value of the nearest grid point in the coordinate system of the original image as the density value of the pixel to be interpolated. Second, the linear interpolation method uses linear interpolation of the density values of the four nearest grid points as the density value of the pixel to be interpolated. Third, the cubic interpolation method uses an approximation formula for the sinc function that appears in the continuous signal sampling theorem, and performs interpolation using the cubic formula using density values at grid points around the pixel to be interpolated. . The interpolation formula used at this time was mainly ``cubic convolution.'' The formula is as shown in equation 1.

[0004] Number 1

[Math 1]

However, in such conventional digital image enlarging/reducing methods and devices, the sampling frequency decreases during reduction, and the image signal contains frequencies that cannot be reproduced, resulting in aliasing errors (
This causes problems such as jaggies becoming noticeable in diagonal lines of reduced pixels due to aliasing (aliasing phenomenon), resulting in poor quality. Furthermore, even though the sampling interval becomes smaller during enlargement and it is now possible to reproduce higher frequency components, the original image does not contain these high frequency components and the frequency components do not change. Therefore, there is a problem that as the image size increases, the image appears blurred.

[0006]

SUMMARY OF THE INVENTION The present invention solves the above-mentioned problems that occur in conventionally known methods and devices for enlarging/reducing digital images, that is, diagonal lines in interpolated images during reduction. This was done to solve image deterioration caused by jaggies and image blurring during enlargement.

[0007]

[Means for Solving the Problems] As a result of intensive studies to solve the above-mentioned problems, the present inventors have designed a filter using the FIR digital filter method for enlarging and reducing digital images, and have designed a filter using the FIR digital filter method. It has been discovered that the above drawbacks can be solved by a method for adaptively enlarging/reducing a digital image, which is characterized by using an optimal interpolation filter created using spline interpolation, and an apparatus using the same.

That is, in the present invention, in the case of digital image reduction, an optimal low-pass filter created using cubic spline interpolation is Interpolation is performed using an interpolation filter to generate a reduced image.

[0009] Furthermore, in the case of enlarging a digital image,
An enlarged image is generated by interpolating an image signal using an optimal interpolation filter created using cubic spline interpolation for a high-frequency emphasis filter designed using the FIR digital filter method.

0010

[Operation] "Reducing" an image means that the sampling frequency included in the image becomes smaller, and therefore, high-frequency frequency components that cannot be reproduced in the image signal are included. This results in image quality deterioration such as aliasing errors. Therefore, in the digital image reduction method of the present invention, the original image signal is passed through an optimal low-pass filter according to the reduction ratio.
This method removes unreproducible high-frequency components as seen from the sampling theorem, and performs interpolation using cubic spline interpolation corresponding to the impulse response of this image signal to generate a reduced image.

[0011] Furthermore, with respect to enlarging, it is possible to reproduce more information than the original image by making the sampling interval smaller, but if the original image signal is simply enlarged, the image quality becomes blurred. Therefore, in the digital image enlarging method according to the present invention, the original image signal is passed through a high-frequency emphasis filter to emphasize high frequency components to give sharpness, and interpolation is performed using cubic spline interpolation corresponding to the impulse response of this image signal. This method generates an enlarged image.

[0012] Incidentally, when the order of the interpolation filter is increased, the performance of the filter is improved, but the amount of calculation increases and the processing time becomes longer. Therefore, it is preferable to determine the order of the filter by considering a trade-off between processing time and filter performance that affects image quality.

[0013]

[Examples] The present invention will be explained in more detail with reference to Examples below, but the present invention is not limited by these Examples.

The method and apparatus for enlarging/reducing a digital image according to the present invention will be explained based on the drawings. 1 and 2 are block diagrams of the present invention.

In the case of image reduction, a low-pass filter is first used, as shown in FIG. Ideally, it is desired that this low-pass filter has characteristics as shown in FIG. 3, but since this is not realistically possible, this is approximated by an FIR digital filter. The DC gain (
Standardize the coefficients (impulse response) of the digital filter so that the amplitude value at frequency 0 becomes 1, and find a cubic spline function to interpolate the impulse response of the filter obtained as a result. asked for 3
A reduction interpolation filter as shown in FIGS. 4 and 5 is designed and generated using the next-order spline function. Thereafter, the image signal is filtered using this reduction interpolation filter to generate an actual reduced image. Furthermore, the image quality is checked, the desired frequency characteristics are changed according to the image quality, and the optimal interpolation filter is determined by feedback.

[0016] As a result, high frequency components that cannot be reproduced as seen from the sampling theorem due to sampling at a longer sampling interval than the original image are removed, so jags and the like due to aliasing errors (aliasing phenomenon) are removed. Image quality deterioration can be improved.

In the case of image enlargement, as shown in FIG. 2, a high-frequency emphasis filter is first used. Ideally, this high-frequency emphasizing filter is desired to have characteristics as shown in FIG. 6, but since this is not realistically possible, this is approximated by an FIR digital filter. Using the required enlargement ratio as a guide, the coefficients (impulse response) of the digital filter are normalized so that the DC gain (amplitude value at frequency 0) is 1 so that the tone of the original image does not change. , a cubic spline function for interpolating the impulse response of the filter obtained as a result is obtained, and the obtained cubic spline function is used to design and generate an expansion interpolation filter as shown in FIGS. 7 and 8. After that, the image signal is filtered using this enlargement interpolation filter,
Generate the actual enlarged image. Furthermore, as in the case of the reduced image, the image quality of the image is checked, the desired frequency characteristics are changed according to the image quality, and the optimal interpolation filter is determined by feedback.

[0018] As a result, the high frequency components of the image signal are emphasized and sharpened, so that blurring of the image quality can be eliminated.

Note that the low-pass filter shown in FIG. 5 is designed as a ninth-order FIR digital filter, and the high-pass emphasis filter shown in FIG. 8 is designed as a third-order FIR digital filter. Examples of interpolation equations used here include equations 2 and 3.

In the case of "reduction" (50% reduction, low pass
(using a filter) The interpolation function in this case is as shown in equation 2,
It is used in place of C(x) in Equation 1 above.

[0021] Number 2

[Math 2]

In the case of "enlargement" (one of the following filters can also be used at any enlargement rate).The interpolation function in this case is as shown in Equation 3.

[0023] Number 3

[Math 3]

[0024]

[Effects of the Invention] With the above-described structure, the present invention can prevent image quality deterioration such as jags due to aliasing errors (aliasing phenomenon) when reducing an image, and can provide a highly preferable digital reduced image. When enlarging an image, blurring of the image quality can be eliminated, and a high-quality digital enlarged image can be provided.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a digital image reduction device according to the present invention.

FIG. 2 is a block diagram showing the configuration of a digital image enlarging device according to the present invention.

FIG. 3 is a characteristic diagram showing the characteristics of an ideal low-pass filter for reducing digital images.

FIG. 4 is a characteristic diagram showing a kernel of a reduction interpolation filter in the digital image reduction method according to the present invention.

FIG. 5 is a characteristic diagram showing the amplitude response characteristics of the reduction interpolation filter of the digital image reduction method according to the present invention.

FIG. 6 is a characteristic diagram showing the characteristics of an ideal high-frequency emphasis filter for enlarging a digital image.

FIG. 7 is a characteristic diagram showing the kernel of the enlarging interpolation filter of the digital image enlarging method according to the present invention.

FIG. 8 is a characteristic diagram showing the amplitude response characteristics of the enlarging interpolation filter of the digital image enlarging method according to the present invention.

FIG. 9 is a schematic diagram showing coordinate transformation of an image in enlarging a digital image.

FIG. 10 is a schematic diagram showing coordinate transformation of an image in reducing a digital image.

Claims (6)

[Claims] Claim 1. A method for enlarging or reducing a digital image, characterized in that an optimal interpolation filter created using cubic spline interpolation is used for a filter designed by the FIR digital filter method. How to enlarge or reduce digital images. Claim 2: A method for enlarging a digital image, comprising:
An adaptive digital image enlarging method characterized by using an optimal enlarging interpolation filter created using cubic spline interpolation with respect to a high-frequency emphasis filter designed by the FIR digital filter method. Claim 3: A method for reducing a digital image, comprising:
An adaptive digital image reduction method characterized by using an optimal reduction interpolation filter created using cubic spline interpolation for a low-pass filter designed by the FIR digital filter method. Claim 4: A digital image enlarging device, comprising:
An adaptive digital image enlarging device comprising an optimal enlarging interpolation filter created using cubic spline interpolation for a high-frequency emphasis filter designed by the FIR digital filter method. 5. A digital image reduction device, comprising:
An adaptive digital image reduction device comprising an optimal reduction interpolation filter created using cubic spline interpolation for a low-pass filter designed by the FIR digital filter method. 6. A digital image enlarging/reducing device, comprising: an optimal reduction interpolation filter created using cubic spline interpolation for a low-pass filter designed by the FIR digital filter method; and an FIR digital filter. 1. An adaptive digital image enlarging/reducing device comprising: an optimal enlarging interpolation filter created using cubic spline interpolation for a high-frequency emphasis filter designed by the method.