JPS63127671A - Resampling method - Google Patents

Abstract

PURPOSE:To reduce the deterioration in picture quality by deciding the position of a picture element subject to resampling in a television picture coding so that the sum of the square of the length of two diagonal lines of a quadrilateral formed by adjacent resampled picture elements and the square of the sides is minimized. CONSTITUTION:A resampling device consists of a one-picture element register 20 storing a resampled picture element, a resamplec clock controller 21 and a resampled picture element location calculator 22. A scanning line outputted from a sampling device 10 is resampled once per n-scanning lines by using a vertical resampled parameter (n) and a scanning line synchronizing signal 201 and a scanning line synchronizing signal 301 is outputted from the resampled clock controller 21 at that time. Through the constitution above, since the resampled picture element position calculator 22 decreases the difference between the distance of the sides and the length of diagonal lines of a quadrilateral formed by the adjacent resampled picture elements, a resampled picture element 300 outputted from the register 20 has a small difference in the correlation among picture elements.

Description

[Detailed Description of the Invention] Industrial Application Field The present invention relates to still images such as photographic images and television images,
Alternatively, the present invention relates to a resampling method for image encoding that digitizes and stores or transmits moving images.

Conventional technology An image encoding device that suppresses the redundancy of images and reduces the number of codes in order to digitize still images or moving images such as photographic images and television images, and efficiently store or transmit them has a high compression rate. A resampling method is known in which a sampled image is sampled again and the resampled image is encoded by the image encoding device. Grid sampling, line offset sampling, etc. are known as resampling methods. Academic Society Image Engineering Study Group Material IE8
4-50 (1984))).

The positions of the sample points in the grid-like sampling described above are shown in Figure 3 (Fig. 3(b)).The positions of the sampling points in the line offset sampling are shown in Figure 3(b). , the pixels in the shaded area are the pixels to be resampled. In both cases, the number of pixels is reduced by half in both the horizontal and vertical directions by resampling.
This is the case when it is reduced to .

Problems to be Solved by the Invention Now, in the conventional resampling method, the resampling pixel position is fixed, so the resampling may change depending on the percentage of the number of pixels reduced by horizontal and vertical resampling. There are cases where the distance between resampled pixels in an image to be resampled differs greatly between the horizontal and vertical directions. For this reason, the correlation between resampled pixels is different, and encoding methods that use correlation between pixels may reduce coding efficiency, or pixels that have not been resampled may be replaced using correlation between resampled pixels. When an image is reproduced through interpolation, there are problems such as deterioration in the quality of the reproduced image.

The present invention solves the above-mentioned problems of the prior art, and provides a resampling method suitable for using an encoding method or an interpolation method using correlation between pixels.

Means for Solving the Problems The present invention provides a method for reducing the number of pixels of a drifted image according to a given ratio in both the horizontal and vertical directions. The position in the image is minimized by the sum of the square of the difference in length between the two diagonals of the quadrilateral formed by adjacent resampled pixels and the square of the difference in length between the shortest side and the longest side. The above objective is achieved by establishing the following.

According to the above-described configuration, the present invention can perform adjacent resampling in the 1Ll-+i image to be resampled, even if the ratio of the number of pixels to be reduced by resampling is different in the horizontal and vertical directions. A resampling method that is suitable for encoding methods or interpolation methods that use correlation between pixels because the difference in the distance between the sides and diagonal length of the quadrilateral formed by pixels is small, so the difference in correlation between pixels does not become large. It is.

Embodiment Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an apparatus that implements a resampling method in an embodiment of the present invention. In FIG. 1, reference numeral 10 denotes a sampler, which inputs an image signal 10 for each scanning line in synchronization with a scanning line synchronization signal 101 manually inputted to a clock controller 11.
Sample 0. The sampled image signal 200, together with the scanning line synchronization signal 201 and both signal synchronization clocks 202 from the sampling clock control section 11, is manually inputted to a resampling device surrounded by a broken line in FIG. The resampling device comprises a one-pixel register 20 for holding resampled pixel values, a resampling clock controller 21, and a resampling pixel position calculator 22. The resampling pixel position calculator 22 calculates the horizontal resampling parameter m.

From the vertical resampling parameter n, a resampling parameter R that minimizes D expressed by the following formula is calculated. (however,
The horizontal resampling parameter m and the vertical resampling parameter n are respectively horizontal (this image is 17 rn G
This is resampled to 1/n in the vertical direction. In addition, as shown in FIG.
'' is a value indicating the relationship between the pixel positions of the resampled pixels of the resampled scan line. That is, the 2j-1 resampled scan line (''
=1,2,3゜4,...), pixel 1, m+
1, 2m+1,...

pm+1.・・・・・・(p=1.2.3.4...
) are resampled, then in the second resampling scan line, pixels R+1 . m+R+1.2m+R+1. ...
−.

Indicates that p m + R+ 1, . . . are to be resampled. ). -(m-2n)2+(Meng door review-↓
Office;] A) In Equation 2, the first term on the right side is the square of the difference in the length of the diagonal lines of the quadrilateral formed by adjacent resampled pixels, and the second term is the square of the difference in the length of the sides. Indicates the power.

The operation of the above configuration will be explained below.

In the following explanation, the operations of the sampler 10 and the sampling clock control section 11 will be omitted, and the resampling device shown in the dotted line G in FIG. 1 will be mainly described.

First, the scanning line output from the sampler 10 is determined by the vertical resampling parameter n and the scanning line synchronization signal 201.
Resampling is performed once every n scanning lines, and the scanning line synchronization signal 301 at that time is output from the resampling clock controller 21. Further, the resampling clock controller 21 uses the resampling parameter R from the resampling pixel position calculator 22, the scanning line synchronization signal 201 output from the clock controller 11, and Image signal synchronization clock 2
02, and when the 2nd j-1 resampled scan line is output from the sampler 10, pixels 1, m+
1, 2m+ 1,...tprn+
1 t... (p=1, 2, 3...)
are sequentially held in one pixel memory (register), and the value 30
0 along with the resampling synchronization clock 302. On the other hand, when the 21st resampled scanning line is output from the sampler 10, pixels R+ 1 , m + R+
1.

2m+R+1.・-・・-, pm-1-R+1.・
---Cp=1+2.3 .

According to the configuration shown in FIG. 1, the resampling pixel position calculator 2
2, k is calculated based on the above formula, that is, the distance between the sides of the quadrilateral formed by adjacent resampled pixels and the difference in the diagonal length become small, so the resampled pixel value output from register 2o 300 has a small difference in correlation between pixels.

In the description of this embodiment, the horizontal direction start position (initial value) of resampling is assumed to be the first pixel of the image, but it can be easily changed by manually inputting a signal indicating the initial value to the resampling clock controller 21. can be controlled. The same applies to the resampling start scan line position in the vertical direction.

Effects of the Invention As described above, the present invention determines the position of a pixel to be resampled in an image to be resampled using the square of the length of two diagonals of a quadrilateral formed by adjacent resampled pixels. By setting the sum of the squares of the lengths of the sides to be the minimum, the difference in the correlation between pixels will not become large, thereby reducing the deterioration of image quality when performing interpolation using the correlation between pixels. It can be done, and the effects are great.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of a device that implements the resampling method in an embodiment of the present invention, and Fig. 2 shows the position of a resampled pixel by the same device and the position used to calculate the resampled pixel position. Parameters m, n. FIG. 3 is a conceptual diagram showing the relationship between R and R, and FIG. 3 is a diagram showing resampled pixel positions by each conventional resampling method. 10... Sampling device, 11... Sampling clock control unit, 20... Register, 21...
. . . resampling clock controller, 22 . . . resampling pixel position calculator. Figure 1 ■――-■--■-1'+S N e+ Qiao Ku Figure 3 (21) Old raccoon ki but picture 4b) 1 shoe change nail

Claims (1)

[Claims] The position of the pixel to be resampled in the image is calculated by calculating the sum of the squares of the lengths of the two diagonals of the quadrilateral formed by adjacent resampled pixels and the square of the difference between the lengths of the shortest and longest sides. A resampling method that determines the minimum value.