JPH02141176A - Picture data compression and expansion device - Google Patents

Abstract

PURPOSE:To reduce the deterioration of a picture by applying compression deciding coefficient of an equation of higher degree so that a data string representing graduation is made nearly coincident with values on a curve represented by the equation of higher degree and applying expansion calculating the coefficients and gradation values from the equation of higher degree representing the curve. CONSTITUTION:Values representing gradation of a picture element 24 at the end of a picture element string 12 and values representing the gradation of the picture element 25 at the end of a picture element string 13 are regarded as values of an element next to the picture element 24 and the picture element strings 12, 13 are treated as data strings representing a series of gradation. Then the coordinate and gradation data of the picture elements on the picture element string 31 are expressed by four values of coefficients A-D of a cubic equation. Thus, the values representing the gradation of a picture are expressed by 4 coefficients A-D. Thus, the picture data is compressed so that the value calculated by an equation of N=4/(number of picture elements of a picture) equal to the compression rate N, where 4 is the number of coefficients in a cubic equation.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to compression and expansion of image data of two-dimensional still images.

[Prior art] A conventional image data compression/decompression method was proposed by Df31p et al. (Reference: E, J, Delp and O, PMit
cell, :”engineering Oomampress
ion Using Block Trunca
tionOoding", 工EKE'TranS,
, 00M-27.9 t pI), 1335-1342
(Sep, 1979)) The original image is divided into blocks of size nXn, and the average value X and standard deviation σ of pixel gradation within each block are determined. Using this average value as a threshold, each pixel value within the block is binarized to create a bit plane. On the compression side, the average value, standard deviation, and bit plane are secured as compressed data for each block.

On the decompression side, the gradation of each pixel was determined from the average value, standard deviation, and bit plane according to two formulas.

[Problems to be Solved by the Invention] However, in the above compression/expansion method, the average value, standard deviation, etc. must be determined for each block, and as the blocks are made smaller, the amount of calculation increases and calculation time is required. Furthermore, if the block size is chosen too large, there will be a big difference between the image before compression and the image after expansion. Also,
The compression ratio is not that large (at most, it is only a few tenths of a second).Also, the data after compression may have different characteristics such as block size, average value, standard deviation, etc.
The size of the entire data varies depending on the block size of one image screen, and the communication protocol (commitment regarding communication data) when transmitting compressed data via communication becomes complicated, and the recording format when storing compressed data becomes difficult. When it comes to complexity, there are five problems.

Therefore, the present invention is intended to solve these problems, and its purpose is to compress and expand image data with a high compression ratio and with almost no change in image quality even if the compressed image data is expanded. It's about providing a method.

[Means for Solving the Problems] The image data compression/expansion method of the present invention provides a two-dimensional image with a nearly orthogonal coordinate axis, and sets pixels on the image as pixel rows in arbitrary coordinate axis directions extending from one end of the image to the other, This pixel row is provided from end to end in the direction of the other coordinate axes of the image, and the start end pixel or end pixel of the pixel row adjacent to the aforementioned pixel row is placed at the start end pixel or the end pixel of any pixel row. continue,
Create a data string that represents the gradations of a series of images. This allows the data string representing the gradation value to be compressed to determine the coefficients of the higher-order equation so that it almost matches the value on the curve represented by the higher-order equation, and the coefficients of the higher-order equation representing the aforementioned curve and the higher-order It is characterized by performing expansion by calculating values representing gradations from equations.

[Operation] The value of the position of a pixel constituting a two-dimensional image and the value of the gradation of the pixel have a one-to-one correspondence for each pixel.

According to the present invention, in a two-dimensional image, pixels on the image can be considered as rows of pixels along the vertical or horizontal direction of the image, and the image is constructed by arranging many of these rows. .

Here, one end pixel of two adjacent columns, for example, the end pixel of one column and the end pixel of another column, are regarded as adjacent data, and the columns are connected to form a long column. By doing this for all columns in the same way, the pixel data of the image can be made into one long column. This pixel data string consists of a string of values representing the gradation of each pixel, and it is possible to obtain a curve expressed by a higher-order equation that passes through the vicinity of this value. The coefficients of the higher-order equation representing the curve are determined so that the value representing the gradation on this column almost matches the point on the curve. With this, a group of values representing the gradation of pixels on a column can be expressed as coefficients on a curve. In other words, image data can be replaced as coefficients of a curve, and a large amount of data representing the gradation of each pixel on an image can be replaced as a small number of coefficients.

This allows the number of data representing an image to be reduced and the image to be compressed.

Further, by substituting the compressed data replaced by the coefficients into the same high-order equation as above and restoring the value representing the gradation of the pixel, the image data before compression can be expanded.

[Example] FIG. 1 is a diagram of an image display in an example of the present invention. In this image in which a landscape painting is displayed on a rectangular screen, 11 is set as the image origin, and the Y-axis and Y-axis are determined to be approximately orthogonal to the direction shown in the figure.

This image is composed of a collection of many pixels. Pixels are composed of several hundred to several tens of pixels in each of the X and Y axis directions, and are arranged in a sock-like manner, and each pixel has a gradation, thereby creating an image.

Here, pixels in the X-axis direction are considered as one column. A certain column and an adjacent column are respectively referred to as a pixel column 12 and a pixel column 16. FIG. 2 shows an enlarged view near the ends of the pixel rows 12 and 13. 21 and 22 are pixels. The pixel row 12 is composed of pixel rows arranged in the direction of the arrow 23. Thereby, the pixel string can be regarded as a data string composed of a string of values representing the gradation corresponding to each pixel. Here, pixel column 12
The value representing the 240th gradation of the pixel at the end of pixel row 13 and the value representing the 250th gradation of the pixel at the end of pixel column 13 are regarded as the pixel 240th-order value, and the value representing the continuous gradation of pixel row 12 and pixel column 13 is Treated as a data column.

In this way, at the end of each pixel column, the first pixels or the last pixels of adjacent pixel columns are connected to each other, creating a continuous data string for the entire image. An example of this data string is shown in FIG. 531. The order of the data is alternated as shown by arrows 32 in FIG. As described above, one data string 51 can constitute a set of the position of a pixel and a value representing the gradation corresponding to this pixel.

FIG. 4 shows a graph in which pixel positions are plotted on the horizontal axis and values representing pixel gradations are plotted on the vertical axis. Here, it is assumed that the positional interval between the end pixel 24 of the pixel row 12 shown in FIG. 2 and the end pixel 25 of the adjacent pixel row 13 is the same as that of the pixel row 12020 and the adjacent pixel 22. As a result, the intervals between all pixels in the pixel column 51 become the same. Also, the reason why the end pixel of pixel row 13 is set as the next pixel after the end pixel of pixel row 12 is because the gradation of nearby pixels in the image has less change and is continuous than the gradation of pixels further away. Therefore, there is little discontinuity in the data of values representing nine gradations.

Here, attention is paid to the combination of the pixel position shown in FIG. 4 and the value representing the floor "5". Let the value of the pixel position be X1 and the value of gradation be C. As many pairs of X and C as the number of pixels forming the image are obtained. In order to approximate this set of X and C by a higher-order equation (cubic equation) c = Ax3 + Bx2 + Dx + E, coefficients A, B, D, and E of the sixth-order equation are determined by the method of least squares. FIG. 5 shows the hexagonal curve 51 and points of gradation values corresponding to the coordinate values.

The coordinate values and gradation data of pixels on the pixel row 31 can be expressed by a cubic equation and its four coefficients A, B, O, and D. As a result, the value representing the gradation of the image is A
, B, O, and D.

Thereby, the image data is compressed so that the value calculated by the following equation becomes the compression ratio N.

N = 4 / (number of pixels in the entire image) 4 is the number of coefficients of the cubic equation. Generally, the number of pixels in the entire image ranges from tens of thousands to several million. Therefore, the compression ratio N becomes - from -10,000 to -1,000,000.

Also, in order to restore the image, B is added to the coefficient.

D, E and the three-dimensional equation 0=A°x3-)Bx2+Dx10E, the pixel position coordinate value X1 t
Q is substituted to obtain CI lC2tcs ·OH, and as image data, pixel position X1 is associated with gradation level C0, pixel position x2 is associated with gradation level 02, etc., and the image is restored.

As a result, the compressed image is decompressed and returned to the original image.

[Effects of the Invention] As described above, according to the present invention, an image can be compressed by a simple operation of replacing image data with coefficients of a higher order equation such as a 6th order curve.The compression ratio of this operation is from -10,000 to It was extremely large, measuring 1/1,000,000 parts, and was able to provide an excellent compression method with almost no image deterioration.

Also, the decompression method can be easily done by performing the opposite of the compression method.

By using the image compression and decompression method according to the present invention, it is possible to significantly reduce the amount of communication in image data communication, improve the data transfer speed, and increase the amount of data transferred on the same line, making it possible to support the future advanced information society. You can contribute a lot. Furthermore, the amount of image data stored in the storage medium can be reduced, allowing a large amount of image data to be stored in a storage medium with the same capacity as the current one, contributing to an improvement in the storage capacity of information.

[Brief explanation of the drawing]

FIG. 1 is an image display diagram in an embodiment of the present invention. FIG. 2 is an enlarged view of a pixel row. FIG. 3 is a pixel row diagram. Figure 5 shows the relationship between the quintic curve, the pixel position, and the gradation of each pixel. 21.22, 24.25...Pixel 23.32
・・・・・・Arrow 61・・・・・・・・・Data string 51・・・・・・・・・Sixth order curve or above Applicant Seiko Epson Corporation agent Patent attorney Masayoshi Kamiyanagi (1 other person) ) Figure 3 Figure 4

Claims (1)

[Claims] In a two-dimensional still image data compression/expansion device that has pixel position information and gradation information corresponding to the position information, a coordinate axis that is approximately perpendicular to the two-dimensional image is provided, and pixels on the image can be arbitrarily moved from one end of the image to the other. means for providing a pixel column in the coordinate axis direction, and providing the pixel column from end to end in the other coordinate axis direction of the image; A means for creating a data string representing a series of gradations of an image by continuing pixels or terminal pixels, and a means for creating a data string representing the gradation values so as to substantially match values on a curve expressed by a higher-order equation. An image data compression/expansion apparatus comprising: compression by means of determining coefficients of the higher-order equation; and expansion by means of means for calculating a value representing a gradation from the coefficients of the higher-order equation representing the curve and the higher-order equation.