JPS6046473B2 - Image encoding method - Google Patents

Description

[Detailed description of the invention] The present invention relates to an image encoding method.

2. Description of the Related Art Techniques are currently being actively researched in which an image is regarded as a set of pixels, each pixel is represented by a limited signal level, and this is encoded as a so-called digital image.

Many research results have been published regarding efficient encoding methods under the condition that the original digital image can be accurately reproduced even after the image is encoded. However, depending on the purpose of encoding and storing or transmitting an image, after encoding the image, it may not be possible to accurately reproduce the original digital image, and only an approximation of the original digital image can be reproduced. There are cases where there is no problem even if it gets old. If approximate images are allowed in this way, a more efficient encoding method can be obtained using this acceptance condition, but in conventional encoding methods, introducing the condition that approximate images are allowed The disadvantage is that the encoding process becomes complicated. The present invention eliminates the above-mentioned drawbacks of the conventional methods, and although it cannot accurately reproduce the original digital image, it is capable of reproducing an approximation image that does not pose a problem in practical use. It is an object of the present invention to provide an encoding method that is efficient and relatively simple in encoding processing.

In the encoding method according to the present invention, a predetermined number of mutually distant pixels are treated as one pixel block, an image is considered to be composed of an array of pixel blocks, and the total number of pixel patterns that a pixel block can have (i.e., If the number of signal levels that one pixel can take is N and each pixel block is composed of m pixels, the total number of pixel patterns that a pixel block can take is Nm). the other pixel patterns are represented by approximate pixel patterns among the selected pixel patterns;
This selected pixel pattern is represented by a binary code,
The image is represented by replacing the array of pixel blocks constituting the image with the array of binary codes described above, and the array of binary codes is further divided into arrays of each digit and encoded.

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

FIG. 1 is an explanatory diagram for explaining the principle of the present invention, and does not show an example of an original binary image to be stored or transmitted. In the figure, a small square surrounded by a dotted line and a solid line is one pixel, and a square containing four pixels and surrounded by a solid line is one pixel block. Also, the entire image shown in Figure 1 is 5×5
=2? It consists of an array of pixel blocks. Assuming that the signal level of each pixel is binary: r1 (black, i.e., the diagonally shaded square in Figure 1) and ROJ (white, i.e., the blank square in Figure 1), it is composed of four pixels. The total number of pixel patterns that a pixel block can have is 7=16.

Fig. 2 is an explanatory diagram that does not show an example of the pixel pattern that a pixel block can take, and the pixel block consisting of four pixels has a total number of different types as shown in Fig. 2 a-p. Possible pixel pattern. In the method of this invention, the second
Usable pixel patterns smaller than 16 are selected from the total number of types of pixel patterns shown in the figure, and other pixel patterns are represented by pixel patterns that are similar to the selected pixel patterns.

FIG. 3 is an explanatory diagram showing an example of usable pixel patterns selected from the pixel patterns shown in FIG.
Eight types of pixel patterns shown in i, m, C, g, e, j, and k are selected and shown in Fig. 3 as 0, 1, 2, 3, 4, 5, and 6, respectively.
, 7, and out of the remaining pixel patterns, Fig. 2 B, d
, n is Fig. 3 0, Fig. 2 p is Fig. 3 3, Fig. 2 h is Fig. 3 4, Fig. 2 e
The one shown in FIG. 3 is approximated by 6 in FIG. 3, the one shown in FIG. 2 f is approximated by FIG. 3, and the one shown in FIG. 2 f is approximated by FIG. 3, respectively. FIG. 4 is an explanatory diagram showing an embodiment of the present invention, and is an image in which the binary image shown in FIG. 1 is expressed using an approximation of the relationship between FIGS. 2 and 3. (However, in order to improve the effect of data compression, approximations other than those shown in Figures 2 and 3 are used in one part.) The images in Figure 1 and the images in Figure 4 are different images, but many It can be seen that in many cases the same meaning can be conveyed and therefore the approximation of the relationship between FIGS. 2 and 3 is acceptable in many cases.

In addition, to represent the image shown in FIG. 1, a 4-bit binary code is required for each pixel block, but a 4-bit binary code is required to represent the image shown in FIG. To represent the image shown in the figure, a 3-bit binary code (generally speaking, a 3-digit binary code) is used for each pixel block.
All you have to do is assign it. FIG. 5 is an explanatory diagram showing binary codes assigned to each pixel pattern in FIG. 3, and shows an example in which a three-digit gray code is assigned.

FIG. 6 is an explanatory diagram showing an embodiment of the present invention, in which the pixel pattern of each pixel block in the binary image approximated as shown in FIG. 4 is expressed by binary codes assigned according to FIG. This is an expression.

In the figure, the solid square corresponds to each pixel block, the lower right of the square is the first digit of the binary code, the center is the second digit, and the upper left is the third digit. It can be seen that there are three sets of binary arrays for each digit of the binary code.

Now, the binary image in Figure 1 can also be regarded as a binary array if the white pixel ROJ is represented and the black pixel is represented by 0.
Comparing with the figure, we can see that Figure 1 is approximated by Figure 4 (Figure 2 is approximated by Figure 3), a binary code is assigned to each pixel pattern in Figure 3 according to Figure 5, and It can be seen that in the process of converting into a diagram, the total number of binary symbols is reduced to 314.

Next, the three sets of binary arrays shown in Figure 6 are encoded separately using a conventionally known method.As an example of this encoding, if an encoding format as shown in Figure 7 is used. This will be explained using an example.

That is, if the binary array of each digit in Figure 6 is scanned horizontally (in the drawing) and sequentially read from the uppermost left symbol to the lower right symbol, for example, the third digit is ROOOOOO, OOOlO,
OlOlO, OlOll, lOOOlJ, but the symbol R
The symbol pattern between the divisions immediately after lJ and up to three symbols is converted into a code bit pattern according to the regulations shown in FIG. The result is ROOllOllllllOllllllOlOlOlO
OLOJ, which is a 24-bit code. Also, the second digit is ROOOOOO, OlllO, OOOOO, lOll, O
lllOJ, which is encoded as ROOlOlO
lOOOOlO1111011010100yo becomes 25
This is the sign of the bit. Also, the first digit is ROOlOO, Ol
If you encode it with OOO, OOOOO, llllO, OOOOOj, you get RllllOlOOOOlllllOl
OLOOOOJ, resulting in a 19-bit code.

In other words, the total number of binary symbols in Figure 6 is 75, but when encoded using the rules in Figure 7, it becomes 24 + 25 +
19 = 68 bits of code, compared to the original binary image shown in Figure 1, it is 314 x 68175...0.6
This means that the data can be compressed and encoded to a number of bits of 8. As described above, according to the encoding method of the present invention, the effects of both the first stage of creating an approximate image and the second stage of encoding it are multiplied, which improves the efficiency of encoding. We can expect that.

At the stage of creating an approximate image by limiting the number of pixel patterns, it is sufficient to make the pixel block as large as possible within a range that allows image quality deterioration due to approximation.

For example, if you take a pixel block to a size of 8 x 8 = 6 pixels and limit the number of pixel patterns used to 210 = 1.024, you can compress the data to 10164' - 0.16. . Furthermore, when assigning a binary code to a pixel pattern, the number of code bits may be as small as possible for the next encoding.

To this end, attention should be paid to the correlation between pixel patterns of adjacent pixel blocks in the arrangement of pixel blocks. For example, a pixel pattern with a large proportion of black pixels tends to be adjacent to a pixel pattern with a large proportion of black pixels, and a pattern with a small proportion of black pixels tends to be adjacent to a pattern with a small proportion of black pixels. For example, by arranging usable pixel patterns in the order of the proportion of black pixels and assigning the gray code to this, the arrangement of the binary symbols of each digit can be made into an arrangement suitable for run-length encoding or the like.
By the way, in the above embodiment, the number of signal levels that a pixel can take is 2, the size of a pixel block is 4 pixels, and the number of usable pixel patterns is 8. is a value chosen for convenience of explanation, and it goes without saying that the same process can be performed even if any other numerical value is chosen. Furthermore, it goes without saying that the method of assigning a binary code to a pixel pattern and the method of encoding an array of binary codes for each digit are not limited to the above-described embodiments. Furthermore, in the above-described embodiment, the binary array of each digit shown in FIG. 6 is encoded separately for each digit, but it is also possible to concatenate and encode these digits.
FIG. 8 is a block diagram showing an embodiment of the present invention, and is a block diagram showing the configuration of a transmitting and receiving device when the method of the present invention is applied to image transmission.

In the figure, 11 is an image reading section, 12 is a pixel pattern memory 13 is an approximate pattern selection section, 14 is an encoding section, and 15
1 is a decoding section, 16 is a pixel pattern memory, 17 is an approximate image restoration section, and 18 is an image display section. The image read by the image reading section 11 is compared with available patterns in the pixel pattern memory 12 for each pixel block, and converted into a binary code of the approximate pattern selected by the approximate pattern selection section 13.

After being encoded for each binary symbol array of each digit in the encoding unit 14, the data is returned to a binary code again in the decoding unit 15 via a transmission path. A pixel pattern is read out from the pixel pattern memory 16 in accordance with this binary code, an approximate image is formed in the approximate image restoration section 17, and the image is transmitted by displaying this on the image display section 18. As described above, in the method of the present invention, an image is regarded as an array of pixel blocks made up of a plurality of pixels, and each pixel block is approximated from a limited number of pixel patterns to which a binary code is assigned. It can be realized relatively easily by selecting and assigning the objects and encoding the binary code array formed as a result.
In addition, it is possible to obtain an extremely efficient encoding method.

[Brief explanation of the drawing]

Figure 1 is an explanatory diagram for explaining the principle of this invention, Figure 2
The figure is an explanatory diagram showing an example of a pixel pattern that a pixel block can take, FIG. 3 is an explanatory diagram showing an example of a usable pixel pattern selected from the pixel patterns in FIG. 2, and FIG. FIG. 5 is an explanatory diagram showing an example of FIG. 3, and FIG. 5 is an explanatory diagram showing binary codes assigned to each pixel pattern in FIG.
FIG. 6 is an explanatory diagram showing an embodiment of the present invention, FIG. 7 is an explanatory diagram showing an example of coding, and FIG. 8 is an explanatory diagram showing an embodiment of the invention. In the figure, 11 is an image reading section, 12 is an image pattern memory, 13 is an approximate pattern selection section, and 14 is an encoding section.

Claims (1)

[Claims] 1. A digital image constituted by an array of pixels is regarded as an array of pixel blocks composed of a plurality of pixels, and the plurality of pixels in this pixel block are considered to have one of the signal levels that can be taken in the digital image. a step of predetermining usable pixel patterns whose number is smaller than the total number of pixel patterns that can be displayed by the pixel block when taking an arbitrary signal level; and a step of determining in advance a binary code representing the type of the usable pixel patterns. A step of determining in advance the pixel pattern shown by each pixel block constituting the digital image and the above usable pixel pattern, and comparing the pixel blocks having the same or similar pixel pattern to the usable pixel pattern. a step of allocating the binary code predetermined in advance, and a step of encoding an array corresponding to the digital image of the binary code representing the arrangement of the pixel blocks by resembling the assigned binary code. An image encoding method comprising: