JPH03266564A - Method and device for encoding picture - Google Patents

Abstract

PURPOSE:To raise the compression rate and to keep a high picture quality by selecting an encoding order of high compression rate in accordance with the distribution of picture information to be subjected to orthogonal conversion to encode information and using compressed and encoded data and data, which discriminates the encoding order, as encoded data of picture information. CONSTITUTION:The coefficient of a quantized DCT(discrete cosine transform) of each block is inputted to a scan table selecting part 16 to select a scan table. The code outputted from the scan table selecting part 16 is sent to a scan table 18 and a multiplexing part 21 and is multiplexed and is outputted to an output terminal 22. Thus, picture information is properly encoded though components after orthogonal conversion are not concentracted to a specific part, and the compression rate of the whole of the picture is raised and a high picture quality is kept.

Description

Detailed Description of the Invention [Field of Industrial Application] The present invention relates to an image encoding method, particularly an image encoding method that performs data compression using orthogonal transformation, which is used in the field of digital processing devices for image signals, and It is related to the device.

[Prior Art] In a conventional image encoding method using orthogonal transform, a method is known in which data after orthogonal transform is quantized and then the quantized data in a block is encoded in a predetermined order. ing.

Hereinafter, a conventional example will be explained using FIG. 5. The image data inputted from the input terminal 51 is divided into blocks of 8×8 pixels by the blocking means 52, and then divided into blocks of 8×8 pixels.
3 to D CT (Discrete Cosine Tr
(ansform) processing is performed. Each DCT coefficient is
Quantization means 54 according to quantization matrix table 55
quantized by . An example of data quantized by the quantization means 54 is shown in FIG.

In FIG. 6, the right side is the higher frequency component in the horizontal direction of the block, and the lower side is the higher frequency component in the vertical direction. Due to the nature of the image, frequency components are biased towards the low frequency side, so generally the components are concentrated near the shaded area in FIG. 8(a), and most of the other coefficients are O.

In the intra-block scanning section 56, the scan table 5
7, rearrange the block data into one column. A conventional example of the scan table 57 is shown in FIG. 7(a). Each number in FIG. 7(a) represents the coding order of the coefficients in the corresponding position. FIG. 7(b) shows the encoding order using arrows. The data rearranged into one column is efficiently encoded by entropy encoding 58. Specifically, the first data of each block is a DC component, and the difference value from other blocks is Huffman encoded. The second and subsequent components are AC components, which are sequentially encoded by a combination of a non-O coefficient and the number of 0s preceding it. The coefficient values are in the variable length encoding table 59
The length of the variable length code and the number of O's are 2.
Huffman encoding is performed using a combination of two.

Note that this Huffman code contains E OB (End
ofB 1ock) is also included. The encoded data is output from the output terminal 60.

As mentioned above, the data after orthogonal transformation and quantization is generally concentrated in one part as shown in Figure 8(a), whereas the data in Figure 7(b) By performing entropy encoding in order from the part where the components are concentrated, it is possible to perform efficient encoding on the component with a coefficient of O, and it is possible to increase the compression rate of the entire image.

[Problems to be Solved by the Invention] However, in the conventional example described above, in blocks that include lines or contours among the divided blocks, the coefficients after quantization do not necessarily have the distribution shown in FIG. 8(a). However, such a block requires a large amount of encoding. As a result, the characters.

This method has the disadvantage that the compression ratio decreases for images that contain many outlines.

In addition, when processing to keep the compression rate constant by making the quantization step size variable, etc., the image quality of images containing many lines and contours was degraded compared to general images. .

The present invention eliminates the above-mentioned conventional drawbacks and provides an image encoding method that increases the compression rate and maintains high image quality for all images.

In particular, the present invention provides an image encoding method that increases the compression rate and maintains high image quality for images that include many lines and contours.

The present invention also provides an image encoding device that implements the above method.

[Means for Solving the Problem] In order to solve this problem, the image encoding method of the present invention is an image encoding method that compresses and encodes orthogonally transformed image information, which comprises: According to the distribution of the image information, a coding order with a high compression ratio is selected and encoded, and the compression-encoded data and data identifying the coding order are used as coded data of the image information.

Further, the image encoding device of the present invention is an image encoding device that compresses and encodes image information for each block consisting of a plurality of pixels, and includes orthogonal transformation means that performs orthogonal transformation for each block, and the orthogonal transformation. coding order selection means for selecting the coding order of the data in each block based on the distribution of the data in each block; and encoding means for encoding data.

Furthermore, a second encoding means for encoding the information on the encoding order is provided.

[Operation] In such a configuration, attention is paid to the distribution of coefficients having a value of O among the coefficients after orthogonal transformation, the coding order of the coefficients is adaptively changed with respect to this distribution, and a code representing the coding order is added. do. By providing such a means, it is possible to perform appropriate encoding even if the components after orthogonal transformation do not have a distribution that concentrates in a specific part, increasing the compression rate of the entire image. .

[Example] Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1A is a block diagram of an image encoding apparatus according to an embodiment of the present invention, and FIG. 1B is a diagram showing an example of the hardware configuration of the image encoding apparatus according to this embodiment.

First, an example of the hardware configuration of the image encoding apparatus of this embodiment will be described with reference to FIG. 1B.

100 is C for calculation and control to control the image encoding device.
In the PU, registers 101 and 1 that store numerical values j and j representing the data distribution after orthogonal transformation for selecting the encoding order.
02.

200 is a ROM that stores the program for control procedure 1 of the CPU 100; 300 is a RAM that stores calculation aids or each table;
1. It has a variable length encoding table 302, scan tables 303A, 303B, . . . Furthermore, i, j
may be stored on the RAM 300, or each table may be stored on the ROM 200. 400 is a CC that inputs unencoded image data to this device.
D is an image data input unit such as an input interface from a sensor or a computer, and 500 is an encoded data output unit that outputs encoded data compressed and encoded by this device.6 The encoded data output here is stored in a memory. or transmitted via communication means.

Next, the operation procedure of the image encoding apparatus of this embodiment will be explained in detail with reference to FIG. 1A.

Image data input from the input terminal 11 is divided into blocks of 8×8 pixels by the blocking means 12.

The data of each block is subjected to DCT by the DCT section 13, and the 64 DCT coefficients are quantized by the quantization means 14 according to the quantization matrix table 15.

The 64 quantized coefficients of each block are scanned and
The data is input to the table selection section 16, and a scan table is selected.

In the scan table selection section 16, a scan table is selected according to the flowchart shown in FIG. In FIG. 2, first, in step S21, two numerical values j and i representing the characteristics of each block are determined. As shown in FIG. 3, j is a value representing how many columns from the left there are coefficients other than O, and i is a value representing how many rows from the top there are coefficients. In step S22, j and i
If the absolute value 1j-if of the difference is less than the threshold Th, go to step S24, and if it is more than the threshold Th, go to step S23.
Branch to .

In this embodiment, Th=3, and when the process branches to step S24, Code=O. When branching to step S23, if J is not 1, branching to step S25 and setting Code=j; if j<i, step S26
Proceed to and set Code=i + 10.

This code is Huffman encoded in step 5207. When the original image in a block includes vertical edges or horizontal edges, the distribution of non-O components of the coefficients after quantization will be as shown in Fig. 8 (b) or (C), and Fig. 2 Then, the process branches to step 5206 or step 5205, respectively, and Code=i+10 or Code=j.

The code output from the scan table selection section 16 in FIG. 1A is sent to the scan table 18 and the multiplexing section 21. Scan table 18 outputs a table corresponding to the sent symbols. For example, if a code corresponding to Code=O is sent, the scan table shown in FIG. 7(a) is selected, and if a code corresponding to Code=3 is sent, the scan table shown in FIG. 4(a) is selected. ) is selected. Further, when a code corresponding to Code=12 is sent, the scan table shown in FIG. 4(c) is selected. Figures 7(b) and 4(b) and (d) illustrate the respective encoding orders.
It is. In this example, the order within the scan range is the seventh
It is the same as the figure.

Based on the table output from the scan table 18, the intra-block scan section 17 rearranges the data into one column. The data rearranged into one column is encoded by entropy encoding means 19. The entropy encoding method is the same as in the conventional example, and the second and subsequent coefficients of the block are encoded using a combination of a non-O coefficient and the number of O's preceding it. Coefficients other than O are variable-length coded based on the variable-length coding table 20, and Huffman coding is performed by combining the code length of the variable-length code and the number of O's.

If the code length of the variable-length code is determined for the code of the entire image to be subjected to Huffman encoding, and a histogram is taken for the number of O's, the distribution in this embodiment is as shown in FIG. 9(b). On the other hand, in the conventional example, the distribution is as shown in FIG. 9(a). Since the total number of codes is the same, it can be seen that the distribution as shown in FIG. 9(b) of this embodiment reduces entropy and performs efficient encoding.

The output from the entropy encoding means 19 and the output from the scan table selection section 16 are multiplexed by a multiplexing section 21 and output to an output terminal 22.

In this embodiment, the multiplexer 21 performs scanning and
A Huffman-encoded code for selecting table 18 is added to the front of each block of data as shown in Figure 10(a), but for example, it is added all at once as shown in Figure 10(b). You may do so. Furthermore, when adding each block, Huffman encoding may not be performed, and a mere selection signal of predetermined bits (2 bits) may be used.

Further, in this embodiment, the threshold value Th is set to 3, but other values may be used. That is, the method of detecting the characteristics of a block is not limited to the above embodiment.

Further, the scan range of each table in the scan table 18 may be set in other forms as shown in FIGS. 11(a) to 11(d), for example.

Tables corresponding to FIGS. 1I (a) to (d) are shown in FIGS. 12(a) to (d).

Alternatively, the scan table 18 may not be stored in memory, and the table may be generated each time a scan method is determined. Further, although DCT is used for orthogonal transformation, similar effects can be obtained by using Hadamard transformation or other transformation methods. Furthermore, it may be combined with means for keeping the compression rate constant (fixed length encoding means). Also, the size of the block to be handled is 16X
The number of pixels may be 16 pixels or other number of pixels, and the shape of the block may also be other shapes.

Furthermore, although the encoding order within the scanning range was in accordance with the scan table shown in FIG. 7 in this embodiment, another encoding order is selected depending on the nature of the image.

The above-described image encoding method can be applied to both still images and moving images. For example, in addition to still image processing devices such as image file systems and storage devices such as still video cameras, image recording devices such as copying machines, and image transmission devices such as facsimiles and video telephones, as well as still image processing devices such as video cameras and video telephones that handle moving images, etc. The present invention is also effective for moving image processing devices.

Furthermore, the present invention is applicable not only to black and white images (but also to color images. For example, the present invention can be applied to color images (R, G,
B), (Y, M, C'), (X, Y.

2) and (Y, I, Q), (L”, “, b”), (L
, u, v), (L, R, B), etc., may be encoded using the encoding algorithm of the present invention. In particular, (Y.

When the luminance signal and chromaticity signal are separated as in I, Q), even if the chromaticity signal is compressed to a smaller amount of code, the image deterioration will not be noticeable to the human eye. do not have.

Furthermore, entropy encoding is not limited to Huffman encoding, and may be MH encoding, MMR encoding, arithmetic encoding, or the like. Furthermore, the encoding is not limited to entropy encoding, and may be run-length encoding or the like.

As explained above, by providing a means to adaptively change the encoding order for data of each block that has undergone orthogonal transformation and quantization, it is possible to adapt even images containing many characters, line 1 outlines, etc. High compression ratio can be obtained.

Furthermore, when used in combination with means for keeping the compression rate constant, higher image quality can be achieved compared to conventional methods with the same compression rate, especially in the above-mentioned images.

[Effects of the Invention] According to the present invention, it is possible to provide an image encoding method that increases the compression rate and maintains high image quality for all images.

In particular, it is possible to provide an image encoding method that increases the compression rate and maintains high image quality for images that include many lines and contours.

Furthermore, it is possible to provide an image encoding device that implements the above method.

[Brief explanation of drawings]

FIG. 1A is a block diagram of an image encoding device according to an embodiment of the present invention, and FIG. 1B is a block diagram showing an example of the hardware configuration of the image encoding device of this embodiment. is a flowchart showing the processing procedure of the scan table selection section; FIG. 3 is a diagram showing the correspondence between quantized data and j and iTo; FIGS. 4(a) and (C) are scan tables of this embodiment. 4(b) and 4(d) are diagrams showing the order of FIGS. 4(a) and 4(c) with arrows. FIG. 5 is a block diagram of a conventional image encoding device. Figure 6 shows an example of data after quantization, Figure 7 (a
) is a diagram showing the conventional scan table, Figure 7(b) is a diagram showing the order of Figure 7(a), and Figures 8(a), (b), and (c) are after quantization. O of the data of
Figure 9 (a) and (b) are diagrams showing the distribution of coefficients that are not
Histogram showing the frequency with respect to the number of items, Figure 10 (a), (b) is a diagram showing an example of the structure of the encoded data string, Figure 11 (a), (b), (c), (d)
Figure 12 (a) is a diagram showing another example of the scan range.
) to (d) are diagrams showing examples of scan tables corresponding to FIGS. 11(a) to (d). In the figure, 11... input terminal, 12... blocking means, 13... DCT unit, 14... quantizing means, 15...
... Quantization matrix table, 16... Scan table selection section, 17... Intra-block scan section,
18... Scan table, 19... Entropy encoding means, 20... Variable length encoding table, 21
... Multiplexing section, 22 ... Output terminal, 100 ・CP
U, 101 ・L/register, 102...Register j, 200...ROM, 300...RAM, 30
1... Quantization matrix table, 302... Variable length encoding table, 303A, 303B... Scan table, 400... Image data input unit, 500
...This is an encoded data output section. Fig. 2 (a) Fig. 7 (a) (b) Fig. 8 (b) (C) 1 5 00 @ soft 1 00 number ( ) (b) Fig. 9 ) ( ) (a) (b) ( ) (b) [' ( ) (d) II2 Figure 11

Claims (3)

[Claims] (1) An image encoding method that compresses and encodes orthogonally transformed image information, which performs encoding by selecting a coding order with a high compression rate according to the distribution of the orthogonally transformed image information, An image encoding method characterized in that the compression-encoded data and data identifying the encoding order are used as encoded data of the image information. (2) An image encoding device that compresses and encodes image information for each block consisting of a plurality of pixels, comprising orthogonal transformation means that performs orthogonal transformation for each block, and within each block of the orthogonally transformed data. coding order selection means for selecting the coding order of data in each block based on the distribution of the data, and coding means for coding the orthogonally transformed data according to the selected coding order of each block. An image encoding device comprising: (3) The image encoding apparatus according to claim 2, further comprising a second encoding means for encoding the information on the encoding order.