JPH0273778A - Compressor and decoder for picture data - Google Patents

Abstract

PURPOSE:To attain higher speed of data decoding processing by using each decoding section of a picture decoder so as to process each inputted code in parallel based on a code quantity. CONSTITUTION:A code quantity calculation means 4 of a picture data compressor calculates code quantity information representing the code quantity of a reference gradation coded by coding means 1-3, a difference value and a resolution component and a code quantity transmission means 5 sends the code quantity information before the coded reference gradation value, a difference value and a resolution component are sent. A code quantity detection means 9 of a picture data decoder receiving a picture data decodes the coded code quantity information, detects the code quantity of the reference gradation, the difference and the resolution component sent from now and a signal distribution means 10 distributes inputted codes so that the code is processed in parallel with decoding sections 6-8 based on the code quantity. Thus, the decoding processing of the picture data is implemented in parallel at a high speed.

Description

[Detailed description of the invention] 〔table of contents〕 overview Industrial application field (Figure 6) Conventional technology (Fig. 7, Fig. 8, Fig. 9).

Fig. 10) Means for solving the problems to be solved by the invention (Fig. 1) Working examples (Fig. 2, Fig. 3, Fig. 4, Fig. 7) Effects of the invention [Summary] Image Regarding data compression devices and decompression devices, especially regarding image data compression devices that encode large amounts of image data and image data decompression devices that decompress the encoded image data, there are differences in image data decoding processing. The aim is to speed up the data decoding process by performing data processing in parallel.The multilevel image is divided into blocks consisting of multiple adjacent pixels, and each block is The difference between the maximum pixel and the minimum pixel in a block is compared with a threshold value to obtain one or more representative gradation values for displaying the image of this block, and this representative gradation value is used as the reference gradation value in each block and this Displays the difference value between the standard gradation value and the representative gradation value, calculates the resolution component that indicates the distribution of the representative gradation value in each block, and encodes the reference gradation value, the difference value, and the resolution component. In the image data compression device and the restoration device that restores the image data, the image data compression device uses a reference gradation value that encodes code amount information indicating the amount of each data code after compressing the image data. , before transmitting the difference value and the resolution component, and the image restoration device is configured so that each input code can be processed in parallel by each decoding unit based on the above-mentioned code amount.

[Industrial application field]

The present invention relates to an image data compression device and a decompression device, and more particularly to an image data compression device that encodes a large amount of image data, and an image data decompression device that decompresses the encoded image data.

[Conventional technology]

For example, there is a multi-gradation adaptive block coding method (Preliminary Draft 6 of the National Conference of the Institute of Image Electronics Engineers, 1988) as a compression method for highly efficient image data.

This multi-gradation adaptive block coding method (General
ized Block Truncation Cod
ing (hereinafter abbreviated as GBTC) divides the original image into blocks consisting of NxN pixels, and each pixel (X i
j ) within the maximum and minimum pixel levels within the block
(n=o.

1.2. ...) level, expresses the quantization level of each pixel in a bit plane format, and encodes gradation information and bit plane information.

This is N=4. The case where n=2 will be described in detail. FIG. 7 shows the GBTC algorithm. Each block has a maximum pixel level within the block (MA
The encoding modes are classified into the following three encoding modes (Mode A, Mode B, and Mode C) based on the difference between the minimum pixel level (MINL) and the encoding parameters T t and T 2 (T 1 < T 2 ).

Mode A: If MAXL-MINL≦T1, the pixels in the block are quantized to levels (po).

Mote B: T 1 < M AX L M I N
If L≦T 2 , the pixels within the block are quantized to two levels (pt, P2).

Mode C: If T2<MAXL-MINL, the pixels within the block are quantized into four equally spaced levels (Q, to Q6.).

The quantization level is determined by the block reference level La, the level interval Ld, and the level designation signal for each pixel (φ,), , (φ
2) Described by 1j. If the average value processing is AVE(), each element required for encoding is calculated as follows.

Mode A: Po=AVE (Xij)=La(φ+1=
O1φ2) iJ=O (for all i, j) Mode B: P 1= A V E (X t J≧
(MAXL+MINL)/2) P 2 = A VE (X I J < (M A
X L +MINL)/2) La= (Pi +P2)/2 Ld=P1-P2 (φt)tj=O: (However, in the case of Xij≧(MAXL+MINL)/2) (φ□), J= 1: (However, if Xij< (MAXL+MINL)/2) (φ2)tj=o: (for all i+J) Mode C: Q1=AVE (X i j≧(3MAX
L+MINL)/4) Q4 =AVE (X i j < (MAXL+3M
INL)/4) La = (Ql + Q4) / 2La=2(Q
, -Q, + )/3 (φ1)ij=0. (φ2)1J=0: (However, Xi
When j≧La+Ld/2) (φt);j=O+(φ2)fJ=1 (However, La+Ld/2>Xi j≧
(In case of La) (φt)tj=t, (φ2)i;=O: (However, La
>X ij >= L a −L d / 2) (φx)tJ=t, (φz)tJ=1: (Tatashi, L
a − L d / 2 > Xij) The resolution component (φ1.φ2) is connected between blocks and
The images are converted into two bitmaps and encoded using the MMR encoding method, which is a standard encoding method for binary images. Ld is variable length coding after nonlinear quantization, La is DPCM
After nonlinear quantization of the pre-difference (ΔLa) using encoding,
Perform variable length encoding.

In the GBTC method, ISO contribution ISO/TC97/SC2/We N510 ”Ge
nerizedBlock Truncation
As described in ``Coding'', a progressive build-up function can be realized using internal stages as shown in Figure 1O. The process was carried out as shown in the figure.That is, from the first stage to the second and third stage.
Up to the 7th stage, the image data is transmitted, and the person using this image data is able to stop acquiring the image data as necessary, thereby shortening the image data transmission time etc. This is what I did.

At this time, an END detection signal indicating the end of the data is added to the end of each element data, as shown in FIG.

FIG. 6 shows this image data restoring device.

In the figure, 10 is a buffer memory for temporarily storing the transmitted image data, 11 is a reference value decoding unit 12, a difference value decoding unit 13, and a resolution component decoding unit for decoding the reference value, difference value, and resolution component in the image data. a signal distribution unit that distributes to 14;
Reference numeral 17 indicates a data control section which operates the signal distribution section based on an end detection signal of the image data width. In addition, 18 in the same figure is a difference value storage unit 1 for storing the decoded difference value.
9 is a resolution component storage unit that stores the decoded resolution component, 2o is a mode detection unit that uses the mode of the block from the resolution component, and 21 is a unit that calculates a representative gradation value from the difference value, resolution component, and mode. 22 indicates an update memory that stores the reference value and the representative gradation value. A decoding unit (12, 13, 14) for each element data.
After detecting the end detection signal, control is returned to the control unit 17 and switched to the decoding unit for the next data.

That is, the image data restoration device converts the initial data into Y component ΔLa(1
), the data control section 17 selects the reference value decoding section 12, sets the data in the buffer memory 10, and starts decoding processing. The reference value decoding unit 12 uses the reference value La
The process of decoding (1) and writing the result to the update memory 22 is repeated. The reference value decoding unit 12 performs the decoding until it detects a predefined END detection signal (see FIG. 9). Upon detecting the END detection signal, the reference value decoding unit 12 immediately returns control to the data control unit 17. Assuming that ΔLa(1) of the U component is from the code data next to the END detection signal in the buffer memory 10, the data after this data is set in the reference value decoding unit 12, and the decoding process is restarted.

Thereafter, the encoded data is decoded as U component La(1) and written into the update memory 22 in the same manner as described above. Repeat the above decoding operation, Y.

After restoring La of the U and V components (after detecting the 12th END detection signal), control returns from the reference value decoding section 12 to the data control section 17. At this time, the control unit 17 concludes the decoding of all La data, sets the data in the resolution component decoding unit 14 as data of the Y component φ from the next code data in the buffer memory 10, starts decoding, and outputs the result. It is written into the resolution component storage section 19. Furthermore, the mode detecting section 20 is activated, and the mode of each block is detected from the data written in the resolution component storage section 19. Similar to the reference value decoding section 12, E
Decoding of the Y component φ1 continues until the ND detection signal is detected.

The resolution component decoding unit 14 outputs the END detection signal (13th E
When the ND detection signal) is detected, control returns to the data control section 17. The data control unit 17 regards the next code data in the buffer memory 10 as Y component Ld data, sets the data in the difference value decoding unit 14, and starts decoding. The restored Ld data is stored in the difference value storage section 18.

Thereafter, until the END detection signal is detected, the difference value decoding unit 1
3 is repeated, and the result is stored in the difference value storage section 18. When the difference value decoding unit 13 detects the END detection signal, the representative gradation value calculation unit 21 is activated, and calculates a representative gradation value while referring to the detection result stored in the mode detection unit 20, The data in the update memory 22 is also added to reconstruct the image.

When the detection mode detection result is A mode, the representative gradation calculation unit 21 sets the value to be added to the update memory 22 as 0. In addition, in the case of mode, the difference value storage unit 18
The above-mentioned gradation level is calculated from the La value within, and the calculated value is added to the update memory 22. After referring to the final data of the mode detection section, control is returned to the data control section 17.

Using the same procedure, the Y component φ2, the U component φ0, Ld, and the ■ component φ1, Ld are decoded in this order, and the restored image in the update memory 22 is updated.

[Problem to be solved by the invention]

By the way, according to the conventional image data compression and decompression method described above, when restoring image data, the END detection signal added to each encoded data is detected, and each time the data control unit executes the next data processing procedure. I'm trying to get it running. Therefore, it is not possible to decode each piece of data in parallel, and there is a limit to the speed improvement of data decoding processing.

SUMMARY OF THE INVENTION Accordingly, the present invention provides an image data compression device and an image data restoration device that can perform data decoding processing at higher speed by processing different data in parallel in the image data decoding processing. With the goal.

[Means to solve the problem]

In the first invention of the present application, the means for solving the above problem is to divide a multivalued image into blocks each consisting of a plurality of adjacent pixels, as shown in FIG. For each block, the difference between the maximum pixel and minimum pixel in the block is compared with a threshold value to obtain one or more representative gradation values for displaying the image of this block, and this representative gradation value is used as the reference gradation value for each block. , display the difference value between this standard gradation value and the representative gradation value, calculate the resolution component that shows the distribution of the representative gradation value in each block, and calculate the above-mentioned standard gradation value, difference value, and resolution component. In an image data compression apparatus comprising reference tone value encoding means 1, difference value encoding means 2, and resolution component encoding means 3, each of the aforementioned encoding means 1, 2.3
code amount calculation means 4 that calculates code amount information indicating the code amount of the reference tone value, difference value, and resolution component encoded by the encoded reference tone value;
The present invention also includes code amount transmission means 5 that transmits the difference value and resolution component before transmitting them.

As shown in FIG. 1 (2), the means for solving the above problem in the second invention of the present application is to divide a multilevel image into blocks each consisting of a plurality of adjacent pixels, and to The difference between the maximum pixel and the minimum pixel in the block is compared with a threshold value to obtain one or more representative gradation values for displaying the image of this block, and this representative gradation value is used as the reference gradation value in each block. The difference value between this standard gradation value and the representative gradation value is displayed, a resolution component indicating the distribution of the representative gradation value in each block is calculated, and the above-mentioned standard gradation value, difference value, and resolution component are requested. In the image data restoration device for restoring image data encoded by the image compression device according to item 1, the reference gradation value decoding means 6, the difference value decoding means 7,
resolution component decoding means 8; code amount detection means 9 for decoding encoded code amount information and detecting the code amount of the reference gradation value, difference value, and resolution component to be transmitted; The present invention also includes a signal distribution means 10 that distributes the code to each decoding unit so that the code is processed in parallel based on the code amount.

[Effect]

According to the present invention, in the image data compression device, the code amount calculation means calculates code amount information indicating the code amount of the reference gradation value, the difference value, and the resolution component encoded by the encoding means,
The code amount transmission means transmits the code amount information before transmitting the coded reference gradation value, difference value, and resolution component.

In the image data restoration device that receives this image data, the code amount detection means decodes the encoded code amount information and detects the code amount of the reference gradation value, difference value, and resolution component that will be transmitted from now on. Since the signal distribution means distributes each input code to each decoding unit for parallel processing based on the code amount, image data can be decoded in parallel at high speed.

〔Example〕

Embodiments of an image data compression device and a decompression device according to the present invention will be described below with reference to the drawings.

0 Image Data Compression Device An embodiment of the image data compression device according to the first invention will be described below. FIG. 2 shows an embodiment of the image data compression device according to the first invention. In the figure, 51 is an input terminal for image data, and 52 is an input terminal for inputting image data into luminance component color difference (Y) data 53, color difference component (U) data, and color difference (
Y) Data separation unit (DMPX) 5 that separates component data
6 is a multiplexer that sequentially outputs these data 53, 54, and 55, and 57 is the maximum gradation of image data (1max
) and the minimum gradation value (1 min), and 59 indicates a gradation number determining unit that determines the number of gradations based on the threshold value from the threshold value memory 58. Also, in the same figure, 60 indicates the maximum gradation (1max) and the minimum gradation value (1min
) and the number of gradations, a representative gradation value determination unit 61 generates a reference gradation value (L, ) and a difference value (Ld) for determining one or more representative gradation values. 62 is a representative gradation value buffer memory that temporarily stores this value; 63 is a difference value encoding unit that encodes a difference value; 64 is a coder for this encoding. The figure shows a difference value buffer memory that stores the calculated difference values. Furthermore, 6
5 is a gradation value storage unit that stores representative gradation values; 66 is a comparison unit that compares image data and representative gradation values; 67.70 is a resolution component storage buffer; and 68.71 is used to encode resolution components. The resolution component encoding unit 69.72 indicates a resolution component code buffer that stores encoded resolution components. 73 is a code amount calculation unit that calculates the code amount of the coded reference gradation value, difference value, and resolution component, respectively; 74 is a code amount memory that stores this code amount; It shows a multiplexer that sends out the code amount before the code of the reference gradation value, difference value, and resolution component.

Next, the operation of this image data compression device will be explained.

First, select the original image (one image of each component image) and
As shown in FIG. 7, the image is divided into blocks each consisting of 4×4 pixels, and the image data X, , of one of these blocks is processed by a multiplexer (MPX) as the luminance (Y) component image data 5.
Either of 31 color difference (U) component image data 542 color difference (V) component image data 55 is selected and sequentially read out.

The gradation number determination unit 59 calculates a difference value L a = 1max − 1m1n based on the intra-block maximum gradation value 1max and minimum gradation value 1m1n from the gradation change amount detection unit 57, and calculates the difference value L a = 1max − 1m1n. The threshold value Tl for representative gradation determination which is the output of the threshold value memory 58 (TIY for Y component, T for UV component
luv) to determine whether the number of representative gradations is 1 or 2 or more. When it is determined that the number of representative gradations is 2 or more, the difference value is compared with a second threshold T2 for determining the number of representative gradations in the threshold memory 58, and it is determined whether the number of representative gradations is set to 2 or 4.
Decide what to do.

According to the number of representative gradations determined by the number of gradations determining section 59,
The representative tone value determining unit 60 determines the tone within the block by linear or sublinear quantization.

When the number of representative gradations is 1, the average value within the block is set as the number of representative gradations, and the reference value is set as a. Furthermore, the resolution component φ1.
As φ2, 0 is assigned to all pixels.

If the number of representative gradations is 2, the maximum gradation value in the block is 1.
The intermediate value 1m1d between max and the minimum gradation value 1m1n is calculated, and the average value P of the gradation values greater than or equal to the intermediate value is calculated. 1
The average value P2 of the number of gradations is calculated. Furthermore, the reference value L8
The average value of Pl and P2 (p++P2)/2 is calculated as follows.

Assuming that the resolution component is φ1, 0 is assigned to a pixel whose representative tone value is P□, and 1 is assigned to a pixel whose representative tone number is 22. φ2 assigns 0 to all pixels. Further, the difference PI P2 between P1 and P2 is defined as a difference value Ld.

If the number of representative gradations is 4, the maximum gradation value in the block is 1.
Divide the area between max and minimum gradation value 1m1n into 4 equal parts and divide the upper 1/
The average value Q1 of the gradation values in the range 4 (including the boundary) and the average value Q4 of the gradation values in the lower 1/4 range (excluding the boundary) are determined. Standard value.

calculates the average value (Q1+Q4)/2 of the Q-work and Q4. Also, from the above Q□ and Q4, 2 (Ql −Q4
)/3 is the difference value Ld. Furthermore, Q,, Q, between 3
Calculate the gradation values to be divided into equal parts. The gradation value that divides the space between Q and Q into three equal parts is taken as the representative gradation value of the block. The resolution components, which are defined as Qyu, Q2°Q, , Q4, are assigned as follows according to the representative gradation value.

For pixels whose representative gradation value is Ql, φ1=O1φ2=0 For pixels with representative gradation value or Q2, φ1=0. φ2=1 For the pixel whose representative gradation value is Q3, φ1=1. φ2=0 For the pixel whose representative gradation value is Q4, φ1=1. φ2=1 shall be.

As described above, after the representative gradation value determination section 60 determines the representative gradation value, reference value, and difference value within the block, the representative gradation value is stored in the gradation value storage section 65.

The comparison unit 66 converts one block of multivalued image data XfJ into one
Each pixel is read out and compared with the representative gradation value in the gradation value storage section 65. This determines which representative gradation value should be used for approximation, and the resolution component φ1 corresponding to the representative gradation value.
φ2 is output, and the output resolution components are stored in resolution component storage buffers 67 and 70, respectively.

The contents of the resolution component storage buffer 67.70 are read out to an encoding section 68.71, encoded by known MMR encoding, and stored in a resolution component buffer 69.72. Further, the code amount calculation unit 73 calculates the code amount in the resolution component buffer 69, 72, and stores the code amount value in the code amount memory 74.

The reference value L8 and the difference value Ld calculated by the representative gradation value determining unit 60 are variable-length encoded by a code generating unit 61.63 and stored in buffer memories 62.64, respectively. Further, as in the case of the resolution component described above, the code amount calculation unit 73 calculates the code amount in each of the buffer memories 62 and 64, and stores the value in the code amount memory 74.

Perform the above operation for the luminance component and the two color difference components,
The code amount data of each component is stored in the code amount memory 74.

As described above, after the completion of the multi-gradation image of all components, the code amount data Lla+La, φ1.
The encoded data of φ2 is sequentially read out via a multiplexer (MPX) 75 and sent out as transmission data.

The transmission data consists of a parameter section (FIG. 3 (1) in this embodiment) indicating the amount of code data and a code data section (FIG. 3 (2)) following the parameter section (FIG. 3 (2)). The parameter section indicates, for example, the end position of each code data (the number of bytes from the beginning of the code data).

■Image data restoring device Hereinafter, an embodiment of the image data restoring device according to the second invention will be described. FIG. 4 shows an embodiment of the image data restoration device. In the figure, 31 is a buffer memory that temporarily stores transmitted image data, 32 is a signal distribution unit that distributes each component of the image data to each decoding unit under the control of a data control unit 42, which will be described later, and 33 is reference value decoding. 34 is a difference value decoding unit, 35 is a resolution component decoding unit, 36 is a difference value storage unit, 37 is a resolution component storage unit, 38 is a mode detection unit,
Reference numeral 39 indicates a representative tone value calculation section, and reference numeral 40 indicates an update memory. In this embodiment, a code amount memory 41 is connected to the signal distribution section to store the first transmitted code amount of image data. This code amount memory has a storage area for storing the code amount of each data as shown in FIG. The code amount memory 41 is provided with a data control section 42 that controls the signal distribution section by referring to the contents of the code amount memory.

Next, the operation of the image data restoration device according to this embodiment will be explained.

The image data restoring device according to this embodiment stores image data in a buffer memory 31, and separates a parameter section at the beginning of the data by a signal distribution section 32 under the control of a data control section 42, and stores the image data in a code amount memory 31. Store in. When the data in the buffer memory exceeds the first value in the code amount memory 41, the data control unit 41 operates the signal distribution unit 32 and sets the data in the buffer memory 31 in the reference value decoding unit 33. do. The reference value decoding unit 33 decodes the data into Y components, and writes the decoded data into the update memory 40.

When the data transmission progresses and the data in the buffer memory 31 exceeds the second value in the code amount memory 41, the signal distribution unit 32, with the support of the data control unit 41, transfers the buffer memory to the reference value decoding unit 33. Set the data in 31 and press U.
It is decoded as component La (1) and the result is written to update memory 40.

Thereafter, in the same manner, each La encoded data is decoded and the image is restored on the update memory 40.

When the data in the buffer memory 31 exceeds the 13th value in the code amount memory 41, the data control unit 41 operates the signal distribution unit 32 to transfer the code data in the buffer memory 31 to the resolution component decoding unit 35. , and decoding of the resolution component is started. This is the Y component φ.

Thereafter, by repeating the same procedure, each component is restored and the image is reconstructed.

As shown above, when the amount of transmitted codes exceeds the amount required for restoration, the bits are sequentially set in the decoding unit and decoding is started. Therefore, if processing can be performed by parts that can operate in parallel, such as the resolution component decoding section 35 and the difference value decoding section 34, parallel processing can be performed, and speeding up of the decoding process can be realized.

Note that the code data amount description part, which is unique to the present invention, is 66 bytes in the above example. Since the conventional termination code is not required, the present invention hardly increases the amount of transmission.
The above-mentioned high-speed image restoration can be achieved.

〔Effect of the invention〕

As explained above, according to the present invention, since the image data compression device and the image data restoration device are configured as described above, the image data can be restored in parallel when restoring the image data. This has the effect that image data restoration processing can be performed at higher speed.

[Brief explanation of the drawing]

FIG. 1 is a principle diagram of the present invention, FIG. 2 is a block diagram showing an embodiment of an image data compression device according to the present invention, and FIG. 3 (1)
, (2) is a diagram showing the output data of the image data compression device shown in FIG. 2, FIG. 4 is a block diagram showing an embodiment of the image data restoration installation according to the present invention, and FIG. 6 is a block diagram showing the conventional image data restoring device, FIG. 7 is a diagram showing the GBTC algorithm, and FIGS. 8 and 9 are FIG. 10, a diagram showing conventional image data, is a diagram showing the contents of each stage of image data. 1... Reference gradation value encoding means 2... Difference value encoding means 3... Resolution component encoding means 4... Code amount calculation means 5... Code amount transmission means 6... Reference Gradation value decoding means 7...Difference value decoding means 8...Resolution component decoding means 9...Signal distribution means 10...Code amount detection means

Claims (2)

[Claims] (1) Divide a multilevel image into blocks consisting of a plurality of adjacent pixels, compare the difference between the maximum pixel and minimum pixel in each block with a threshold value, and display the image of this block. Find the representative gradation value, display this representative gradation value as the standard gradation value in each block, and the difference value between this standard gradation value and the representative gradation value, and calculate the distribution of the representative gradation value in each block. a reference gradation value encoding means (1), a difference value encoding means (2), and a resolution component encoding means (1), a difference value encoding means (2), and a resolution component encoding means (1), which calculates a resolution component indicating the above-mentioned reference gradation value, a difference value, and a resolution component; 3), a code amount calculation means (4) for calculating code amount information indicating the code amount of the reference gradation value, difference value, and resolution component encoded by each of the above-mentioned encoding means; , a code amount transmission means (5) that encodes the code amount information and transmits the encoded reference gradation value, difference value, and resolution component before transmitting the code amount information. . (2) Divide a multivalued image into blocks consisting of a plurality of adjacent pixels, compare the difference between the maximum pixel and minimum pixel in each block with a threshold value, and display the image of this block. Find the representative gradation value, display this representative gradation value as the standard gradation value in each block, and the difference value between this standard gradation value and the representative gradation value, and calculate the distribution of the representative gradation value in each block. An image data restoring device for calculating a resolution component indicating a reference gradation value, and restoring image data encoded by the image compression device according to claim 1 using the reference gradation value, the difference value, and the resolution component. Value decoding means (
6), a difference value decoding means (7), a resolution component decoding means (8), and codes for the reference gradation value, difference value, and resolution component that are to be transmitted after decoding the encoded code amount information. An image characterized by comprising a code amount detection means (9) for detecting the amount of code, and a signal distribution means (10) for distributing each input code to each decoding unit for parallel processing based on the code amount. Data recovery device.