JP2000299866A - Method for encoding image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain encoding/decoding method that improves encoding/decoding efficiency even such an image in which image quality is locally different by evaluating correlation strength among ambient pixels of an encoding object pixel, deciding a reference pixel value on the basis of evaluated results and calculating the differential value between the reference pixel value and the encoding object pixel value. SOLUTION: A predicting part has a prediction direction selecting part 310 and a reference image value setting part 320. The pixel value 50 of ambient pixels is inputted to the part 310 from an image buffer. The part 30 evaluates which direction is stronger of correlation as for the ambient pixels according to a preliminarily decided evaluation expression and outputs results as a predictive direction signal 51 to the part 320. The part 320 outputs the value 50 of the ambient pixels used for deciding the signal 51 and a preliminarily defined pixel value of ambient pixels to the signal 51 outputted from the part 310 or a pixel value calculated from the ambient pixels as a reference pixel value 52.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a method for lossless encoding and decoding of image information.

[0002]

2. Description of the Related Art In the prior art, an irreversible encoding method and a reversible encoding method are defined in the international standard JPEG method for encoding and decoding still images and moving images. DPCM + entropy coding is generally known as a reversible coding method. Here, DPCM stands for differential pulse-code modula.
) is a method of reducing the signal bandwidth by transmitting only the difference between successive pixels. Entropy coding is a coding scheme in which a short codeword is assigned to a code having a high appearance probability, and a long codeword is assigned to a code having a low appearance rate. Can be reduced.

In the reversible encoding method, the original image can be completely reproduced by decoding even after encoding, that is, compression, but there is a disadvantage that the compression ratio is small. In the lossless encoding method in JPEG, difference value data between a pixel to be encoded and a prediction value calculated using pixels surrounding the pixel to be encoded is set as a prediction error, and the prediction error is encoded. Basic.

FIG. 1 is a diagram for explaining a prediction method according to the prior art (JPEG system). Pixel X is a pixel to be encoded,
A, B, and C are surrounding pixels used for prediction of the pixel X, and are pixels that have already been encoded. There are seven difference generation formulas (prediction calculation formulas) for calculating the difference between the predicted value and the encoding target pixel X, as shown in (1) to (7) of FIG. A prediction operation formula is selected, and a prediction error is obtained from the predicted value obtained by the selected prediction operation expression and X. ("International standard for multimedia coding"
Edited by Hiroshi Yasuda, Maruzen Co., Ltd .; p. 28 See Figure 1.11. )

[0005]

However, the prediction formulas in the above-mentioned prior art are closely related to the properties of local images, and the magnitude of the compression ratio is influenced by the selection of the prediction formula. . In the DPPEG encoding method of JPEG, since the prediction operation formula is fixed for each scan,
When the properties of the image are locally different, there is a problem that the coding efficiency is deteriorated.

Further, it is necessary to transmit and receive additional information such as which arithmetic expression is used for encoding, and this also causes a reduction in encoding efficiency. SUMMARY OF THE INVENTION It is an object of the present invention to provide an encoding / decoding method for improving encoding / decoding efficiency even in an image having a locally different image property in an image reversible encoding method.

[0007]

To achieve the above object, the present invention is configured as follows. The present invention relates to an encoding method for encoding image information, comprising: an evaluation step of evaluating the strength of correlation between surrounding pixels of an encoding target pixel; and determining a reference pixel value based on a result of the evaluation. Steps and
Obtaining a difference value between the reference pixel value and the encoding target pixel value, and encoding the difference value.

[0008] In the above configuration, the evaluation step includes:
Comparing a difference step between at least a difference value between pixels in the vertical direction and a difference value between pixels in the horizontal direction in a plurality of peripheral pixels adjacent to the encoding target pixel, and comparing the plurality of difference values obtained by the difference step And the step of performing. To achieve the above object, the present invention can also be configured as follows.

The present invention relates to a decoding method for decoding coded image information, comprising the steps of restoring coded data to obtain a restoration value, and determining the strength of correlation between surrounding pixels of a decoding target pixel. An evaluation step of evaluating, a step of determining a reference pixel value based on the evaluation result, a step of adding the reference pixel value and the restored value to obtain decoded data, and using the decoded data as surrounding pixel values And

According to the present invention, the reference pixel value of the encoding target pixel value is determined based on the result of evaluating the pixel value of each peripheral pixel of the encoding target pixel. The reference pixel value is determined. Therefore, it is not necessary to transmit and receive the additional information that was conventionally required when decoding the used arithmetic expression or the like, and it is possible to improve the encoding efficiency.

Further, according to the present invention, the pixel to be encoded is compared with its surrounding pixels, and the pixel to be encoded is
In the vertical direction, if necessary, it is determined which pixel in the diagonal direction has a strong correlation, and based on the result, it is determined which pixel to use as the predicted value, so that the Since the difference value can be made a relatively small value, the coding efficiency can be further improved.

Further, according to the decoding method of the present invention, the image information encoded as described above can be efficiently decoded.

[0013]

FIG. 2 is a configuration diagram of an encoding apparatus according to an embodiment of the present invention. As shown in the figure, the present encoding device includes an image buffer 110, a prediction unit 120, a differentiator 130, and an encoding unit 140. The operation of the present encoding device will be described based on the flowchart shown in FIG.

First, as step 1, image data 1
(Each pixel value) is stored in the image buffer 110. Note that image data can be stored in the image buffer as long as pixels used for encoding are stored in the image buffer, and various methods can be used. For example, in the case of scanning in the horizontal direction, data for several rows may be stored. Next, as step 2, the surrounding pixel value 3 for the encoding target pixel value 2 in the image data 1 is input from the image buffer 110 to the prediction unit 120. In step 3, the prediction unit 120 evaluates which direction of the surrounding pixels has a strong correlation using a predetermined evaluation formula, and outputs the encoding reference pixel value 4 to the differentiator 130 based on the evaluation result. .
The details of the evaluation in the prediction unit 120 will be described later.

The differentiator 130 has an encoding reference pixel value of 4
And the encoding target pixel value 2 is input from the image buffer 110. In step 4, a value obtained by subtracting the encoding reference pixel value 4 from the encoding target pixel value 2 is obtained.
And input to the encoding unit 140. Here, in the present embodiment, the differentiator 130 reads the encoding target pixel value 2 from the image buffer 110,
Before storing the image data in the differentiator 10
May be entered.

In the encoding unit 140, as step 5,
The encoding target data 5 input from the differentiator 130 is encoded, and the compressed data 6 is output as step 6. As an encoding method, entropy encoding such as Huffman encoding is used. FIG. 4 shows a configuration example of the prediction unit 120 described above. As shown in the drawing, the prediction unit 120 includes a prediction direction selection unit 310 and a reference pixel value setting unit 320. The operation of the prediction unit 120 will be described with reference to the flowchart shown in FIG.

In step 10, a prediction direction selection unit 31
0, the pixel value 50 of the surrounding pixels from the image buffer 110
Is input as a step 11 to evaluate which direction of the surrounding pixels has a strong correlation by a predetermined evaluation formula.
The value is output to the reference pixel value setting unit 320 as 1. The details of the prediction direction selection unit 310 will be described later.

The reference pixel value setting section 320 stores the pixel values 50 of the surrounding pixels used to determine the prediction direction signal 51.
And the prediction direction signal 51 output from the prediction direction selection unit 310, and as a step 13, a pixel value of a predetermined surrounding pixel or a pixel value obtained from the surrounding pixel is determined with respect to the prediction direction signal 51. It is output as a reference pixel value 52.

FIG. 6 shows a detailed configuration of the prediction direction selection unit 310 described above. As shown in FIG.
10 has a plurality of prediction direction calculators 311 to 313 and a minimum value determination unit 314. The operation of the prediction direction selection unit 310 is as follows. The pixel value of a predetermined pixel among the surrounding pixels of the pixel to be coded (the pixel at the position already coded) is input to the prediction direction calculators 311 to 313, respectively. The prediction direction calculators 311 to 313 obtain the strength of the correlation between the peripheral pixels using a correlation evaluation formula set in advance for each calculator, and output it as an evaluation value. The correlation evaluation formula will be described later.

Next, the operation of the prediction direction selection unit 310 in FIG. 6 and the prediction unit shown in FIG. 4 will be described more specifically with reference to FIG. In the present embodiment, as will be described later with reference to FIG. 9, a case will be described in which the image encoding process is performed from the upper left direction to the lower right direction of the image. A, B, C, and X in FIG. 7 represent pixels, respectively. When X is a pixel to be encoded, A, B, C
Are surrounding pixels. D is a virtual pixel obtained from pixels A and C, which is also used for evaluation. here,
The surrounding pixels A, B, and C are pixels at positions already encoded. That is, when the image encoding process is performed from the upper left direction to the lower right direction of the image, the pixel A located immediately to the left of X, the pixel C located immediately above X, and the pixel A located immediately above X (that is, the pixel C
(The pixel immediately to the left) is a pixel surrounding the pixel X. Further, the pixel D in FIG. 7 is a virtual pixel provided for evaluating the correlation in the oblique direction, and the pixel D is also a pixel X
Surrounding pixels.

In this embodiment, it is assumed that the pixel value of the pixel D has a value obtained by the following equation (1). Equation 1 D = (A + C) / 2 In the prediction direction calculator 311 in FIG. 6, the values of the pixel A and the pixel B are input from the image buffer 110, and the strength of the correlation between the pixel A and the pixel B is calculated based on the two pixel values. Obtained as a difference value. That is, the correlation evaluation formula in the BA direction is | AB |. If the difference value is small, it means that the characteristics (for example, luminance) of both pixels are similar.

Similarly, in the prediction direction calculator 312,
The values of pixel B and pixel C are input from image buffer 110,
The difference between the two pixel values is obtained. That is, the correlation evaluation formula in the BC direction is | CB |. The prediction direction calculator 313 inputs the values of the pixels A, B, and C from the image buffer 110, obtains the value of the virtual pixel D by Expression 1, and obtains the value of the virtual pixel D.
And the difference between the two pixel values of the pixel B. That is, the correlation evaluation equation in the BD direction is | DB |.

The difference value obtained by each prediction direction calculator is compared in the minimum value determination unit 314, and a pixel value to be taken as a difference from the pixel X is determined depending on which direction of the surrounding pixels has a strong correlation. FIG. 8 shows a prediction direction corresponding to the direction in which the correlation is the strongest, that is, the direction in which the evaluation value is minimum, that is, a reference pixel value to be taken as a difference from the pixel X. The prediction direction is determined based on this correspondence table.

For example, when | AB | is determined to be the minimum, the minimum value determination unit 314 of FIG. 6 selects the value of the pixel C as a reference pixel for encoding the pixel X. The prediction direction signal 51 is generated and output to the reference pixel value setting unit 320 shown in FIG. The reference pixel value setting unit 320 in FIG. 4 selects and outputs one of the pixel values of the surrounding pixels input from the image buffer 110 as an encoded reference pixel value 52 based on the prediction direction signal 51. For example, when | AB | is determined to be the minimum by the minimum value determination circuit 314, the pixel value of the pixel C becomes the encoding reference pixel value 52.
When the prediction direction signal 51 specifies the virtual pixel D, the reference pixel value setting unit 320 sets the pixel A and the pixel C
The reference pixel value is calculated by Expression 1 using the pixel value of, and is output as the reference pixel value 52 for encoding.

As described above, the encoding reference pixel value 52 is uniquely determined from each peripheral pixel value of the target pixel. Here, in the pixel arrangement shown in FIG.
In the case of the present embodiment in which F, B, C, E, A, and X are encoded in this order, the reference pixel value of X is determined by the method described above, but G, H, I, F, E Since the conditions of surrounding pixels are different from those in the above example, the prediction method is different from the above method. FIG. 10 is a diagram for explaining a prediction method for each region of a pixel in an encoding target image. The prediction method for each region in the present embodiment will be described with reference to FIG.

In the figure, the value of a pixel determined based on the evaluation of the correlation described with reference to FIGS. 7 and 8 is used as a reference pixel value for a region IV. In the case of the region II, that is, in the pixel on the uppermost side of the image, there are no pixels corresponding to the pixels B and C shown in FIG. 6 as peripheral pixels, and thus prediction is performed so that the pixel value on the left is always selected. The direction signal 51 is output.

In the case of the region III, that is, in the pixel on the leftmost side of the image, there are no pixels corresponding to the pixels B and A shown in FIG. 7 as peripheral pixels, so that the pixel value immediately above is always selected. To output the prediction direction signal 51. Further, in the case of the pixel in the region I, that is, in the pixel to be encoded first, an intermediate value (0 to 2) of values that can be taken by the encoding target pixel.
A fixed value determined in advance such as 128 for 55 gradations of 55 is output as the encoding reference pixel value 52.

FIG. 11 is a flowchart showing the procedure of the method for determining the reference pixel value of the pixel to be coded as described above. In step 21, the prediction direction selection unit 310 illustrated in FIG. 4 determines the region of the encoding target pixel. If the region is other than IV, in step 22,
A predetermined prediction direction signal is output, and in step 23,
The reference pixel value setting unit 320 outputs the corresponding reference pixel value.

If the region is IV, step 24
In step, correlation evaluation in each direction is performed.
Then, the prediction direction signal in the direction in which the evaluation values are compared and becomes the minimum is output from the prediction direction selection unit. In step 26, the reference pixel setting unit outputs a reference pixel value corresponding to the prediction direction signal. Next, a decoding process for decoding the encoded image data will be described. FIG. 12 is a diagram showing the configuration of the decoding device in the present embodiment. As shown in the drawing, the present decoding device includes a decoding unit 210, a prediction unit 220,
An adder 230 and a decoded image buffer 240 are provided. Next, the operation of the present decoding apparatus will be described using the flowchart shown in FIG.

In step 30, the compressed data 6, which is encoded image data, is input to the decoding unit 210. Subsequently, as Step 31, the decoding unit 210 decodes the compressed data 6, and the encoding target data 5 in FIG. 2 is restored. That is, a difference value between the encoding target pixel value 2 (target pixel X) and the encoding reference pixel value 4 is restored. Then, it is output as decoded data 8.

In the case of the pixel arrangement shown in FIG. 9, the encoding apparatus starts with the pixel G as the head, and outputs H, I, F, B, C, E,
A, the data is coded in the order of A and transmitted as compressed data, and the decoding device sends H, I, F, B, C,
E and A are received in this order, decoded in that order and output as image data, and the pixel value of the subsequent pixel is stored in the image buffer 240 for restoration processing. For example, for the pixel B, a pixel value is obtained from the pixels F, G, and H that have been subjected to the decoding processing before the decoding of the B, and are stored in the decoded image buffer 240. Similarly, for the pixel A, the pixel values obtained by the pixels E, F, and B are stored in the image buffer. The same applies to the pixel C.

As described above, when the decoded data 8 corresponding to the pixel of interest X is output from the decoding unit 210, the other pixels A, B, and C have already been decoded and the respective pixel values are stored in the decoded image buffer 240. Is stored. Prediction unit 2
20 has the same configuration as the prediction unit 120 of the encoding apparatus, and decode data corresponding to the pixel X of interest is decoded by the decoding unit 210.
, The decoding reference pixel value 11 that is the same as when the pixel of interest X is encoded, based on the pixel values of the pixels A, B, and C stored in the decoded image buffer 240.
Is output from the prediction unit 220.

Next, as step 32, the adder 230
Then, the decoded data 8 and the reference pixel value for decoding 11 are added, and are stored in the decoded image buffer 240 to decode the next pixel value to be processed in step 33, and are output as image data 9 in step 34. You. In the description of the above embodiment, the prediction direction signal is determined by three evaluation expressions. However, the prediction direction signal is determined by only the two evaluation directions of the BA direction and the BC direction in FIG. 7, that is, the prediction direction signal is not considered in the BD direction. Even when a signal is obtained, coding efficiency can be improved as compared with the conventional method. In this case, it is possible to reduce the calculation resources such as the prediction direction calculator and the processing time as compared with the above embodiment.

Further, the pixels in the regions II and III in FIG. 10 have been described in the embodiment in which the pixel values immediately above and immediately to the left are used as reference pixel values, but fixed values specified in advance may be used. The fixed value may be set as a parameter for each image to be encoded. FIG. 14 is a diagram showing a case where prediction errors of the conventional method and the present invention are compared with a block of 4 × 4 pixels as an original image. In the conventional method b), the prediction evaluation formula (7) in FIG. 1 is used. Although the pixel value greatly changes in the boundary portion in the image, the prediction error in such a portion is large in the conventional method.

On the other hand, in the present invention, the prediction error can be greatly reduced by adaptively switching the prediction direction by utilizing the strength of the correlation in the horizontal or vertical direction. As described above, since the variance of the prediction error is reduced by the present invention, entropy coding can be used effectively, and as a result, coding efficiency can be improved.

The configuration of the encoding device and the decoding device in the above embodiment is not limited to the above example. Each component and processing procedure are constructed by software (program), and the program is stored in a computer system. When executed, it is possible to encode and decode an image according to the method of the present invention. FIG. 15 is a block diagram showing an example of a hardware configuration of the computer system. The computer system includes a CPU that executes processing.
401, a memory 402 for storing programs and data,
It comprises an external storage device 403 for storing programs and data, a display 404 for displaying data, a keyboard 405 for inputting data or instructions, and a communication processing device 406 for communicating with other computer systems via a network. . The above program is CPU4
01.

The above-mentioned program is stored in a recording medium, and the program stored in the recording medium is executed by a computer system, so that an existing computer system can be coded or decoded to execute the method of the present invention. Can be used as Further, the program recorded on the recording medium can be preinstalled in a computer system.

Such a recording medium is stored in the memory 402 or the external storage device 40 in the computer system described above.
Equivalent to 3. Memory 402 or external storage device 403
A program for executing the processing according to the present invention is stored in
When the CPU executes the program, encoding and decoding of the image according to the present invention are performed. Further, the above-described recording medium can be realized by an electronic memory, a hard disk, or a portable recording medium such as a floppy disk, a magneto-optical disk, or a magnetic tape.

It should be noted that the present invention is not limited to the above-described embodiments, but can be variously modified and applied within the scope of the claims.

[0040]

According to the present invention, if there is no encoding loss when encoding image data and the loss on the transmission line is ignored, the image data 9 output from the decoding device is input to the encoding device. Since the image data is the same as the image data 1 and the image can be completely reproduced, lossless encoding can be realized. In addition, since the reference pixel value used in encoding and decoding is selected and generated depending on the strength of the correlation between the peripheral pixels, the difference from the pixel value of interest can be made smaller than in the conventional method. That is, the prediction performance can be improved.

Therefore, according to the present invention, the value of the data to be encoded can be made smaller than that of the conventional method, and the decoding can be performed uniquely by the peripheral pixel values. do not need. That is, the coding efficiency can be improved as compared with the conventional method. As described above, according to the present invention, the correlation in the vicinity of the encoding target pixel is evaluated, and prediction is performed from the direction in which the correlation is determined to be strong. Can be performed.

Also, since the strength of the correlation is determined from the pixel values that have already been encoded, efficient encoding can be performed without requiring additional information for switching the prediction direction.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining a prediction method in a conventional technique.

FIG. 2 is a configuration diagram of an encoding device according to an embodiment of the present invention.

FIG. 3 is a flowchart illustrating an operation of the encoding device according to the embodiment of the present invention.

FIG. 4 is a configuration diagram of a prediction unit in the encoding device.

FIG. 5 is a flowchart illustrating an operation of a prediction unit.

FIG. 6 is a configuration diagram of a prediction direction selection unit in the prediction unit.

FIG. 7 is a diagram for explaining a method of determining a prediction direction.

FIG. 8 is a correspondence table showing prediction directions corresponding to directions having the strongest correlation.

FIG. 9 is a diagram for explaining a direction in which an encoding process is performed.

FIG. 10 is a diagram illustrating a method of determining a reference pixel value for each region of an image.

FIG. 11 is a flowchart illustrating a reference pixel value determining method.

FIG. 12 is a configuration diagram of a decoding device according to an embodiment of the present invention.

FIG. 13 is a flowchart illustrating an operation of the decoding device according to the embodiment of the present invention.

FIG. 14 is a diagram showing a case where prediction errors of the conventional method and the present invention are compared.

FIG. 15 is a configuration diagram of a computer system.

[Explanation of symbols]

 1, 9 Image data 2 Encoding target pixel 3 Surrounding pixel value 4 Encoding reference pixel value 5 Encoding target data 6 Compressed data 8 Decoded data 10 Decoded pixel value 11 Decoding reference pixel value 50 Pixel value 51 Prediction direction signal 52 Reference pixel value 110 Image buffer 120 Prediction unit 130 Difference unit 140 Encoding unit 210 Decoding unit 220 Prediction unit 230 Adder 240 Decoded image buffer 310 Prediction direction selection unit 320 Reference pixel setting unit 401 CPU 402 Memory 403 External storage device 404 Display 405 keyboard 406 communication processing device

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Noriko Yonehara 223-1 Yamashita-cho, Naka-ku, Yokohama-shi, Kanagawa Prefecture Inside NTT Software Corporation (72) Inventor Takao Soma Yamashita, Naka-ku, Yokohama-shi, Kanagawa 223-1 Machi NTT Software Co., Ltd. F-term (reference) 5C059 KK00 KK11 MA00 MA02 MA04 PP01 SS15 TA16 TA17 TB10 TC02 TC03 TC42 TD05 TD11 UA02 UA05 5C078 BA32 CA02 DA00 DA01 DA02 DB13 9A001 EE04 HH27

Claims (3)

[Claims] An encoding method for encoding image information, comprising: an evaluation step of evaluating the strength of correlation between surrounding pixels of an encoding target pixel; and a reference pixel value based on a result of the evaluation. An encoding method comprising: determining; and calculating a difference value between the reference pixel value and the encoding target pixel value, and encoding the difference value. 2. The evaluation step includes: a difference step of obtaining at least a difference value between pixels in a vertical direction and a difference value between pixels in a horizontal direction in a plurality of peripheral pixels adjacent to the encoding target pixel; And comparing the plurality of difference values obtained by the above method. 3. A decoding method for decoding encoded image information, comprising the steps of: restoring encoded data to obtain a restoration value; and evaluating the strength of correlation between pixels surrounding the decoding target pixel. An evaluation step; a step of determining a reference pixel value based on the evaluation result; a step of adding the reference pixel value and the restored value to obtain decoded data; and a step of using the decoded data as surrounding pixel values. A decoding method comprising: