JPH05199517A - Method and device for coding image - Google Patents

Abstract

PURPOSE:To provide the method and device for coding image with high efficiency for image coding used for the case of compressing, transmitting and recording digital images without degrading picture quality. CONSTITUTION:In the method for coding at every block after dividing the image into blocks consisting of plural picture elements approaching to the image and predicting the image as the object of coding while using reference images hourly in the past and future, an independent block detector 1 is provided to detect the block, which is not present on the reference image in the past rather than the coding object image and is not present on the reference image in the future rather than the coding object image although being present in the coding object image, included in the image area of the coding object image as an independent block and in the detected independent block, a coding is independently processed so as to remove the after-image noise of a mobile object without depending on the reference images in the past and future.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding method and an image coding device used for compressing, transmitting and recording a digital image.

[0002]

2. Description of the Related Art Conventionally, as a method of encoding a digital image, an image is divided into blocks made up of a plurality of adjacent pixels, and a plurality of images positioned temporally before and after the target image are used as reference images. A method of predicting an image and coding the prediction error for each block is used. In this method, the pixel level variance of the block to be encoded is compared with the variance of the error signal of the prediction error block to directly encode the target block or to encode the prediction error. Is selected.

[0003]

However, the above-mentioned method has a problem that afterimage noise of the moving body is generated and image quality is deteriorated. This is because the movement of an object in the image causes an image that is present in the target image but does not exist in the images that are temporally located before and after, and it is difficult to predict the region from the reference images before and after. This is because the coding noise increases and becomes afterimage noise. In addition, the afterimage noise generated in the region where the amount of change in the pixel level of the image is small is particularly noticeable, and causes a remarkable deterioration of the image quality.

It is an object of the present invention to provide an image coding method and an image coding apparatus that removes afterimage noise without degrading image quality and that is highly efficient.

[0005]

According to the present invention, an image is divided into blocks composed of a plurality of adjacent pixels, and an image to be coded is predicted using a temporally past reference image and a future reference image. In the method of encoding for each, in the image to be encoded, but does not exist in the image from the image to be encoded to the past reference image, and from the image to be encoded to the future reference image Of the image to be encoded, which does not exist in the image of the target image, is detected as an independent block, and the independent block is independently coded without depending on the past reference image and the future reference image. It is configured to perform a chemical conversion process.

Further, the image coding apparatus of the present invention is an apparatus for realizing the above-described image coding method, and divides an image into blocks composed of a plurality of adjacent pixels, and a temporally past reference image and In a device that predicts an encoding target image using a future reference image and encodes it for each block, it exists in the encoding target image, but exists in the images from the encoding target image to the past reference image. Independent block detector that detects, as an independent block, a block including an image region in the encoding target image that does not exist in an image from the encoding target image to a future reference image, and the independent block detector In the independent block detected by the above, the independent block does not depend on the past reference image and the future reference image, and is constituted by means for independently performing encoding processing. That.

[0007]

As a result, an image region which does not exist in the images before and after the image to be encoded but exists in the image to be encoded is detected, and in the region, the encoding process is performed independently without depending on the reference image. By doing so, the afterimage of the moving body is removed and the image quality is improved.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the image coding method and the image coding apparatus of the present invention will be described below in detail with reference to the drawings. Here, the image is processed on a frame-by-frame basis, and past frames and future frames are used as reference images used for prediction, as compared with the image to be encoded. Also, motion detection is performed using each of the reference images, and a target block to be encoded is predicted by a prediction signal block obtained by performing motion compensation for each block from a reproduced image of the reference images that have already been decoded. Then, a method of encoding the prediction error signal for each block will be described.

FIG. 1 shows a first embodiment of the basic configuration of the present invention. In FIG. 1, the past reference block is
The blocks in the reference image past the encoding target image at the same position as the encoding target block, and the blocks from the front block 1 to the front block N are located between the encoding target image and the past reference image. This is a block located at the same position as the encoding target block in each frame when N existing frames are arbitrarily designated. Further, the future reference block is a block in the reference image in the future of the encoding target image at the same position as the encoding target block, and the blocks from the backward block 1 to the backward block M are the encoding target image and the encoding target image. It is a block located at the same position as the encoding target block in each frame when M frames which are present between future reference images are arbitrarily designated. The prediction error block is a block of the prediction error signal when the coding target block is predicted by the prediction signal. Further, in the selector 2, when the inputs are a, b, and c in FIG. 2, the output d is a when c = 1, and the other outputs d are b.

First, past reference blocks, 1 to N forward blocks, future reference blocks, 1 to M backward blocks, and blocks to be encoded are input to the independent block detector 1, and the blocks to be encoded are past blocks. Detects a reference block, a future reference block, and an independent block that does not depend on any of the front block and the rear block. When the block to be coded is detected as an independent block, the block 2 to be coded is input to the coder (not shown) by the selector 2 and coded independently of the reference image. That is, intraframe coding is performed.

On the other hand, when the block to be coded is not an independent block, the prediction error block is input to the coder by the selector 2 and interframe coding is performed. In this way, by detecting whether the reference image is independent or dependent on the reference image, it is possible to remove the afterimage noise of the moving object by adaptively performing the intraframe coding or the interframe coding, thereby improving the image quality. can get.

Next, a second embodiment of the present invention will be described with reference to FIG. In FIG. 3, as in the first embodiment, the past reference block and the future reference block are a block in the past reference image and a block in the future reference image that are at the same position as the encoding target block, respectively. To N forward blocks and 1 to M backward blocks are coding target blocks in a frame between a coding target image and a past reference image and in a frame between a coding target image and a future reference image, respectively. The blocks at the same position, and the prediction error block is a block of the prediction error signal when the target block to be coded is predicted by the prediction signal. Further, 22 and 23, and 24 are a logical sum device and a logical product device for calculating and outputting a logical sum and a logical product, respectively.

The change amount measuring devices 3 to 9 have the same configuration and calculate and output the pixel level variance as the change amount of the pixel level in the block. Also, the subtracters 10 to 15
Outputs the magnitude of the difference between the two input changes. The comparators 16 to 21 are the comparators 2 in FIG.
5 has the same configuration as that of FIG.
When a is larger than b with respect to b, the output c is 1, and otherwise the output c is 0.

First, the level change amount of pixels in the past reference block, the future reference block, each of the front block and the back block, and the block to be encoded is measured by each change amount measuring device. Next, regarding the measured level change amount in the past reference block, the future reference block, and each of the front block and the rear block, the magnitude of the difference between the measured level change amount in the encoding target block is individually determined. And the difference between each
Compare with T0.

If at least one of the level change amounts in the past reference block or the preceding block and the level change amount in the encoding target block is larger than CONST0, and the level in the future reference block or the backward block is large. If at least one of the changes is greater than CONST0 in the level change difference in the block to be coded, the block to be coded is recognized as an independent block, and the block to be coded is further selected by the selector 2. It is input to the encoder and is subjected to intraframe coding.
If the above two cases do not hold,
The prediction error block is input to the encoder for interframe coding.

When there is no frame between the past reference image and the image to be encoded or between the future reference image and the image to be encoded, N = 0 or M = 0 is substituted, and the forward block or The rear block does not exist. In addition, a predetermined amount CONST
A value of 0 is preferably around 500 when the pixel level is 256 gradations. This is because the block with a small change in pixel level has a variance of 30 or less, and the complex block with a large change in pixel level has a variance of more than 500. However, CONST0 is not limited to the above value, and efficient coding can be performed by adaptively changing the value depending on the image. By adaptively performing the intra-frame coding in this way, the afterimage noise of the moving body is removed and the image quality is improved.

Next, a third embodiment will be described with reference to FIG. In FIG. 5, the past block indicates the past reference block described in the first and second embodiments or a plurality of blocks in which the past reference block and the forward block are combined, and the future block is described in the first and second embodiments. Of the future reference block or a plurality of blocks in which the future reference block and the backward block are combined, and like the first and second embodiments,
The prediction error block is a block of the prediction error signal when the target block to be encoded is predicted by the prediction signal.

The change amount measuring devices 26 to 29 have the same configuration and calculate and output the pixel level variance as the change amount of the pixel level in each block. Also, the comparator 30-
34 is the same as the comparator 25 of FIG. Also, 3
Reference numerals 5 to 37 and 38 are an AND unit and an OR unit which calculate and output an AND and an OR, respectively. The selector 2 is the same as that of the first embodiment, and the block to be coded is input to the encoder by the output of the logical adder 38 to perform intra-frame encoding, or the prediction error is input to the encoder. It is for selecting whether to input and perform interframe coding. Further, CONST1, CONCT2, and CONST3 are predetermined constants.

In the conventional method, the intra-frame coding or the inter-frame coding is selected by comparing the pixel level change amount of the coding target block with the pixel level change amount of the prediction error block. That is, the change amount measuring device 28,
29, comparators 33 and 34, a logical product 37, and a selector 2. Further, the comparator 34 is for selecting inter-frame coding more often than selecting intra-frame coding. This is because when the Discrete Cosine Transform (DCT) coding method is used, the amount of information required for coding the DC component of the DCT is large in intra-frame coding, which may lead to a decrease in coding efficiency. There is to prevent this.

First, the variance of the pixel level of the past block is calculated by the change amount measuring device 26 and compared with the predetermined amount CONST1 by the comparator 30. Similarly, the pixel level variance of the past block is calculated by the change amount measuring device 27, and the comparator 31
Is compared with the predetermined amount CONST1. In addition, the variance of the pixel level of the block to be encoded is calculated by the change amount measuring device 28, and is compared with the predetermined amount CONST2 by the comparator 32. When the variance of the past block is greater than CONST1, the variance of the future block is greater than CONST1, and the variance of the encoding target block is less than CONST2, the encoding target block is an independent block independent of any reference image. Recognized, the target block is input to the encoder by the selector 2, and intra-frame encoding is performed.

In this case, it means that the portion where the pixel level change is complicated in the past frame and the future frame is a flat portion where the pixel level change is small in the current frame. In such a block including the portion of the current frame, it is highly possible that both the past reference block and the future reference block to be referred to include an image that does not exist in the block of the current frame, and the prediction error signal is encoded. If the inter-frame coding is performed, afterimage noise is generated and the image quality is deteriorated. However, if the image is encoded independently of the reference frame, the afterimage noise can be removed.

Next, the past block and the future block are blocks in which the pixel level changes in a complicated manner, but when the variance of the block to be coded is larger than CONST2 and it is not recognized as a flat part in which the pixel level change is small, Similarly to the conventional method, the variance of the target block and the variance of the prediction error block are compared in the comparator 33, and the intra-frame coding or the inter-frame coding is performed according to the magnitude. However, it does not depend on the output of the comparator 32. If the reference block is not a block in which the pixel level changes intricately, intra-frame coding or inter-frame coding is selected and coded as usual.

In FIGS. 6A and 6B, areas for selecting intra-frame coding and inter-frame coding are shown. In this way, intra-frame coding is performed on a region in which afterimage noise is particularly generated, in which a complicated region in which the pixel level change amount is large in the reference image is a flat region in which the pixel level change amount is currently small. As a result, the image quality can be improved without causing afterimage noise.

The predetermined amount CON in the third embodiment
ST1, CONST2, and CONST3 have a pixel level of 2
56, 500, 30, 1000 respectively
Before and after is suitable, because the variance of a flat block having a small amount of level change of pixels is 30 or less, and the variance of a complex block having a large amount of level change is over 500. However, the present invention is not limited to this, and efficient coding can be performed by adaptively changing the image.

Although the image is handled in frame units, it may be handled in field units. Further, using the past block and the future block is not a necessary condition, and only the past block or the future block may be used. Furthermore, in the change amount measuring device of the embodiment, the variance of the pixel level in the block is calculated and output, but not limited to this, for example, the sum of the absolute values of the differences between the pixel level in the block and its average value, Any scale may be used as long as it is a scale indicating the amount of change in the pixel level, such as the sum of the difference values from the surrounding pixel levels.

Since the reference image is already encoded when the image to be encoded is encoded, it is not necessary to measure the amount of change in the pixel level in the block again, and the reference image is measured at the time of encoding. The amount of change may be delayed by a delay element before use. Further, the independent block may be detected only when a plurality of pixel levels of pixels in the encoding target block are within a predetermined amount range. This is because there is a pixel level range where afterimage noise is easily noticeable to human eyes, and processing is performed only within this range. As an example of this, the average value of the pixel levels of the pixels in the target block may be measured, and the independent block may be detected only when the average value is within a predetermined range. Further, not only the average value but also the pixel levels of a plurality of representative pixels may be detected within a predetermined range. When the pixel level has 256 tones, the predetermined range is appropriately 120 to 180, but is not limited to this.

When the image signal is composed of a luminance signal and a color signal, independent blocks may be detected and processed, or an independent block may be detected only by the luminance signal.
If the luminance signal is an independent block, the chrominance signal block may be processed as an independent block. Also, only when it is recognized as an independent block in the luminance signal,
Independent signals may also be detected and processed in the color signal.

[0028]

As described above, the present invention divides an image into blocks composed of a plurality of adjacent pixels, predicts an encoding target image using temporally past and future reference images, and encodes each block. In the encoding method, the reference image exists in the encoding target image, but does not exist in images past the encoding target image and past reference images, and is a future reference image in the future from the encoding target image. To a block including an image area in the image to be coded that does not exist in the images up to the independent image, the independent block does not depend on the past reference image and the future reference image, By configuring the encoding processing to be performed, the afterimage noise of the moving object can be removed, and the image quality can be improved.

Further, in the apparatus for dividing an image into blocks made up of a plurality of pixels adjacent to each other, predicting an encoding target image temporally using past and future reference images, and encoding each block, It exists in the target image but does not exist in images past and past the reference image from the encoding target image, and does not exist in images up to the future reference image in the future from the encoding target image. An independent block detector that detects a block including an image region in an image to be encoded as an independent block, and an independent block detected by the independent block detector depends on the past reference image and the future reference image. It is possible to remove the afterimage noise of the moving body by configuring it with a means for independently performing encoding processing without
Image quality improvement can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram of an image coding apparatus according to a first embodiment of the present invention.

FIG. 2 is a block diagram of a selector according to the first embodiment of the present invention.

FIG. 3 is a block diagram of an image coding apparatus according to a second embodiment of the present invention.

FIG. 4 is a block diagram of a comparator according to a second embodiment of the present invention.

FIG. 5 is a block diagram of an image coding apparatus according to a third embodiment of the present invention.

FIG. 6 is an explanatory diagram of selected regions for intra-frame coding and inter-frame coding according to the third embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Independent block detector 2 Selector 3-9 Change amount measuring device 10-15 Difference device 16-21 Comparator 22 Logical sum device 23 Logical sum device 24 Logical product device 26-29 Change amount measuring device 30-34 Comparator 35 ~ 37 logical ANDer 38 logical OR

Claims (10)

[Claims] 1. An image is divided into blocks made up of a plurality of adjacent pixels, and an encoding target image is predicted using a past reference image temporally located in the past and a future reference image located in the future, and each block is predicted. In the encoding method, any image up to the past reference image in the past from the encoding target image including the past reference image is set as a forward image, and the encoding target image including the future reference image is set in the future. When an arbitrary image up to a certain future reference image is set as a rear image,
A block including an image region in the encoding target image which is present in the encoding target image and which is not present in the forward image and the backward image is detected as an independent block independent of the past reference image and the future reference image, and the independent block is detected. An image coding method characterized in that a block is independently coded without depending on either the past reference image or the future reference image. 2. A block located at the same position as a target block in an encoding target image in a front image is defined as a front block, and a block located at the same position as a target block in the encoding target image in a rear image. As a rear block, the difference between the level change amount of the pixel included in the front block and the level change amount of the pixel included in the target block is measured as a front difference, and the level change amount of the pixel included in the rear block and the The difference from the level change amount of the pixels included in the target block is measured as a rear difference, and when the front difference is larger than a first predetermined amount and the rear difference is larger than a second predetermined amount, the target block is independent. The image coding method according to claim 1, wherein the image is detected as a block. 3. A block located at the same position as the target block in the encoding target image in the front image is defined as a front block, and a block located at the same position as the target block in the encoding target image in the rear image is defined. As a rear block, the level change amount of pixels included in the target block in the encoding target image is equal to or less than a third predetermined amount, and the level change amount of pixels included in the front block is equal to or greater than a fourth predetermined amount. The image coding method according to claim 1, wherein the target block is detected as an independent block when the level change amount of pixels included in the rear block is equal to or larger than a fifth predetermined amount. 4. The image coding method according to claim 2, wherein independent block detection is performed on the target block when the pixel levels of a plurality of pixels in the target block are within a predetermined range. .. 5. When the image signal is composed of a luminance signal and a color signal, the independent block is always detected for the luminance signal, but is always detected for the color signal or the independent block is detected in the luminance signal. Only when the independent block is detected, or the independent block is not detected in the chrominance signal and the block located at the same position as the independent block of the luminance signal is set as the independent block in the chrominance signal. 1. The image coding method described in 1. 6. An image is divided into blocks made up of a plurality of adjacent pixels, and a coding target image is predicted by using a past reference image temporally located in the past and a future reference image located in the future, and each block is predicted. A device for encoding, wherein any image up to the past reference image that is past the encoding target image including the past reference image is a forward image, and the encoding target image including the future reference image A block including an image area in the encoding target image which is present in the encoding target image and is not present in the forward image and the backward image when an arbitrary image in the future and up to the future reference image is set as a rear image An independent block detector that detects as an independent block independent of the past reference image and the future reference image, and an independent block detected by the independent block detector. In addition, the image coding apparatus is provided with an encoder that independently performs coding processing without depending on any of the past reference image and the future reference image. 7. The independent block detector comprises a change amount measuring device for measuring a level change amount of a pixel included in a block in an image, and a target block in an encoding target image obtained by the change amount measuring device. A front differencer that outputs the difference between the level change amount and the level change amount of the front block in the front image located at the same position as the target block, and the target block in the encoding target image obtained by the change amount measuring device And a rear differencer that outputs the difference between the level change amount of the rear block in the rear image located at the same position as the target block and the output of the front difference device and the sixth predetermined amount. A front comparator, a rear comparator that compares the output of the rear differencer with a seventh predetermined amount, the output of the front differencer is greater than a sixth predetermined amount and the output of the rear differencer is Seventh The image coding apparatus according to claim 6, wherein the target block is detected as an independent block when the target block is larger than a predetermined amount. 8. The independent block detector is a change amount measuring device for measuring a level change amount of a pixel included in a block in an image, and a level change amount of a target block in an encoding target image is an eighth predetermined amount. A first comparator for detecting that the level change amount of the pixel included in the front block in the front image located at the same position as the target block is larger than a ninth predetermined amount. And a rear comparator for detecting that a level change amount of a pixel included in a rear block in a rear image at the same position as the target block is larger than a tenth predetermined amount, the level of the target block When the change amount is smaller than the eighth predetermined amount, the level change amount of the front block is larger than the ninth predetermined amount, and the level change amount of the rear block is larger than the tenth predetermined amount, the pair The image coding apparatus according to claim 6, wherein the elephant block is an independent block. 9. The independent block detector comprises means for detecting that the pixel levels of a plurality of pixels in the target block are within a predetermined range, and means for detecting that the target block is an independent block. 9. The image coding apparatus according to claim 7, wherein an independent block is detected for the target block when the pixel levels of the plurality of pixels are within a predetermined range. 10. When the image signal is composed of a luminance signal and a color signal, the independent block is always detected for the luminance signal, but is always detected for the color signal or the independent block is detected in the luminance signal. Only when the independent block is detected, or the independent block is not detected in the chrominance signal and the block located at the same position as the independent block of the luminance signal is set as the independent block in the chrominance signal. 6. The image encoding device according to item 6.