WO2010035993A2

WO2010035993A2 - Apparatus and method for image encoding/decoding considering impulse signal

Info

Publication number: WO2010035993A2
Application number: PCT/KR2009/005362
Authority: WO
Inventors: 김하윤; 문주희; 이영렬; 김해광; 전병우; 김기백; 박형미
Original assignee: 에스케이텔레콤 주식회사
Priority date: 2008-09-25
Filing date: 2009-09-21
Publication date: 2010-04-01
Also published as: WO2010035993A3

Abstract

The present invention relates to an apparatus and a method for image encoding/decoding considering an impulse signal. The method for image encoding according to the present invention comprises: an MxN residual block generating step of estimating the current block to generate an estimated block, and performing a subtraction between the estimated block and the current block to generate an MxN residual block; and an AxB encoding step of encoding the AxB residual block containing the residual signal of an impulse component in the MxN residual block to generate a bit stream. The apparatus and the method of the present invention improve encoding efficiency by efficiently encoding or decoding the residual signal of the impulse component in encoding or decoding images.

Description

Image Encoding / Decoding Apparatus and Method Considering Impulse Signals

The present invention relates to an image encoding / decoding apparatus and method considering an impulse signal. More particularly, the present invention relates to an apparatus and method for improving encoding efficiency by efficiently encoding or decoding a residual signal of an impulse component in encoding or decoding an image.

Moving Picture Experts Group (MPEG) and Video Coding Experts Group (VCEG) have developed video compression techniques that are superior to the existing MPEG-4 Part 2 and H.263 standards. The new standard is called H.264 / AVC (Advanced Video Coding) and was jointly released as MPEG-4 Part 10 AVC and ITU-T Recommendation H.264.

In such H.264 / AVC (hereinafter abbreviated as 'H.264'), integer discrete cosine transform (DCT) is called, and motion estimation and compensation of transform block size (Variable Block Size) Motion Estimation and Compensation, Quantization, Entropy Coding.

The encoding method of the image data according to H.264 / AVC can be largely classified into intra prediction encoding and inter prediction encoding according to the prediction method. Intra prediction predicts pixels of a current block to be encoded in one reference picture by using pixels of a neighboring block located in the periphery thereof. The inter prediction predicts the pixels of the current block by using the pixels of the block most similar to the current block in the reference picture in one or both directions.

Most image compression techniques such as H.264 compress image data by encoding only the difference between the prediction pixel predicted through inter prediction or intra prediction and the original pixel, thereby removing temporal redundancy and spatial redundancy of the image. As the difference between the prediction pixel and the original pixel is smaller, the size of the compressed image data is smaller, thereby improving the compression efficiency.

Therefore, in the field of image compression, various techniques such as a technique for determining whether to perform inter prediction or intra prediction according to characteristics of an image, a prediction technique for increasing the accuracy of prediction, and the like to improve compression efficiency by increasing the accuracy of prediction. Prediction and coding techniques are being developed.

However, for various reasons, a pixel may not always be accurately predicted. In this case, the difference between the pixel that is not accurately predicted and the original pixel is larger than that of other pixels. As such, a difference between the original pixel and the predicted pixel, that is, a specific residual signal larger than other residual signals of the residual signal is called a residual signal of an impulse component, and the residual signal of the impulse component has a problem of lowering compression efficiency. .

In order to solve the above-mentioned problems, the present invention has a main object in improving encoding efficiency by efficiently encoding or decoding a residual signal of an impulse component in encoding or decoding an image.

According to an aspect of the present invention, there is provided a method of encoding an image, the method comprising: generating an MxN residual block by predicting a current block to generate a prediction block and subtracting the current block and the prediction block to generate an MxN residual block; And an AxB encoding step of encoding the AxB residual block including the residual signal of the impulse component in the MxN residual block to generate a bitstream.

According to another object of the present invention, there is provided an apparatus for encoding an image, comprising: a prediction unit for predicting a current block to generate a prediction block; A subtraction unit for generating an MxN residual block by subtracting the current block and the prediction block; And an AxB encoder that generates a bitstream by encoding an AxB residual block including a residual signal of an impulse component in the MxN residual block.

According to still another object of the present invention, there is provided a method of decoding an image, comprising: a decoding step of decoding a bitstream to extract a quantized frequency coefficient sequence; Generating an AxB residual block by performing inverse scan, inverse quantization, and inverse transformation on the quantized frequency coefficient sequence in units of AxB blocks; Generating an MxN residual block by further combining one or more residual signals to the AxB residual block to generate an MxN residual block; A prediction step of predicting a current block to generate a prediction block; And a reconstruction step of reconstructing the current block by adding the prediction block and the MxN residual block.

According to still another object of the present invention, there is provided an apparatus for decoding an image, comprising: a decoder configured to decode a bitstream to extract a quantized frequency coefficient sequence; An AxB residual block generator configured to inversely scan, inverse quantize, and inversely convert the quantized frequency coefficient sequence into AxB block units to generate an AxB residual block; An MxN residual block generator for further combining one or more residual signals to the AxB residual block to generate an MxN residual block; A prediction unit for predicting a current block to generate a prediction block; And an adder configured to reconstruct the current block by adding the prediction block and the MxN residual block.

As described above, according to the present invention, in encoding or decoding an image, encoding efficiency can be improved by efficiently encoding or decoding a residual signal of an impulse component.

1 and 2 are exemplary views illustrating a prediction according to intra prediction and inter prediction.

3 is a block diagram schematically illustrating an electronic configuration of a video encoding apparatus;

4 is an exemplary diagram showing a residual signal before conversion and a frequency coefficient after conversion;

5 is an exemplary view showing how an impulse affects a frequency domain;

6 is a block diagram schematically illustrating a video encoding apparatus according to a first embodiment of the present invention;

7 is an exemplary view showing the configuration of an AxB residual block according to the present invention;

8 is an exemplary view illustrating setting an arbitrary value to the remaining residual signal according to the present invention;

9 is an exemplary view illustrating a process of converting a frequency domain in units of AxB blocks according to the present invention;

10 is an exemplary view illustrating a scan method according to an AxB block form according to the present invention;

11 is a flowchart for explaining a video encoding method according to a first embodiment of the present invention;

12 is a block diagram schematically illustrating a configuration of an image decoding apparatus according to a first embodiment of the present invention;

13 is a flowchart illustrating a video decoding method according to a first embodiment of the present invention;

14 is a block diagram schematically illustrating a configuration of a video encoding apparatus according to a second embodiment of the present invention;

15 is a flowchart for explaining a video encoding method according to a second embodiment of the present invention;

16 is a block diagram schematically showing the configuration of a video encoding apparatus according to a third embodiment of the present invention;

17 is an exemplary view showing the configuration of an AxB residual block in a diagonal block form according to a third embodiment of the present invention;

18 is a block diagram schematically illustrating the configuration of a video encoding apparatus according to a fourth embodiment of the present invention;

19 is a flowchart for explaining a video decoding method according to a fourth embodiment of the present invention;

20 is a flowchart illustrating a video encoding method according to a fifth embodiment of the present invention.

Hereinafter, some embodiments of the present invention will be described in detail through exemplary drawings. In adding reference numerals to the components of each drawing, it should be noted that the same reference numerals are assigned to the same components as much as possible even though they are shown in different drawings. In addition, in describing the present invention, when it is determined that the detailed description of the related well-known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

In addition, in describing the component of this invention, terms, such as 1st, 2nd, A, B, (a), (b), can be used. These terms are only for distinguishing the components from other components, and the nature, order or order of the components are not limited by the terms. If a component is described as being "connected", "coupled" or "connected" to another component, that component may be directly connected or connected to that other component, but between components It will be understood that may be "connected", "coupled" or "connected".

1 and 2 are exemplary views illustrating the prediction according to intra prediction and inter prediction.

Intra prediction includes intra 4x4 prediction, intra 16x16 prediction, and intra 8x8 prediction, and each intra prediction includes a plurality of prediction modes. 1 illustrates nine prediction modes in intra 4x4 prediction.

Referring to FIG. 1, intra 4x4 prediction includes vertical mode, horizontal mode, direct current mode, diagonal down-left mode, diagonal down-right, and vertical. There are nine prediction modes, including the right-right mode, horizontal-down mode, vertical-left and horizontal-up modes.

Although not shown, intra 8x8 prediction has the same prediction mode as intra 4x4 prediction, and intra 16x16 prediction includes vertical mode, horizontal mode, direct current mode, and plane mode. There are four prediction modes.

Inter prediction predicts pixels using a motion estimation technique and a motion compensation technique. Referring to FIG. 2, an image is composed of a series of still images. These still images are classified in units of group of pictures (GOP). Each still image is called a picture or frame. One picture group includes an I picture, a P picture, and a B picture. The I picture is a picture that is encoded by itself without using a reference picture, and the P picture and the B picture are pictures that are encoded by performing motion estimation and motion compensation using the reference picture. In particular, a B picture is a picture that is encoded by predicting a picture of a past picture and a picture of a future in a forward direction and a reverse direction, respectively.

Motion estimation and motion compensation for encoding a P picture use an I picture as a reference picture. Motion estimation and motion compensation for encoding a B picture use an I picture and a P picture as reference pictures.

3 is a block diagram schematically illustrating an electronic configuration of a video encoding apparatus.

The image encoding apparatus 300 may include a predictor 310, a subtractor 320, a transformer 330, a quantizer 340, a scan unit 350, an encoder 360, an inverse quantizer 370, It may be configured to include an inverse transform unit 380 and an adder 390.

The prediction unit 310 generates a prediction block by predicting a current block to be currently encoded in an image. That is, the prediction unit 310 predicts the pixel value of each pixel of the current block to be encoded in the image according to a predetermined optimal prediction mode, and predicts the predicted pixel value of each pixel. A predicted block having a pixel value) is generated. In addition, the prediction unit 310 may transmit the information about the prediction mode to the encoder 360 so that the encoder 360 may encode the information about the prediction mode.

The subtraction unit 320 generates a residual block by subtracting the prediction block from the current block. That is, the subtractor 320 calculates a difference between the pixel value of each pixel of the current block to be encoded and the predicted pixel value of each pixel of the prediction block predicted by the prediction unit 310 to obtain a residual signal in the form of a block. Create a residual block with).

The converter 330 converts the residual block into the frequency domain to convert each pixel value of the residual block into a frequency coefficient. Here, the transform unit 330 uses various transformation techniques for transforming an image signal of a spatial axis into a frequency axis, such as a Hadamard transform and a Discrete Cosine Transform Based Transform (DCT based Transform). The residual signal can be converted into the frequency domain, and the residual signal converted into the frequency domain becomes a frequency coefficient.

The quantization unit 340 quantizes the residual block having the frequency coefficient transformed by the transformer 330 into the frequency domain. Here, the quantization unit 340 converts the transformed residual block into Dead Zone Uniform Threshold Quantization (DZUTQ), a quantization weight matrix, or an improved quantization technique. Can be quantized using

The scan unit 350 scans the quantization frequency coefficients of the residual block quantized by the quantization unit 340 according to various scan methods such as a zigzag scan to generate a quantization frequency coefficient sequence.

The encoder 360 outputs the bitstream by encoding the quantized frequency coefficient sequence generated by the scan unit 350 using an entropy coding technique or the like. In addition, the encoder 360 may encode the information about the prediction mode in which the predictor 310 predicts the current block.

The inverse quantizer 370 inverse quantizes the residual block quantized by the quantizer 340. That is, the inverse quantization unit 370 inversely quantizes the quantized frequency coefficients of the Angularized residual block to generate a residual block having frequency coefficients.

The inverse transform unit 380 inverse transforms the residual block inversely quantized by the inverse quantizer 370. That is, the inverse transformer 380 inversely transforms frequency coefficients of the inverse quantized residual block to generate a residual block having a pixel value, that is, a reconstructed residual block. Here, the inverse transform unit 380 may perform inverse transform by using the transformed method used by the transform unit 330 as the inverse.

The adder 390 reconstructs the current block by adding the prediction block predicted by the predictor 310 and the residual block reconstructed by the inverse transform unit 380. The reconstructed current block is stored in the prediction unit 310 as a reference picture and used as a reference picture when encoding the next block of the current block or another block in the future.

Although not shown in FIG. 3, a deblocking filter unit (not shown) may be further connected between the predictor 310 and the adder 390. The deblocking filter unit deblocks filtering the current block restored by the adder 390. Here, the deblocking filtering refers to an operation of reducing block distortion generated by encoding an image in block units, and applying a deblocking filter to a block boundary and a macroblock boundary, or applying a deblocking filter only to a macroblock boundary or a deblocking filter. You can optionally use one of the methods that does not use.

On the other hand, when the transformer 330 converts the residual signal of the residual block into the frequency domain to generate a frequency coefficient, the residual signal is decomposed into a low frequency component and a high frequency component. As described above, when the decomposed frequency coefficient is quantized and the zigzag scan is performed by the scan unit 350, as shown in FIG. 5A, when the impulse component does not occur in the residual block, the frequency coefficient of the low frequency component is' 0. It has a non-value and is driven to the position of the frequency coefficient of the DC component, and the frequency coefficients of the high frequency component are '0' and can be ignored. Therefore, since only the Molly quantization frequency coefficient is encoded around the frequency coefficient of the DC component, as shown in Fig. 4, a high compression effect can be obtained. 4, it can be seen that the distribution of the frequency coefficient obtained by converting the residual signal, but the value other than '0' is driven around the frequency coefficient of the DC component.

However, when the accuracy of prediction is deteriorated and the residual signal of the impulse component occurs around the residual block, when the residual signal of the impulse component is converted into the frequency domain, it is converted into the frequency coefficient of the high frequency component, and 5B shown in FIG. Likewise, frequency coefficients of the high frequency component occur. The frequency coefficient of the high frequency component is located at the end of the quantization frequency coefficient sequence when zigzag scanning after performing the quantization process, resulting in a non-zero value at the end of the quantization frequency coefficient sequence. The amount of data in the stream increases. That is, if there is no residual signal of the impulse component, the data size of the encoded bitstream can be reduced, but as the residual signal of the impulse component occurs, the frequency coefficient of the high frequency component has a large value, and the amount of data of the encoded bitstream Increase and hence the efficiency of the compression decreases. Therefore, when the residual signal of such an impulse component is included in the residual block, it should be able to be efficiently encoded.

The present invention provides an image encoding / decoding method and apparatus for efficiently encoding when a residual signal of an impulse component is included in a residual block. In the present invention, the residual signal of the impulse component refers to a residual signal whose absolute value is greater than or equal to a preset value among the residual signals in the residual block.

6 is a block diagram schematically illustrating a video encoding apparatus according to a first embodiment of the present invention.

The image encoding apparatus 600 according to the first embodiment of the present invention includes a predictor 610, a subtractor 620, a converter 630, an MxN converter 640, an AxB converter 642, and an MxN quantization. 650, AxB quantizer 652, scan unit 660, encoder 670, MxN inverse quantizer 680, AxB inverse quantizer 682, MxN inverse transform unit 690, AxB inverse The conversion unit 692 and the adder 694 may be configured.

The video encoding apparatus 600 may be a personal computer (PC), a notebook computer, a personal digital assistant (PDA), a portable multimedia player (PMP), or a PlayStation Portable (PSP). ), A communication device such as a communication modem for communicating with various devices or a wired / wireless communication network, a memory for storing various programs and data for encoding an image, and executing a program. Means a variety of devices including a microprocessor for operation and control.

Since the predictor 610 and the subtractor 620 perform the same or similar functions as those of the predictor 310 and the subtractor 320 of the image encoding apparatus 300 described above with reference to FIG. 3, a detailed description thereof will be omitted. do. In addition, the MxN transform unit 640, the MxN quantizer 650, the MxN inverse quantizer 680, the MxN inverse transform unit 690, the AxB transform unit 642, the AxB quantizer 652, and the AxB inverse quantizer 682 and AxB inverse transform unit 692 are the same as the transform unit 330, the quantizer 340, the inverse quantizer 370, and the inverse transform unit 380 of the image encoding apparatus described above with reference to FIG. 3. We transform the residual block into the frequency domain, quantize the transformed frequency coefficients, and inversely quantize and inversely transform the quantized frequency coefficients.

However, the MxN transform unit 640, the MxN quantizer 650, the MxN inverse quantizer 680, and the MxN inverse transform unit 690 include a transform unit 330, a quantizer 340, and an inverse quantizer 370. Like the inverse transform unit 380, transformation, quantization, inverse quantization, and inverse transformation are performed in the same MxN block unit as the block size of the current block, and the transformation, quantization, inverse quantization, and inverse transformation are performed in AxB block units. The names of the AxB transform unit 642, the AxB quantizer 652, the AxB inverse quantizer 682, and the AxB inverse transform unit 692 are only different. On the other hand, the AxB transform unit 642, the AxB quantizer 652, the AxB inverse quantizer 682, and the AxB inverse transform unit 692 transform, residual, and quantize the residual block. 330, quantization unit 340, inverse quantization unit 370, and inverse transform unit 380, but different in that the residual block is transformed and quantized in AxB block units, and then inversely quantized and inversely transformed.

Hereinafter, in describing the AxB transform unit 642, the AxB quantizer 652, the AxB inverse quantizer 682, and the AxB inverse transform unit 692, the transform unit 330, the quantizer 340, The differences from the inverse quantizer 370 and the inverse transform unit 380 will be described.

The encoding mode determiner 630 determines whether to encode the residual block among the MxN residual block and the AxB residual block according to whether the residual signal of the impulse component is included in the MxN residual block transmitted from the subtractor 620. The residual block is transferred to the MxN converter 640 or the AxB converter 642.

That is, the encoding mode determiner 630 determines whether to encode the MxN residual block transmitted from the subtractor 620 in units of AxB blocks or in units of MxN blocks, thereby MxN residual blocks or AxB residual blocks, respectively. The data is transmitted to the converter 640 or the AxB converter 642. That is, the encoding mode determiner 630 determines whether there is at least one residual signal that is larger than an impulse component of a predetermined magnitude among all residual signals in the MxN residual block transmitted from the subtractor 620. When there is at least one residual signal of an impulse component, an AxB residual block including a residual signal of at least one impulse component is configured, and the AxB residual block is transmitted to the AxB converter 642. Here, the encoding mode determiner 630 determines whether there is a residual signal of the impulse component, and determines an absolute value of all residual signals in the MxN residual block (a residual value of the impulse component having a predetermined size). The residual signal of the impulse component in the residual block by checking whether the absolute value of the at least one residual signal is equal to or greater than the preset value and whether the residual signal equal to or greater than the preset value is included in the residual block. You can check whether there is.

As shown in FIG. 7, the AxB residual block may be configured as a rectangular block including all residual signals of an impulse component in the MxN residual block. In the AxB residual block, A and B each represent the number of rows and columns of the rectangular block. Referring to FIG. 7, when the residual signal of the impulse component is present at both ends of the bottom row of the MxN residual block, the size of the AxB residual block is 1 × 4. In this way, if the rectangular block is configured to include the residual signals of all impulse components, as shown in 7B, 7C, and 7D, the sizes of the AxB residual blocks are 4x3, 2x2, and 3x4, respectively.

The residual signal in the AxB residual block configured as described above is generated as a bitstream through AxB encoding, which is transformed, quantized, scanned and encoded in units of AxB blocks, and the remaining residual signals not included in the AxB residual block among the residual signals in the MxN residual block are It is assumed that the signal is negligibly smaller than the residual signal of the impulse component, and it is encoded by setting to an arbitrary value, and the arbitrary value may be '0'. Referring to FIG. 8, as shown in 8A, an MxN residual block including residual signals x1 to x16 includes an AxB residual block configured to have a size of 3x3 to include all residual signals x6, x9, and x15 of impulse components hatched. x5 to x7, x9 to x11, x13 to x15 are transformed, quantized and scanned and encoded in 3x3 block units, and the remaining residual signals x1 to x4, x8, x12 and x16 are set to an arbitrary value such as '0' and encoded. Can be. However, '0' is merely an example of an arbitrary value, and an arbitrary value does not necessarily have a value of '0', and may be set to various values such as '-1' and '1'. .

In other words, it is assumed that the residual signal except the AxB residual block in the MxN residual block is significantly smaller than the residual signal of the impulse component, so that the corresponding block, macroblock, slice, picture, etc. may not be used as it is. When a video decoding apparatus generates and restores an MxN residual block by adding it to an AxB residual block, the image decoding apparatus may find and determine a value that may have optimal performance, and determine a residual signal having the value as an alternative residual signal. May be encoded or may be included in the bitstream as replacement residual signal information as it is. When reconstructing the MxN residual block, the image decoding apparatus may extract the replacement residual signal information from the bitstream and use the residual signal except for the AxB residual block as a value of the replacement residual signal identified by the replacement residual signal information. As described above, the value of the replacement residual signal may be determined by finding a value that can achieve optimal performance in the image encoding apparatus, but may be set to a value such as '0', '-1', or '1'. In this case, the replacement residual signal information may not be generated or included in the bitstream, and the image decoding apparatus may restore the MxN residual block by using the preset value as the remaining residual signal.

If a predetermined value (a) is appropriately set to determine whether the residual signal is an impulse component, a residual signal (b) smaller than the predetermined value (a) is -a <b <a (where a is a positive number). It exists in the same range as, and if the transformation, quantization, inverse quantization, inverse transformation, etc. process is added, the quantization error may be added to be a value c other than the value b of the original residual signal. For example, when b is -2, if b is transformed, quantized, inverse quantized, and inversely transformed, a quantization error may be added to restore the value of c equal to -3, and further, the sign is changed to +1. The value may be restored.

As in this example, a residual signal with a small absolute value is restored through transformation and quantization, inverse quantization and inverse transformation, and its value is around '0' and its sign is '+' to '-' or '-' to '+'. Since it can be changed to ', if the preset value (a) is properly set to determine the residual signal of the impulse component, the remaining residual signals not included in the AxB residual block are driven to the value of' 0 'on average. There is a tendency. Therefore, the predetermined value (a) is appropriately set to determine the residual signal of the impulse component, and the values of the remaining residual signals not included in the AxB residual block are set to '0', '-1', '1', etc. By setting it to a predetermined random value, no loss occurs even if the remaining residual signals are not encoded separately or if only the value of the residual residual signal is found for optimal performance. Therefore, only the residual residual signal information may be transmitted by encoding or encoding the residual residual signals to the image decoding apparatus or by encoding or encoding one alternative residual signal that can replace the remaining residual signals. When the apparatus decodes the encoded bitstream in this manner, the residual residual signal is set to a predetermined value such as '0', '-1', or '1' for the remaining residual signals or is identified by the residual residual signal information. By setting the value to, all original residual signals can be restored.

In addition, when there is a residual signal of an impulse component in the residual block, the encoding mode determiner 630 checks whether the size of the AxB residual block is smaller than the size of the MxN residual block, and thus, the size of the AxB residual block is the MxN residual block. The AxB residual block may be transferred to the AxB converter 642 only when the size is smaller than. That is, an AxB residual block including all residual signals of the impulse component is configured. When the size of the AxB residual block is the same as the size of the MxN residual block, since the residual signal of the impulse component exists throughout the MxN residual block, The conventional method, that is, the MxN residual block is encoded.

In addition, when there is no residual signal of an impulse component among all residual signals in the MxN residual block, the encoding mode determiner 630 transmits the MxN residual block to the MxN converter 640 to encode the block in the same size as the current block. In this case, all residual signals of the MxN residual block may be set to '0', which is an arbitrary value. If there is no residual signal of the impulse component in the residual block, as described above, if all residual signals are recovered after transforming and quantizing and inverse quantizing and inverse transforming, the value is set to '0' because the value is driven around '0'. This is because there may not be a large loss even when encoding.

In addition, when the encoding mode determiner 630 determines to encode the AxB residual block, the encoding mode determiner 630 includes at least one of block shape information of the AxB residual block, block position information of the AxB residual block, and encoding identification information for identifying AxB encoding. Mode information may be generated and transmitted to the encoder 670. Here, AxB encoding refers to an AxB residual block including a residual signal of an impulse component, and predicting, transforming, quantizing, scanning, and encoding the AxB residual block in units of AxB blocks. The encoding mode determiner 630 transmits the block shape information of the AxB residual block and the block position information of the AxB residual block to the AxB converter 642.

For example, the block shape information of the AxB residual block may be information in which block sizes such as 3x4 and 2x3 are specified, and the block position information of the AxB residual block is the position where the residual signal of the impulse component first appears in the MxN residual block. Information such as the vertical component value and the horizontal component value of the coordinates for the AxB residual block or coordinates indicating where the position of the topmost residual signal in the leftmost position of the AxB residual block is located in the MxN residual block may be used. Encoding identification information indicating that encoding is performed in units is a flag having a value of '1' or '0', and if the flag is '1', it means that the encoding is performed in AxB block units, and in case of '0', it is MxN block unit. It may indicate that the encoding.

When the AxB converter 642 receives the block shape information of the AxB residual block, the block position information of the AxB residual block, and the AxB residual block from the encoding mode determiner 630, the AxB transform unit 642 performs frequency conversion in units of AxB blocks to obtain a frequency coefficient. Generates an AxB residual block. The AxB converter 642 may perform the conversion in various conversion schemes. For example, one-dimensional transformation is performed on each residual signal of each row and each column of the AxB residual block.

Referring to FIG. 9, the transformation may be performed by multiplying a matrix such as Y = AXB . In FIG. 9, matrix X is a matrix representing AxB residual blocks having a 3x2 block size, and matrix A is a one-dimensional horizontal basis vector {(a1, a2, a3), (b1, b2, b3), (c1, c2, c3)}, and the matrix B represents a matrix including the one-dimensional vertical basis vectors {(d1, d2), (e1, e2)} of length B. In the matrix, Y is a product of the A matrix, the X matrix, and the B matrix, and represents transform coefficients converted in the vertical direction and the horizontal direction. Since the remaining residual signals having a value smaller than the residual signal of the impulse component are assumed to be small enough to be ignored, it is assumed to be '0'.

The AxB quantizer 652 quantizes the AxB residual block transformed by the AxB transform unit 642 to generate a residual block having quantization frequency coefficients. To this end, the AxB quantization unit 652 may quantize using a quantization step (Q Step: Quantization Step) given according to a quantization parameter (QP).

The scan unit 660 generates a quantization frequency coefficient sequence by using a conventional scan method such as a zigzag scan of the quantization frequency coefficients of the MxN residual block quantized by the MxN quantization unit 650 or the AxB quantization unit 652. The quantized frequency coefficient sequence of the quantized frequency coefficients of the AxB residual block is generated by using a different scanning scheme according to the AxB block type.

For example, when the flag value is '0' and the encoding identification information indicates MxN encoding, the scan unit 660 scans the quantization frequency coefficient of the MxN residual block quantized by the MxN quantization unit 650 in a normal zigzag scan. In the case of scanning according to the scheme, and the flag value is '1', and the encoding identification information indicates AxB encoding, the zigzag scan scheme of the quantized frequency coefficient of the AxB residual block quantized by the AxB quantization unit 652 is converted into an AxB block form. Scan according to the modified scan method. The modified scan scheme may be a scheme as exemplarily illustrated in FIG. 10, and may be modified in various scan schemes according to the AxB block type.

The encoder 670 encodes the quantized frequency coefficient sequence scanned by the scan unit 660 using various encoding techniques such as entropy encoding to generate a bitstream. In this case, the encoder 670 may insert mode information transferred from the conversion mode determiner 630 into the bitstream.

The MxN inverse quantizer 680 inverse quantizes the quantized frequency coefficients of the MxN residual block quantized by the MxN quantizer 650 to generate an MxN residual block having frequency coefficients, and the MxN inverse transform unit 690 performs the MxN inverse An inverse transform of the frequency coefficients of the inverse quantized MxN residual block by the quantization unit 680 generates a reconstructed MxN residual block having the reconstructed residual signal.

The AxB inverse quantizer 682 inverse quantizes the quantized frequency coefficients of the AxB residual block quantized by the AxB quantizer 652 to generate an AxB residual block having frequency coefficients, and the AxB inverse transform unit 692 performs AxB inverse. An inverse transform of the frequency coefficients of the inverse quantized AxB residual block by the quantization unit 682 generates a reconstructed AxB residual block having the reconstructed residual signal. Here, inverse AxB quantization and AxB inverse transform may be performed inversely with the aforementioned AxB quantization and AxB transform.

The MxN residual block generator 694 restores the MxN residual block by substituting the AxB residual block transferred from the AxB inverse transform unit 692 into the corresponding position of the MxN residual block. In this case, the MxN residual block generator 694 may substitute the corresponding position using the block position information of the AxB residual block, and have an arbitrary value (for example, '0') for the remaining residual signals not included in the AxB residual block. ) To generate the final MxN residual block.

The adder 696 reconstructs the current block by adding the reconstructed MxN residual block transmitted from the MxN inverse transform unit 690 or the MxN residual block generator 694 with the predicted block predicted by the predictor 610. The reconstructed current block is accumulated in picture units and stored as a reference picture, and may be used to predict the next block.

11 is a flowchart illustrating a video encoding method according to a first embodiment of the present invention.

When the image to be encoded is input, the image encoding apparatus 600 classifies the image into macroblocks or subblock units of the macroblocks and determines an optimal encoding mode among various encoding modes such as an inter prediction mode or an intra prediction mode, and then determines the encoding. Predict and encode a current block to be encoded according to a mode. In this case, when intra mode is determined as an encoding mode and intra prediction is performed, the image encoding apparatus 600 predicts a current block to be encoded to generate a prediction block, and subtracts the prediction block from the current block to generate a residual block. In operation S1110, if the size of the current block is MxN size, an MxN residual block having an MxN size is generated.

The image encoding apparatus 600 analyzes the residual signal in the MxN residual block to determine whether there is a residual signal of an impulse component (S1120). As a result of checking in step S1120, the image encoding apparatus 600 configures an AxB residual block including the residual signal of the impulse component, and the size of the AxB residual block is smaller than the size of the MxN residual block. In operation S1130, the AxB residual block is encoded only in a small case to generate a bitstream (S1140). A method of encoding the AxB residual block has been described above with reference to FIGS. 6 to 10.

On the other hand, if there is no residual signal of the impulse component in the residual block as a result of checking in step S1120, the image encoding apparatus 600 sets all residual signals of the MxN residual block to '0' (S1150), and encodes the MxN residual block. In operation S1160, a bitstream is generated. Also, as a result of checking in step S1130, the image encoding apparatus 600 proceeds to step S1160 when the residual signal of the impulse component is present in the residual block, but the size of the AxB residual block is not smaller than the size of the MxN residual block. Encode In this case, the MxN residual block to be encoded is a residual block not passed through step S1150, that is, an MxN residual block generated in step S1110.

When encoding in this manner, only the residual signal of the small residual block including the residual signal of the impulse component is encoded, and the remaining residual signals are set to predetermined values such as '0', '-1', and '1'. Since only one replacement residual signal having the best performance is encoded without being encoded or representing the remaining residual signals, the amount of bit streams generated by encoding can be reduced, thereby improving compression efficiency.

* The image encoded in the bitstream by the video encoding apparatus 600 is real-time or non-real-time through the wired or wireless communication network, such as the Internet, local area wireless communication network, wireless LAN network, WiBro network, mobile communication network or the like, cable, universal serial bus (USB It may be transmitted to an image decoding apparatus to be described later through a communication interface such as a universal serial bus, decoded by the image decoding apparatus, and restored and reproduced as an image.

12 is a block diagram schematically illustrating a configuration of an image decoding apparatus according to a first embodiment of the present invention.

The image decoding apparatus 1200 according to the first embodiment of the present invention includes a decoder 1210, an inverse scan unit 1220, an MxN inverse quantizer 1230, an AxB inverse quantizer 1230, and an MxN inverse transform unit ( 1240, an AxB inverse transformer 1242, an MxN residual block generator 1250, a predictor 1260, and an adder 1270.

The video encoding apparatus 1200 may be a personal computer (PC), a notebook computer, a personal digital assistant (PDA), a portable multimedia player (PMP), or a PlayStation Portable (PSP). ), A communication device such as a communication modem for communicating with various devices or a wired / wireless communication network, a memory for storing various programs and data for decoding an image, and executing a program. Means a variety of devices including a microprocessor for operation and control.

The decoder 1210 decodes the bitstream, extracts the quantized frequency coefficient sequence, and delivers the quantized frequency coefficient string to the inverse scan unit 1220. In this case, the decoder 1210 may extract mode information from the bitstream. The mode information may be the same information as the mode information described above with reference to FIG. 6, and is transmitted to the inverse scan unit 1220.

The inverse scan unit 1220 inversely scans the quantized frequency coefficient sequence transmitted from the decoder 1210, and inversely scans the quantized frequency coefficient sequence using a conventional zigzag scan method according to the mode information transmitted from the decoder 1210. Scanning may generate an MxN residual block having quantization frequency coefficients, or an AxB residual block having quantization frequency coefficients may be generated by reverse scanning the quantization frequency coefficient sequence using a scan method modified according to the AxB block shape. For example, when the flag indicating the encoding identification information included in the mode information is '0', the quantization frequency coefficient sequence is reversely scanned according to the existing zigzag scan method, and the flag indicating the encoding identification information is '1'. In this case, the quantization frequency coefficient sequence may be reverse-scanned by a different zigzag scan method according to the shape of the block identified by the AxB block type information included in the mode information. The reverse scanning according to the zigzag scan method according to the type of the block identified by the AxB block type information may be performed by scanning in the reverse order of the scan method described above with reference to FIG. 10. As such, when each residual block is generated by performing different inverse scanning according to the mode information, the inverse scan unit 1220 transfers the generated residual block to the MxN inverse quantization unit 1230 when the generated residual block is an MxN residual block, and AxB. In the case of the residual block, it is transmitted to the AxB inverse quantization unit 1232.

The MxN inverse quantizer 1230, the MxN inverse transform unit 1240, the AxB inverse quantizer 1232, and the AxB inverse transform unit 1242 are the MxN inverse quantizer 680 and the MxN inverse transform unit described above with reference to FIG. 6. 690 and the AxB inverse quantizer 682 and the AxB inverse transform unit 692 perform the same or similar functions, and thus detailed descriptions thereof will be omitted.

The MxN residual block generator 1250 restores the MxN residual block by inserting the AxB residual block transferred from the AxB inverse transform unit 1242 into a corresponding position of the MxN residual block. In this case, the MxN residual block generator 1250 may substitute the corresponding position by using the block position information of the AxB residual block extracted from the bitstream by the decoder 1210 and the remaining residual not included in the AxB residual block. If the signal is assigned a predetermined value (for example, '0', '-1', '1', etc.) or extracted the residual residual signal information from the bitstream, the residual residual signal identified by the alternate residual signal information By setting to the value of, the final MxN residual block is generated.

The predictor 1260 predicts the current block to generate a predicted block, and the adder 1270 predicts the reconstructed MxN residual block transmitted from the MxN inverse transformer 1240 or the MxN residual block generator 1250. The current block is reconstructed by adding the prediction block predicted by the block 1260. The reconstructed current block is accumulated in picture units and output as a reconstructed picture or stored as a reference picture, and may be used to predict the next block.

13 is a flowchart illustrating a method of decoding an image according to the first embodiment of the present invention.

The image decoding apparatus 1200, which receives and stores a bitstream of an image through a wired or wireless communication network or a cable, decodes and restores the image in order to reproduce the image according to a user's selection or an algorithm of another program being executed.

To this end, the image decoding apparatus 1200 extracts a residual block by decoding the bitstream (S1310). In this case, the image decoding apparatus 1200 extracts mode information from the bitstream.

The image decoding apparatus 1200 checks the mode information to determine whether the encoding information includes the encoding identification information indicating the AxB encoding (S1320). If the encoding information includes the encoding identification information indicating the AxB encoding, the AxB residual is determined. By inversely scanning the block according to the block shape information included in the mode information, and inversely transforming by inverse quantization in AxB block units, the AxB residual block is restored (S1330), and the AxB residual restored according to the block position information included in the mode information. The MxN residual block is restored by assigning the block to the MxN residual block and assigning the remaining residual signal to a predetermined value or a value of the replacement residual signal identified by the replacement residual signal information extracted from the bitstream (S1340). On the other hand, the image decoding apparatus 1200 restores by inversely scanning, inversely quantizing, and inversely transforming the MxN residual block according to a conventional scheme when the determination result of step S1320 does not include encoding identification information indicating AxB encoding. (S1350).

The image decoding apparatus 1200 predicts the current block, generates a prediction block, adds the MxN residual block reconstructed in operation S1340 or S1350, and restores the current block in operation S1360. As such, the reconstructed current block is stored in picture units and output as a reconstructed picture or stored as a reference picture to be used to predict the next block.

14 is a block diagram schematically illustrating a configuration of an image encoding apparatus according to a second embodiment of the present invention.

The image encoding apparatus 1400 according to the second embodiment of the present invention is an apparatus for encoding an image, and includes all components of the image encoding apparatus 600 according to the first embodiment of the present invention described above with reference to FIG. 6. In addition, the apparatus further includes a ratio calculator 1410. The video encoding apparatus 1400 according to the second embodiment of the present invention is the same as the video encoding apparatus 600 according to the first embodiment of the present invention described above with reference to FIG. 6, and includes an AxB residual including a residual signal of an impulse component. AxB encoding is performed for encoding a block. However, it is determined whether AxB encoding is to be performed according to an impulse ratio, which is a ratio of residual signals of impulse components among all residual signals in the AxB residual block. Hereinafter, detailed descriptions of the remaining components described above with reference to FIG. 6 will be omitted.

When the AxB encoding is determined by the encoding mode determiner 620 to receive the AxB residual block, the ratio calculation unit 1410 receives an impulse ratio (AxB residual), which is a ratio of the residual signal of the impulse component among all residual signals in the AxB residual block. AxB residual block and mode information only when the number of residual signals of the impulse component in the block / the total number of residual signals in the AxB residual block) is greater than or equal to the threshold by comparing the calculated ratio with a preset threshold value. If it is less than the threshold value, the MxN residual block may be transferred to the MxN transform unit 640 or the encoding mode determiner 620 may transmit the MxN residual block so that the MxN residual block may be encoded as in the conventional method. Control to transfer to the MxN converter 640. Accordingly, it is possible to indicate that AxB encoding is performed by setting a flag indicating encoding identification information to '1' only when the impulse ratio is equal to or greater than a threshold value, and the encoding identification information, block shape information, and block position information are converted into AxB conversion units 642. If the impulse ratio is less than the threshold value, a flag indicating encoding identification information may be set to '0'.

15 is a flowchart for explaining a video encoding method according to a second embodiment of the present invention.

In the image encoding method according to the second embodiment of the present invention, steps S1510 to S1530 are the same as or similar to steps S1110 to S1130 described above with reference to FIG. 11, and steps S1550 to S1570 are described above with reference to FIG. 11. To the same or similar to step S1160. However, according to the image encoding method according to the second embodiment of the present invention, when the size of the AxB residual block is smaller than the size of the MxN residual block in step S1530, the image encoding apparatus 1400 calculates an impulse ratio and preset threshold value. In comparison with S1540, the process proceeds to step S1550 only when the impulse ratio is greater than or equal to the threshold value to encode the AxB residual block, and generates a bitstream if the impulse ratio is less than the threshold, and proceeds to step S1570 to encode the MxN residual block. To generate the bitstream.

The image decoding apparatus according to the second embodiment of the present invention is configured in the same manner as the image decoding apparatus 1200 according to the first embodiment of the present invention described above with reference to FIG. 12 and generates a bitstream by encoding an image. 11 is the same as the image decoding method according to the first embodiment of the present invention described above with reference to FIG. 11. This is because the video encoding apparatus 1400 may change the encoding scheme according to the impulse ratio and the decoding methods are the same.

16 is a block diagram schematically illustrating a configuration of an image encoding apparatus according to a third embodiment of the present invention.

The image encoding apparatus 1600 according to the third embodiment of the present invention includes all components of the image encoding apparatus 600 according to the first embodiment of the present invention described above with reference to FIG. 6. However, the functions of the encoding mode determiner 1610 and the MxN residual block generator 1620 are partially different from those of the encoding mode determiner 630 and the MxN residual block generator 694, respectively.

That is, the encoding mode determiner 1610 of the image encoding apparatus 1600 according to the third exemplary embodiment of the present invention configures an AxB residual block including a residual signal of an impulse component, in a diagonal direction as well as a rectangular block form. An AxB residual block in the form of a diagonal block may be configured to include a residual signal of a located impulse component.

To this end, when configuring the AxB residual block, the encoding mode determiner 1610 determines whether there is a residual signal of the impulse component in the diagonal direction in the MxN residual block, and when there is no residual signal of the impulse component in the diagonal direction. The AxB residual block is configured in the above-described rectangular block form, and when there is a residual signal of the impulse component in the diagonal direction, the AxB residual block is configured in the diagonal block form.

Referring to FIG. 17, assuming that an MxN residual block is a 4x4 block and three residual signals are diagonally residual signals of an impulse component, as shown in 17A, an AxB residual block is configured according to the first embodiment of the present invention. In this case, the AxB residual block becomes a 3x3 block, and six residual signals must be encoded. On the other hand, when configuring the AxB residual block according to the third embodiment of the present invention as shown in 17C, the AxB residual block may be a block of size 1x3, and only three residual signals may be encoded.

Similarly, assuming that the MxN residual block is a 4x4 block and two pairs of three residual signals diagonally are residual signals of an impulse component, as shown in 17B, when configuring an AxB residual block according to the first embodiment of the present invention, In this case, the AxB residual block becomes a 4x4 sized block so that the size of the AxB residual block is the same as that of the MxN residual block, and 16 residual signals must be encoded. On the other hand, when configuring the AxB residual block according to the third embodiment of the present invention as shown in 17D, the AxB residual block becomes a 2x3 sized block so that the size of the AxB residual block becomes smaller than the size of the MxN residual block. Only residual signals may be encoded.

Accordingly, as can be seen from FIG. 17, when the AxB residual block is configured according to the third embodiment of the present invention, the size of the AxB residual block can be reduced, and the number of residual signals to be encoded can be reduced accordingly. Since the size of the AxB residual block becomes the same as the size of the MxN residual block, and when implemented in combination with the second embodiment of the present invention, the impulse ratio can be increased, so that the ratio capable of performing AxB encoding is increased. As a result, the compression efficiency can be further improved.

In addition, the encoding mode determiner 1610 may include diagonal block identification information indicating that the AxB residual block is formed in the form of a diagonal block, and diagonal position information indicating the position of the residual signal of each impulse component constituting the AxB residual block in the form of a diagonal block. It is additionally generated as mode information, and the information is transmitted to the MxN residual block generator 1620.

The MxN residual block generator 1620 generates an MxN residual block when the AxB residual block reconstructed from the AxB inverse quantizer 692 is delivered. The MxN residual block generator 1620 reconstructs the diagonal block identification information transmitted from the encoding mode determiner 1610. Identifies that the AxB residual block is a residual block in the form of a diagonal block, and assigns each residual signal of the restored AxB residual block to the MxN residual block by dividing the diagonal position information into a diagonal position, and the remaining residual signals are predetermined random signals. (Eg, '0', '-1', '1', etc.) or the last reconstructed MxN residual by setting to the value of the replacement residual signal identified by the replacement residual signal information extracted from the bitstream. Create a block.

Although the configuration of the video decoding apparatus according to the third embodiment of the present invention is not shown, the configuration of the video decoding apparatus 1200 according to the first embodiment described above with reference to FIG. 12 is the same. However, the mode information extracted from the bitstream by the decoder 1210 further includes diagonal block identification information and diagonal position information, and the information is transmitted to the MxN residual block generator 1250. The MxN residual block generator 1250 according to the third embodiment of the present invention uses a diagonal block form using diagonal block identification information and diagonal position information, like the MxN residual block generator 1620 described above with reference to FIG. 16. Substitute the AxB residual block in the MxN residual block and assign the remaining residual signal to an arbitrary value or a value of the replacement residual signal identified by the replacement residual signal information extracted from the bitstream to generate a reconstructed MxN residual block.

Meanwhile, in the above-described first to third embodiments of the present invention, the

image encoding apparatuses

600, 1400, and 1600 may predetermine whether all of the quantization frequency coefficients generated by MxN residual blocks are transformed and quantized are '0'. If it is determined that all are '0', MxN encoding may be performed. If at least one is not '0', AxB encoding may be performed. That is, a coefficient discrimination unit is further configured between the scan unit 660, the MxN quantization unit 650, and the AxB quantization unit 652 to quantize the quantized residual block output from the MxN quantization unit 650. When all the quantized frequency coefficients are '0' by checking the frequency coefficient, the quantized residual block output from the MxN quantizer 650 is transmitted to the scan unit 660, and mode information such as encoding identification information indicating MxN encoding is transmitted. By not generating or including in the bitstream, the existing MxN encoding may be performed, and the quantized residual block output from the MxN quantization unit 650 only when at least one or more of the quantization frequency coefficients is not '0', or The quantized residual block output from the AxB quantization unit 652 is transmitted to the scan unit 660, and MxN encoding is performed on AxB encoding or mode information including encoding identification information. MxN encoding or AxB encoding may be performed by including mode information including encoding identification information, block type information, and block position information indicating in the bitstream. Accordingly, in the first to third embodiments, the

image encoding apparatuses

600, 1400, and 1600 may convert the mode information into the bitstream when the quantization frequency coefficients generated by MxN residual blocks are transformed and quantized are all '0'. Without including, mode information may be included in the bitstream only when at least one of the quantization frequency coefficients is not '0'.

Similarly, the above-described image decoding apparatuses 1200 according to the first to third embodiments of the present invention have all the quantization frequency coefficients reconstructed by decoding the bitstream being '0' (or encoding included in the bitstream). Coded block pattern (CBP) can be used to recover the current block by restoring the MxN residual block by inverse quantization and inverse transformation of quantization frequency coefficients without reading mode information from the bitstream. If at least one or more of the quantized frequency coefficients reconstructed by decoding is not '0' (or can be known through a coded block pattern (CBP) included in the bitstream), mode information may be obtained from the bitstream. It reads and judges whether it is MxN encoding or AxB encoding based on encoding identification information, and MxN residual block in case of MxN encoding. The indent restore and read the additional block position information and block form information from the mode information when the restore the current block, AxB encoding can reconstruct the current block by restoring the residual block AxB.

18 is a block diagram schematically illustrating a configuration of a video encoding apparatus according to a fourth embodiment of the present invention.

The image encoding apparatus 1800 according to the fourth embodiment of the present invention includes all components of the image encoding apparatus 600 according to the first embodiment of the present invention described above with reference to FIG. 1820 is further included, and some functions of the encoding mode determiner 1810 are changed. Hereinafter, detailed descriptions of the remaining components described above with reference to FIG. 6 will be omitted.

The video encoding apparatus 1800 according to the fourth embodiment of the present invention is the same as the video encoding apparatus 600 according to the first embodiment of the present invention described above with reference to FIG. 6, and includes an AxB residual including a residual signal of an impulse component. AxB encoding for encoding a block is performed. However, by encoding the MxN residual block according to the conventional method, it is determined whether all the quantization frequency coefficient values of the quantized residual block are '0', and when all are '0', the MxN residual is predicted, transformed, and quantized by the conventional method. Generates a bitstream by scanning and encoding a block by using a conventional zigzag scan method.If one or more is not '0', the bitstream is predicted, transformed, and quantized in AxB block units, and the scan method is changed according to the shape of the AxB residual block. AxB encoding that generates a bitstream by scanning and encoding is performed. In addition, when AxB encoding is performed, it is determined whether the number of pixels of the AxB residual block is equal to or less than a preset number, and when AxB is smaller than or equal to the preset number, AxB encoding is performed. Encode the residual block. For example, when the preset value is 8, the residual block multiplied by A and B is 8 or less performs AxB encoding, and the residual block larger than 8 performs MxN encoding. In the case of residual blocks such as 1x1, 2x3, and 4x2, AxB encoding is performed because the number of pixels of the residual block is 8 or less.

To this end, the encoding mode determiner 1810 not only determines whether to perform AxB encoding as in the encoding mode determiner 630 described above with reference to FIG. 6, but also, when performing AxB encoding, the number of pixels of the AxB residual block. Is determined to be less than or equal to the preset number, and when less than or equal to the preset number, the AxB residual block is transferred to the AxB transform unit 642 to perform AxB encoding. The MxN residual block is transmitted to the MxN converter 640 to encode the block.

The coefficient discrimination unit 1820 determines whether all quantization frequency coefficients of the quantized residual block transmitted from the MxN quantization unit 650 are '0'. And transmitting the quantized MxN residual block to the scan unit 660 to scan in a conventional zigzag scan method without including mode information including encoding identification information indicating MxN encoding in the bitstream, and at least one of '0'. If not, it is determined whether the number of pixels of the AxB residual block is less than or equal to the preset number, and if only the following is included, mode information including encoding identification information indicating the AxB encoding, block shape information, and block position information is included in the bitstream. The quantized AxB residual block transmitted from the AxB quantization unit 652 is transferred to the scan unit 660, and according to the block shape of the AxB residual block. Otherwise the method will be to scan. When the number of pixels of the AxB residual block exceeds the preset number, the mode information including the encoding identification information indicating the MxN encoding is encoded in the conventional manner without being included in the bitstream.

19 is a flowchart illustrating a video encoding method according to a fourth embodiment of the present invention.

The image encoding apparatus 1800 determines whether the quantization frequency coefficients of the quantized MxN residual blocks are all '0' according to an existing encoding scheme encoded in MxN block units (S1910). If the result of the determination in step S1910 is '0', the image encoding apparatus 1800 generates a bitstream by scanning and encoding the quantized frequency coefficients encoded by the MxN block unit using the existing zigzag scan method or the like ( S1920). If at least one of the determination results in step S1910 is not '0', the image encoding apparatus 1800 determines whether the number of pixels of the AxB residual block is less than or equal to the preset number (S1930). As a result of checking in step S1930, if the preset number is exceeded, the process proceeds to S1920 and the method is encoded according to the existing method. If the number is less than the preset number, the bitstream is generated by performing AxB encoding that encodes the AxB residual block in units of AxB blocks. S1940).

According to the fourth embodiment of the present invention, it is determined in advance whether the quantization frequency coefficients of the quantized residual blocks of the MxN residual block are all '0' and whether the number of pixels of the AxB residual block is equal to or less than the preset number. By determining, the number of pixels of the AxB residual block is not set to at least one of the quantization frequency coefficients and the preset number of pixels of the AxB residual block is not included in every residual block including mode information including encoding identification information indicating MxN encoding or AxB encoding. Only in the following cases, encoding identification information, block type information, and block position information indicating AxB encoding may be included in the bitstream. Therefore, the bit amount for mode information including encoding identification information, block shape information, and block position information can be reduced according to the characteristics of the block.

In addition, according to the fourth embodiment of the present invention, since AxB encoding is performed only when the number of pixels of the AxB residual block is less than or equal to the preset number, the number of pixels of the AxB residual block is reduced by reducing the number of pixels of the AxB residual block. Can be reduced.

The image decoding apparatus according to the fourth embodiment of the present invention includes all components of the image decoding apparatus 1200 according to the first embodiment of the present invention described above with reference to FIG. 12. However, when obtaining the quantization frequency coefficient sequence from the bitstream, the decoder 1210 checks whether the number of quantization frequency coefficients is less than or equal to a preset number, and if it is less than or equals, extracts mode information from the bitstream and confirms encoding identification information, and then AxB. In the case of encoding, block type information and block position information are extracted from the mode information. That is, the first embodiment differs from the first embodiment in that the number of quantization frequency coefficients of the quantization frequency coefficient sequence is first checked in order to decode in units of AxB blocks only when the number of pixels of the AxB residual block is equal to or less than a preset number.

For example, the image decoding apparatus according to the fourth exemplary embodiment of the present invention checks the number of quantization frequency coefficients of a corresponding block through total coefficients and total zeros of the quantization frequency coefficient sequences, and then sets 'N'. If there are more than ',' the residual block is determined to be MxN coded (the number of frequencies cannot be more than 'N' in AxB coding), and if 'N' or less, 'N' is also used through MxN coding. Since the following frequency coefficients can be created (for example, when there is only one frequency coefficient of a DC component or when a frequency coefficient of a low frequency component near a DC component occurs), additionally, mode information is extracted from the bitstream and mode information. It is determined whether the AxB encoding is performed based on encoding identification information included in the. Accordingly, mode information is not included in the bitstream when AxB encoding is not performed, and mode information including encoding identification information, block shape information, and block position information is included in the bitstream only when AxB encoding is used.

The image encoding apparatus according to the fifth embodiment of the present invention includes all components of the image encoding apparatus 1800 according to the fourth embodiment of the present invention described above with reference to FIG. 18, and includes a coefficient discrimination unit 1820. Some functions of are changed. Hereinafter, detailed descriptions of the remaining components described above with reference to FIG. 18 will be omitted.

An image encoding apparatus according to a fifth embodiment of the present invention encodes an AxB residual block including a residual signal of an impulse component, like the image encoding apparatus 1800 according to the fourth embodiment of the present invention described above with reference to FIG. 18. AxB encoding is performed. However, by encoding the MxN residual block according to the conventional method, it is determined whether all the quantization frequency coefficient values of the quantized residual block are '0', and when all are '0', the MxN residual is predicted, transformed, and quantized by the conventional method. Generates a bitstream by scanning and encoding a block by using a conventional zigzag scan method.If one or more is not '0', the bitstream is predicted, transformed, and quantized in AxB block units, and the scan method is changed according to the shape of the AxB residual block. AxB encoding that generates a bitstream by scanning and encoding is performed. In addition, the image encoding apparatus according to the fifth embodiment of the present invention determines whether the number of pixels of the AxB residual block exceeds a preset number when at least one or more of the quantization frequency coefficients of the quantized residual block is not '0'. If it is determined, the AxB encoding is performed when the preset number is exceeded, and when the number is less than the preset number, the MxN residual block is encoded by the conventional method. For example, when the preset value is 8, the residual block multiplied by A and B is 8 or less performs MxN encoding, and the residual block larger than 8 performs AxB encoding. That is, in the case of residual blocks such as 3x3, 3x4, and 4x3, AxB encoding is performed because the number of pixels of the residual block exceeds eight.

The coefficient discrimination unit 1820 determines whether all quantization frequency coefficients of the quantized residual block transmitted from the MxN quantization unit 650 are '0'. And mode information including encoding identification information indicating MxN encoding is included in the bitstream while the quantized MxN residual block is transmitted to the scan unit 660, and scanned in a conventional zigzag scan method, and at least one of the quantization frequency coefficients. If the value is not '0', it is determined whether the number of pixels of the AxB residual block exceeds a preset number, and mode information including encoding identification information indicating the AxB encoding, block shape information, and block position information only when the number of pixels in the AxB residual block is exceeded. Is included in the bitstream and the AxB quantized AxB residual block delivered from the AxB quantizer 652 is transferred to the scan unit 660 while According to the block type of the secondary blocks to be scanned by different scanning method. If the number of pixels of the AxB residual block is less than or equal to the preset number, the scan unit 660 includes mode information including encoding identification information indicating MxN encoding in the bitstream and predicts, transforms, and quantizes the MxN residual block in a conventional manner. Encode it by passing it.

The image encoding apparatus according to the fifth embodiment of the present invention determines whether the quantization frequency coefficients of the quantized MxN residual blocks are all '0' according to an existing encoding scheme encoded in MxN block units (S2010). If the result of the determination in step S2010 is '0', the image encoding apparatus generates a bitstream by scanning and encoding the quantization frequency coefficients encoded in MxN block units using an existing zigzag scan method or the like (S2020). If at least one of the determination results in step S2010 is not '0', the image encoding apparatus checks whether the number of pixels of the AxB residual block exceeds a preset number (S2030). If the result of the check in step S2030 is less than or equal to the predetermined number, the process proceeds to S2020 and the existing method is encoded. If the preset number is exceeded, the bitstream is generated by performing AxB encoding that encodes the AxB residual block in units of AxB blocks. (S2040).

According to the fifth embodiment of the present invention, AxB encoding is performed only when the number of pixels of the AxB residual block exceeds a preset number. Thus, compared with the fourth embodiment, the AxB residual block increases the number of pixels in the AxB residual block. Although the number of bits of the data encoding the block may be increased, encoding identification information for indicating MxN encoding or AxB encoding should be included in the bitstream for every block, while encoding block shape information and block position information of the AxB residual block. It can further reduce the amount of bits required. For example, when the preset number is '8', the AxB residual block may be configured as one of 3x3, 4x3, and 3x4, and thus the shape of the AxB residual block is also determined as one of the smaller cases. Since the position is also determined by one of the fewer cases, the block type information and the block position information can be represented with fewer bits.

The image decoding apparatus according to the fifth embodiment of the present invention is configured in the same manner as the image decoding apparatus 1200 according to the first embodiment of the present invention described above with reference to FIG. 12. Accordingly, the image decoding method according to the fifth embodiment of the present invention is also the same as the other image decoding method of the first embodiment of the present invention described above with reference to FIG. 13.

Here, in the fifth embodiment of the present invention, unlike the fourth embodiment, the reason why the image decoding apparatus does not first determine the number of quantization frequency coefficients is that the total coefficient of the quantization frequency coefficient sequence and the total zero are calculated. This is because it is not possible to determine that the residual block is AxB coded if the number of quantization frequency coefficients of the corresponding block that is identified through the number exceeds 'N'. That is, even when MxN coding is performed, the number of quantization frequency coefficients may exceed N. Therefore, it may not be determined whether the residual block is AxB coded or MxN coded based on the number of quantized frequency coefficients. Accordingly, the video decoding apparatus according to the fifth embodiment of the present invention reads the encoding identification information of the mode information from the bitstream and determines whether it is MxN encoding or AxB encoding, as in the first embodiment, and in case of AxB encoding. Reads block type information and block position information of mode information and performs AxB decoding.

Optionally combining one or more of the first to fifth embodiments of the present invention described above can be implemented as the sixth embodiment to be described later. That is, the image encoding apparatus according to the sixth embodiment of the present invention is an apparatus for encoding an image, the prediction unit generating a prediction block by predicting a current block, and the subtraction of generating a MxN residual block by subtracting the current block and the prediction block. The AxB encoder may be configured to encode an AxB residual block including a residual signal of an impulse component in the MxN residual block to generate a bitstream. Here, the AxB encoder may generate and insert mode information including at least one of block shape information of the AxB residual block, block position information of the AxB residual block, and encoding mode information for identifying AxB encoding, into the bitstream.

In addition, the image encoding apparatus according to the sixth embodiment of the present invention provides an MxN residual block and an MxN residual block according to whether the residual signal of an impulse component is included in the MxN encoder that generates the bitstream by encoding the MxN residual block, and the MxN residual block. The apparatus may further include an encoding mode determiner configured to determine which residual block among the AxB residual blocks.

According to the method of encoding an image according to the sixth embodiment of the present invention, the image encoding apparatus predicts a current block to generate a prediction block, subtracts the current block and the prediction block, generates an MxN residual block, and generates an MxN residual. A bitstream may be generated by encoding the AxB residual block including the residual signal of the impulse component in the block. The apparatus for encoding an image further includes generating and inserting mode information including at least one of block shape information of an AxB residual block, block position information of an AxB residual block, and encoding identification information for identifying AxB encoding into a bitstream. Can be done with

In addition, when the size of the AxB residual block is the same as the size of the MxN residual block, the image encoding apparatus may further perform encoding of the MxN residual block to generate a bitstream, wherein the MxN residual block is a residual signal of an impulse component. If it is not included, the step of encoding by setting all residual signals of the MxN residual block to '0' may be further performed.

Here, in performing AxB encoding, the apparatus for encoding an image may scan and encode the quantized frequency coefficients obtained by transforming the AxB residual block and quantizing the quantized frequency coefficients according to the shape of the AxB residual block, and the size of the AxB residual block may be An AxB residual block can be encoded only when it is smaller than the size of the MxN residual block. In addition, the AxB residual block may be in the form of a rectangular block or a diagonal block. If the AxB residual block is in the form of a diagonal block, the image encoding apparatus may additionally generate and insert one or more of the diagonal block identification information and the diagonal block position information into the bitstream.

Also, the image encoding apparatus may encode the AxB residual block only when the impulse ratio of the AxB residual block is greater than or equal to the threshold. Also, the image encoding apparatus may encode the AxB residual block only when the number of pixels of the AxB residual block is equal to or less than a preset number. In particular, the image encoding apparatus encodes an AxB residual block only when at least one or more of the values of the quantization frequency coefficients generated by transforming and quantizing the MxN order block are not '0' and the number of pixels of the AxB residual block is less than or equal to a preset number. can do. In this case, the image encoding apparatus blocks the block of the AxB residual block only when at least one or more of the values of the quantization frequency coefficients generated by transforming and quantizing the MxN order block are not '0' and the number of pixels of the AxB residual block is less than or equal to a preset number. Mode information including one or more of shape information, block position information of the AxB residual block, and encoding identification information for identifying AxB encoding may be generated and inserted into the bitstream.

In addition, the image encoding apparatus may sign the AxB residual block only when the number of pixels of the AxB residual block exceeds a preset number. In particular, the image encoding apparatus may use the AxB residual block only when at least one or more of the values of the quantization frequency coefficients generated by transforming and quantizing the MxN order block are not '0' and the number of pixels of the AxB residual block exceeds a preset number. Can be encoded. In this case, the image encoding apparatus performs the AxB residual block only when at least one or more of the values of the quantization frequency coefficients generated by transforming and quantizing the MxN order block are not '0' and the number of pixels of the AxB residual block exceeds a preset number. Mode information including at least one of block shape information of a block, block position information of an AxB residual block, and encoding identification information for identifying AxB encoding may be generated and inserted into a bitstream.

An apparatus for decoding an image according to a sixth embodiment of the present invention includes a decoder that extracts a quantized frequency coefficient sequence by decoding a bitstream, and inversely scans, inverse quantizes, and inversely transforms the quantized frequency coefficient sequence in units of AxB blocks, thereby causing AxB residuals. AxB residual block generator for generating blocks, MxN residual block generator for generating MxN residual blocks by additionally combining one or more residual signals to AxB residual blocks, prediction unit for predicting the current block to generate prediction blocks and prediction blocks And an adder for adding the MxN residual block to restore the current block.

According to a method of decoding an image according to a sixth embodiment of the present invention, the image decoding apparatus may include extracting a quantized frequency coefficient sequence by decoding a bitstream, performing an inverse scan of the quantized frequency coefficient sequence in units of AxB blocks, Generating an AxB residual block by inverse quantization and inverse transformation, further combining one or more residual signals to the AxB residual block to generate an MxN residual block, and predicting the current block to generate a prediction block; And reconstructing the current block by adding the prediction block and the MxN residual block.

Here, in generating the AxB residual block, the image decoding apparatus inversely scans and inverse quantizes the quantized frequency coefficient sequence in units of AxB blocks based on the block shape information of the AxB residual block included in the mode information extracted from the bitstream. And an inverse transform, to generate an AxB residual block. In addition, in generating the MxN residual block, the image decoding apparatus may generate the MxN residual block based on the block position information of the AxB residual block included in the mode information extracted from the bitstream. In addition, the one or more residual signals may have a predetermined value, in particular, all may be set to a value of '0'. In addition, the apparatus for decoding an image may generate an MxN residual block ring by using, as one or more residual signals, the residual residual signal identified by the residual residual signal information extracted from the bitstream.

In addition, in generating the AxB residual block, the image decoding apparatus may extract mode information from the bitstream, and AxB only when AxB encoding is identified by encoding identification information included in the mode information extracted from the bitstream. Residual blocks can be created. In generating the AxB residual block, the image decoding apparatus may generate the AxB residual block only when the number of quantization frequency coefficients of the quantization frequency coefficient sequence is equal to or less than a preset number.

In addition, the apparatus for decoding an image includes: a decoder which extracts a quantized frequency coefficient sequence by decoding a bitstream; An AxB residual block generator configured to inversely scan, inverse quantize, and inversely convert the quantized frequency coefficient sequence into AxB block units to generate an AxB residual block; An MxN residual block generator for further combining one or more residual signals to the AxB residual block to generate an MxN residual block; A prediction unit for predicting a current block to generate a prediction block; And an adder for reconstructing the current block by adding the prediction block and the MxN residual block.

In the above description, it is described that all the components constituting the embodiments of the present invention are combined or operated in one, but the present invention is not necessarily limited to these embodiments. In other words, within the scope of the present invention, all of the components may be selectively operated in combination with one or more. In addition, although all of the components may be implemented in one independent hardware, each or all of the components may be selectively combined to perform some or all functions combined in one or a plurality of hardware. It may be implemented as a computer program having a. Codes and code segments constituting the computer program may be easily inferred by those skilled in the art. Such a computer program may be stored in a computer readable storage medium and read and executed by a computer, thereby implementing embodiments of the present invention. The storage medium of the computer program may include a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like.

In addition, the terms "comprise", "comprise" or "having" described above mean that the corresponding component may be inherent unless specifically stated otherwise, and thus excludes other components. It should be construed that it may further include other components instead. All terms, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. Terms used generally, such as terms defined in a dictionary, should be interpreted to coincide with the contextual meaning of the related art, and shall not be interpreted in an ideal or excessively formal sense unless explicitly defined in the present invention.

The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art to which the present invention pertains may make various modifications and changes without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention but to describe the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the equivalent scope should be interpreted as being included in the scope of the present invention.

As described above, the present invention is applied to fields such as an apparatus and a method for encoding or decoding an image, and when an residual signal of an impulse component is generated due to an inaccuracy of prediction, an image is generated by different encoding and decoding methods. It is a very useful invention that generates the effect of improving the encoding efficiency by efficiently encoding or decoding.

CROSS-REFERENCE TO RELATED APPLICATION

This patent application is related to the patent application No. 10-2008-0094167 filed in Korea on September 25, 2008 and the patent application No. 10-2009-0085571 filed on September 10, 2009. If a) claims priority under section 35 USC § 119 (a), all of which is incorporated by reference in this patent application. In addition, if this patent application claims priority for the same reason as above for a country other than the United States, all the contents thereof are incorporated into this patent application by reference.

Claims

In the method of encoding an image,

An MxN residual block generation step of generating a prediction block by predicting a current block and generating an MxN residual block by subtracting the current block and the prediction block; And

An AxB encoding step of generating a bitstream by encoding an AxB residual block including a residual signal of an impulse component in the MxN residual block

Image encoding method comprising a.
The method of claim 1, wherein the video encoding method comprises:

Adding a mode information insertion step of generating mode information including at least one of block shape information of the AxB residual block, block position information of the AxB residual block, and encoding identification information for identifying AxB encoding, and inserting it into the bitstream; The video encoding method comprising a.
The method of claim 1, wherein the AxB encoding step is performed.

And encoding and quantizing the quantized frequency coefficients obtained by transforming and quantizing the AxB residual block according to a shape of the AxB residual block.
The method of claim 1, wherein the AxB encoding step is performed.

And encoding the AxB residual block only when the size of the AxB residual block is smaller than the size of the MxN residual block.
The method of claim 1, wherein the video encoding method comprises:

And encoding the MxN residual block to generate a bitstream when the size of the AxB residual block is the same as the size of the MxN residual block.
The method of claim 1, wherein the video encoding method comprises:

If the MxN residual block does not include the residual signal of the impulse component, the video encoding method further comprises setting an encoding by setting all residual signals of the MxN residual block to '0'.
The method of claim 1, wherein the AxB residual block,

Image coding method characterized in that the rectangular block form or diagonal block form.
The method of claim 7, wherein the video encoding method,

And inserting at least one of diagonal block identification information and diagonal block position information into the bitstream when the AxB residual block is in the form of the diagonal block, and inserting the diagonal block into the bitstream.
The method of claim 1, wherein the AxB encoding step is performed.

And encoding the AxB residual block only when an impulse ratio of the AxB residual block is greater than or equal to a threshold.
The method of claim 1, wherein the AxB encoding step is performed.

And encoding the AxB residual block only when the number of pixels of the AxB residual block is equal to or less than a preset number.
The method of claim 10, wherein the AxB encoding step is performed.

The AxB residual block is encoded only when at least one or more quantization frequency coefficients generated by transforming and quantizing the MxN residual block are not '0' and the number of pixels of the AxB residual block is less than or equal to a preset number. Image coding method.
The method of claim 11, wherein the video encoding method comprises:

Block shape information of the AxB residual block, AxB only when at least one or more of the quantization frequency coefficients generated by transforming and quantizing the MxN residual block is not '0' and the number of pixels of the AxB residual block is equal to or less than a preset number. And generating mode information including at least one of block position information of the residual block and encoding identification information for identifying AxB encoding, and inserting mode information into the bitstream.
The method of claim 1, wherein the AxB encoding step is performed.

And encoding the AxB residual block only when the number of pixels of the AxB residual block exceeds a preset number.
The method of claim 13, wherein the AxB encoding step is performed.

The AxB residual block is encoded only when at least one or more quantization frequency coefficients generated by transforming and quantizing the MxN residual block are not '0' and the number of pixels of the AxB residual block exceeds a preset number. A video encoding method.
The method of claim 14, wherein the video encoding method comprises:

Block shape information of the AxB residual block only when at least one or more of the quantization frequency coefficients generated by transforming and quantizing the MxN residual block is not '0' and the number of pixels of the AxB residual block exceeds a preset number; And encoding mode information for generating mode information including at least one of block position information of the AxB residual block and encoding identification information for identifying AxB encoding, and inserting the mode information into the bitstream. Way.
The method of claim 1, wherein the video encoding method comprises:

And a representative residual signal information inserting step of inserting, in the bitstream, residual residual signal information identifying an alternative residual signal of a residual signal other than the AxB residual block in the MxN residual block. .
In the apparatus for encoding a video,

A prediction unit for predicting a current block to generate a prediction block;

A subtraction unit for generating an MxN residual block by subtracting the current block and the prediction block; And

An AxB encoder for generating a bitstream by encoding an AxB residual block including a residual signal of an impulse component in the MxN residual block

An image encoding apparatus comprising a.
The video encoding apparatus of claim 17,

An MxN encoder which generates a bitstream by encoding the MxN residual block; And

An encoding mode determiner which determines which residual block of the MxN residual block and the AxB residual block is to be encoded according to whether the residual signal of the impulse component is included in the MxN residual block

The image encoding apparatus further comprises.
The method of claim 17, wherein the AxB encoder,

And encoding mode information including at least one of block shape information of the AxB residual block, block position information of the AxB residual block, and encoding mode information for identifying AxB encoding, and inserting the same into the bitstream. Device.
In the method of decoding an image,

Decoding a bitstream to extract a quantized frequency coefficient sequence;

Generating an AxB residual block by performing an inverse scan, inverse quantization, and inverse transformation on the quantized frequency coefficient sequence in units of AxB blocks;

Generating an MxN residual block by further combining one or more residual signals to the AxB residual block to generate an MxN residual block;

A prediction step of predicting a current block to generate a prediction block; And

Reconstructing the current block by adding the prediction block and the MxN residual block

Image decoding method comprising a.
The method of claim 20, wherein generating the AxB residual block comprises:

And performing inverse scanning, inverse quantization, and inverse transformation on the quantized frequency coefficient sequence in units of AxB blocks based on block shape information of the AxB residual block included in the mode information extracted from the bitstream.
The method of claim 20, wherein generating the MxN residual block comprises:

And generating the MxN residual block based on block position information of the AxB residual block included in the mode information extracted from the bitstream.
The method of claim 20, wherein the one or more residual signals,

The image decoding method having a predetermined value.
24. The method of claim 23, wherein the one or more residual signals are

All are set to a value of '0' video decoding method.
The method of claim 20,

The MxN residual block generation step,

And using the residual residual signal identified by the residual residual signal information extracted from the bitstream as the one or more residual signals.
The method of claim 20, wherein generating the AxB residual block comprises:

And generating the AxB residual block only when AxB encoding is identified by encoding identification information included in mode information extracted from the bitstream.
27. The method of any of claims 21, 22 and 26, wherein the decoding step comprises:

And extracting the mode information from the bitstream.
The method of claim 20, wherein generating the AxB residual block comprises:

And generating the AxB residual block only when the number of quantization frequency coefficients of the quantization frequency coefficient sequence is equal to or less than a preset number.
In the apparatus for decoding an image,

A decoder which extracts a quantized frequency coefficient sequence by decoding the bitstream;

An AxB residual block generator configured to inversely scan, inverse quantize, and inversely transform the quantized frequency coefficient sequence in units of AxB blocks to generate an AxB residual block;

An MxN residual block generator configured to further combine one or more residual signals to the AxB residual block to generate an MxN residual block;

A prediction unit for predicting a current block to generate a prediction block; And

An adder configured to add the prediction block and the MxN residual block to restore the current block;

Video decoding apparatus comprising a.