Method and apparatus for compensating illumination compensation and method and apparatus for encoding moving picture based on illumination compensation, and method and apparatus for encoding moving picture based on illumination compensation

Illuminance compensation method, apparatus and video encoding method using same, and apparatus using same

1 is a block diagram illustrating an illuminance compensating device according to the present invention.

2(a)(b) are diagrams for explaining a method of calculating an illumination compensation parameter according to the present invention.

3 is a flowchart illustrating a method for compensating for illumination according to the present invention.

4 is a block diagram illustrating a video encoding apparatus to which the illumination compensation method according to the present invention is applied.

5 is a flowchart illustrating a video encoding method to which the illumination compensation method according to the present invention is applied.

6 is a block diagram illustrating a video decoding apparatus to which the illumination compensation method according to the present invention is applied.

7 is a flowchart illustrating a video decoding method to which the illumination compensation method according to the present invention is applied.

8 is a diagram illustrating an encoding sequence in the case of an 8x8 mode.

9 is a diagram illustrating an encoding sequence in the case of a 4x4 mode.

<backgroundart><p>The present invention relates to an illumination compensation method and apparatus, and more particularly, to an illumination compensation method and apparatus for video encoding and decoding.</p><p>Multi-view coding (MVC) for a three-dimensional display application is an incompletely calibrated camera, a different perspective projection direction, and a different reflection effect when predicting between adjacent views. effects), etc., causes changes in illumination between adjacent views, which degrades encoding efficiency. In addition, even in the case of a single view, there is a problem in that encoding efficiency is deteriorated according to a change in illuminance in the case of screen change or the like.</p><p>To solve this problem, the H.264 standard uses a technique called weighted prediction. This weighted prediction technique is applied to motion compensation at the slice level, and the illuminance is compensated according to an appropriate weighted factor W and an additional offset O. As an improved method for such a weighted prediction technique, there is illumination change-adaptive motion estimation/compensation (ICA ME/MC).</p><p>According to the illuminance transformation adaptive motion prediction and compensation method, in the case of the Y component, ICA ME/MC is performed in units of 16X16 blocks, and a difference value of illumination change (DVIC) is obtained for each macroblock. lose </p><p>There are two modes in ICA ME/MC. One of the two modes is the IC-inter 16x16 mode using ICA ME/MC, used in P or B slices. Another mode is the IC-direct 16x16 mode that does not use ICA ME, used only in B slices. In order to compensate for the narrow illuminance transformation, a flag of 1 bit for each inter 16x16 and each direct 16x16 block mode, that is, mb_ic_flag is required.</p><p>Since there is a very high correlation between the DVIC of the current block and the DVIC of the adjacent blocks, the DVIC of the current block takes a method of encoding the difference from the DVIC of the adjacent blocks. </p><p>Hereinafter, ICA ME/MC in macroblock units for the inter 16x16 mode will be described.</p><p>For ICA ME/MC, a new sum of absolute difference (SAD) must be defined. Assuming that the pixel at the (i,j) position of the current frame is f(i,j) and the pixel at the (i,j) position of the reference frame is r(i,j), SAD for blocks of size SxT is calculated as in Equation 1 below. Here, SxT may be 16x16, 16x8, 8x16, 8x8, 8x4, 4x8, 4x4, or the like.</p><p><maths num="1"><df><img file="KR100856411B1_D0001.tif" /></df></maths></p><p>Here, (x,y) is a candidate motion vector, and (m,n) is the position of the current block.</p><p>To compensate for the illuminance transformation, a new SAD is needed. This is obtained by Equations 2 and 3 below.</p><p><maths num="2"><df><img file="KR100856411B1_D0002.tif" /></df></maths></p><p>Here, Mcur is the average value of pixels in the current block, and Mref is the average value of pixels in the reference block. Also, (p, q) is the position of the reference block. A new SAD, that is, NewSAD (x,y) is obtained by Equation 3 below.</p><p><maths num="3"><df><img file="KR100856411B1_D0003.tif" /></df></maths></p><p>In the ICA ME/MC method, a block such that NewSAD (x,y) is minimized, for example, a 16x16 block, is searched based on Equation 3, and an MV corresponding thereto is searched for.</p><p>When the motion vector MV(x', y') that makes NewSAD (x,y) minimize is determined, the illuminance compensation residual signal NewR(i,j) is determined by Equation 4 below.</p><p><maths num="4"><df><img file="KR100856411B1_D0004.tif" /></df></maths></p><p>At this time, the 1-bit flag mb_ic_flag is stored in syntax to indicate whether ICA ME/MC is used. The DPCM value of DVIC is also included in the syntax. In this embodiment, when mb_ic_flag is 0, it indicates that ICA MC is not performed on the current block. Also, when mb_ic_flag is 1, it indicates that ICA MC is performed for the current block.</p><p>In addition, when mb_ic_flag is 1, the ICA ME/MC unit of the decoding apparatus obtains a reconstructed pixel using Equation 5 below.</p><p><maths num="5"><df><img file="KR100856411B1_D0005.tif" /></df></maths></p><p>In Equation 5, NewR``(i,j) is a reconstructed illuminance compensation residual signal, and f`(i,j) represents a pixel in the reconstructed current frame.</p><p>In the conventional ICA ME/MC method, since DVIC information must be transmitted, there is a problem in that encoding efficiency is lowered.</p></backgroundart><abstractproblem><p>The technical problem to be achieved by the present invention is to provide an improved illuminance compensation method in which transmission of DVIC information can be omitted by improving the conventional illuminance compensation method, and a video encoding and decoding method and apparatus employing the illuminance compensation method will do </p></abstractproblem>

<p>In order to achieve the above object, a method of compensating for illumination of a reference block according to the present invention includes the steps of: receiving pixel values of reconstructed neighboring pixels of a current block and pixel values of reconstructed neighboring pixels of a reference block; and performing illuminance compensation on the reference block based on the input pixel values of the restored neighboring pixels of the current block and the pixel values of the restored neighboring pixels of the reference block.</p><p>In addition, in order to achieve the above object, the performing of the illuminance compensation comprises a reference block based on the correlation between the pixel values of the reconstructed neighboring pixels of the current block to be encoded of the current frame and the pixel values of the reconstructed neighboring pixels of the reference block. calculating an illuminance compensation parameter for illuminance compensation of ; It is preferable to include, based on the calculated illuminance compensation parameter, generating an illuminance-compensated reference block.</p><p> In addition, in order to achieve the above task, the step of performing illuminance compensation is</p><p>the following formula</p><p><img file="KR100856411B1_D0006.tif" /></p><p>In a, which minimizes the value of J, a<sb>x, y</sb> and b<sb>x, y</sb>determining; </p><p>as determined above<sb>x, y</sb> and b<sb>x, y</sb>using the following formula</p><p><img file="KR100856411B1_D0007.tif" /></p><p>Further comprising the step of generating an illuminance-compensated reference block based on<sb>x, y</sb> and b<sb>x, y</sb> has a different value depending on the motion vector (x,y), is a constant for each motion vector, and f`(i,-1) and f`(-1,j) are pixels of the reconstructed neighboring pixels of the current block. values, and r'x,y(i,-1) and r'x,y(-1,j) are pixel values of reconstructed neighboring pixels of the motion-compensated reference block, and r'x,y(i, j) denotes a motion-compensated reference block, and is preferably an illuminance-compensated reference block.</p><p>In addition, in order to achieve the above object, the performing of the illuminance compensation may include generating an illuminance-compensated reference block by using an average value of a difference between the values of the restored neighboring pixels of the current block and the values of the restored neighboring pixels of the reference block. It is preferable to further include the step of</p><p>In addition, in order to achieve the above object, the performing of the illuminance compensation may include a reference in which illuminance is compensated using a difference between the average value of the pixel values of the reconstructed neighboring pixels of the current block and the average value of the values of the reconstructed neighboring pixels of the reference block. It is preferred to include the step of generating a block.</p><p>Further, in order to achieve the above object, it is preferable that the reference block is a block within one reference frame among reconstructed frames of an adjacent view in multi-view coding.</p><p>In addition, the above object includes an illuminance compensation unit configured to perform illuminance compensation on the reference block based on the pixel values of the restored neighboring pixels of the current block and the pixel values of the restored neighboring pixels of the reference block. can be achieved by the device.</p><p>In addition, the illuminance compensator according to the present invention for achieving the above object is based on the correlation between the pixel values of the reconstructed neighboring pixels of the current block to be encoded of the current frame and the pixel values of the reconstructed neighboring pixels of the reference block. an illuminance compensation parameter calculator for calculating illuminance compensation parameters for illuminance compensation of the block; It is preferable to include an illuminance-compensated reference block generator that generates an illuminance-compensated reference block based on the calculated illuminance compensation parameter.</p><p>The method further includes: performing illuminance compensation on the reference block based on the pixel values of the reconstructed neighboring pixels of the current block to be encoded and the pixel values of the reconstructed neighboring pixels of the reference block; This may be achieved by a video encoding method based on illumination compensation, which includes performing motion prediction based on a reference block on which illumination compensation has been performed.</p><p>In addition, the object includes: an illuminance compensator for performing illuminance compensation on the reference block based on the pixel values of the reconstructed neighboring pixels of the current block to be encoded and the pixel values of the reconstructed neighboring pixels of the reference block; This may be achieved by an illumination compensation-based video encoding apparatus including a motion prediction unit that performs motion prediction based on the reference block on which illumination compensation has been performed.</p><p>The method further includes: performing illuminance compensation on the reference block based on the pixel values of the restored neighboring pixels of the current block to be decoded and the pixel values of the restored neighboring pixels of the reference block; This may be achieved by a video decoding method based on illumination compensation, which includes performing motion prediction based on a reference block on which illumination compensation has been performed.</p><p>In addition, the object includes: an illuminance compensator for performing illuminance compensation on the reference block based on the pixel values of the restored neighboring pixels of the current block to be decoded and the pixel values of the restored neighboring pixels of the reference block; This may be achieved by an illumination compensation-based video decoding apparatus comprising a motion prediction unit that performs motion prediction based on a reference block on which illumination compensation has been performed.</p><p> The method further includes: receiving pixel values of the restored neighboring pixels of the current block and pixel values of the restored neighboring pixels of the reference block; Based on the input pixel values of the reconstructed neighboring pixels of the current block and the pixel values of the reconstructed neighboring pixels of the reference block, for executing a method of compensating for illumination of a reference block including performing illumination compensation on a reference block It may be achieved by a computer-readable recording medium in which the program is recorded.</p><p>The method further includes: performing illuminance compensation on the reference block based on the pixel values of the reconstructed neighboring pixels of the current block to be encoded and the pixel values of the reconstructed neighboring pixels of the reference block; It can also be achieved by a computer-readable recording medium in which a program for executing a moving picture encoding method based on illumination compensation including performing motion prediction based on the reference block on which illumination compensation is performed is recorded.</p><p>The method further includes: performing illuminance compensation on the reference block based on the pixel values of the restored neighboring pixels of the current block to be decoded and the pixel values of the restored neighboring pixels of the reference block; It can also be achieved by a computer-readable recording medium in which a program for executing a motion picture decoding method based on illumination compensation including performing motion prediction based on the reference block on which illumination compensation is performed is recorded.</p><p>Hereinafter, a preferred embodiment according to the present invention will be described in detail with reference to the accompanying drawings.</p><p>1 is a block diagram illustrating an illuminance compensating device according to the present invention.</p><p>The illuminance compensation apparatus 100 includes an illuminance compensation parameter calculator 120 and an illuminance-compensated reference block generator 140 .</p><p>The illuminance compensation parameter calculator 120 calculates an illuminance compensation parameter based on the restored peripheral pixel values of the current block and the restored peripheral pixel values of the reference block, and outputs the illuminance compensation parameter to the illuminance-compensated reference block generator 140 . do. In the case of single-view coding, the reference block is the reference block of the reconstructed previous frame. In addition, in the case of selective multi-view coding, it may be a block within one reference frame among reconstructed frames of an adjacent view.</p><p>Hereinafter, a method of calculating an illumination compensation parameter according to the present invention will be described with reference to FIGS. 2A and 2B . Hereinafter, each case of the 16x16 mode, the 8x8 mode, the 4x4 mode, and the adaptive mode will be described.</p><p>First, a method of calculating the illuminance compensation parameter in the 16x16 mode will be described with reference to FIGS. 2(a) and 2(b). An illuminance compensation method according to the remaining modes will be described later.</p><p>The thick line block of FIG. 2A is a 16x16 Y block that is a reconstructed reference block corresponding to the motion vector (x,y) in the reference frame. The reconstructed pixel value at the position (i, j) in the reference block corresponding to the motion vector (x, y) is r`<sb>x, y</sb>It is denoted by (i,j). Here, the subscript ` means the restored value. On the other hand, pixels displayed in dark colors are reconstructed neighboring pixel values. The reference frame is a previous frame in the case of a single view, and a restored frame of an adjacent view in the case of the MVC method.</p><p>The bold line block in FIG. 2(b) is the current block, the Y block having a size of 16x16. The original pixel value at the (i,j) position in the current block is expressed as f(i,j). The neighboring pixels indicated in dark colors are used for motion estimation compensation. The neighboring pixels used for motion estimation compensation are reconstructed pixels. As in Fig. 2(a), the subscript ` means a restored value.</p><p>As in Equation 6, the predictor of f(i,j) in the current block is r`<sb>x, y</sb>It can be expressed as a linear function of (i,j).</p><p><maths num="6"><df><img file="KR100856411B1_D0008.tif" /></df></maths></p><p>The predictors obtained in Equation 6 are clipped to [0,255] in the case of an 8-bit image.</p><p>As in Equation 6, a<sb>x, y</sb> and b<sb>x, y</sb>When α is determined, a predictor can be calculated. In the present embodiment, paying attention to the fact that the difference between the illuminance of the pixels inside the current block and the illuminance of the neighboring pixels of the current block is not large, the motion vector (x, y) using the neighboring pixels of FIG. 2(a)(b) illuminance compensation parameters corresponding to a<sb>x, y</sb> and b<sb>x, y</sb>to calculate</p><p>That is, using the surrounding reconstruction pixels of the current block and the reference block, a<sb>x, y</sb> and b<sb>x, y</sb> Since we can determine a from encoder to decoder<sb>x, y</sb> and b<sb>x, y</sb>is not transmitted through the syntax, but in the decoder<sb>x, y</sb> and b<sb>x, y</sb>can be calculated. Accordingly, there is an effect of reducing transmission data.</p><p>In the following, a according to the present invention<sb>x, y</sb> and b<sb>x, y</sb> Three embodiments for the determination will be described. Three embodiments are a linear regression method, an Average of Difference based Prediction (ADP) method, and a Difference of Average based Prediction (DAP) method.</p><p>First, a linear regression method will be described with reference to Equations 7 and 8 below.</p><p>According to the linear regression method, a that minimizes the value of Equation 7 below.<sb>x, y</sb> and b<sb>x, y</sb>can be decided</p><p><maths num="7"><df><img file="KR100856411B1_D0009.tif" /></df></maths></p><p>Here, f`(i,-1) and f`(-1,j) are pixel values of neighboring pixels of the current block. Also, r`<sb>x, y</sb>(i,-1) and f`(-1,j) are pixel values of neighboring pixels of the motion-compensated reference block. Also, a<sb>x, y</sb> and b<sb>x, y</sb>The value of is changed according to the motion vector (x,y), and is a constant for each motion vector.</p><p>a to minimize J in Equation 7 through partial differentiation<sb>x, y</sb> and b<sb>x, y</sb>is calculated as in Equation 8 below.</p><p><maths num="8"><df><img file="KR100856411B1_D0010.tif" /></df></maths></p><p>When the size of the current block and the reference block is 16x16, in Equation 8, N is 32, f`(n) means the restored neighboring pixels of the current block, and f`(-1,j) in FIG. Corresponds to any one pixel among f`(i,-1). where i and j are values between 0 and 15. Also, r`<sb>x, y</sb>(n) degree r`<sb>x, y</sb>(i,-1) and r`<sb>x, y</sb>Corresponds to any one pixel among (-1,j).</p><p>Below, with reference to Equation 9, a according to the Average of Difference based Prediction (ADP) method<sb>x, y</sb> and b<sb>x, y</sb> Describe the calculation method.</p><p>In the difference mean-based prediction method, a<sb>x, y</sb> is fixed to 1, and b<sb>x, y</sb> only way to save. i.e. b<sb>x, y</sb>A method of determining as the average of differences between surrounding reconstruction pixels, b<sb>x, y</sb> may be determined by Equation 9 below. </p><p><maths num="9"><df><img file="KR100856411B1_D0011.tif" /></df></maths></p><p>In the following, a according to the difference of average based prediction (DAP) method with reference to Equation 10<sb>x, y</sb> and b<sb>x, y</sb> Describe the calculation method.</p><p>In the mean difference-based prediction method, a<sb>x, y</sb> is fixed to 1, and b<sb>x, y</sb> only way to save. i.e. b<sb>x, y</sb>A method for determining the difference of the average between surrounding reconstruction pixels, b<sb>x, y</sb> may be determined by Equation 10 below. </p><p><maths num="10"><df><img file="KR100856411B1_D0012.tif" /></df></maths></p><p>As such, in the illuminance compensation parameter calculation unit 120, a linear regression method, a difference average-based prediction method, and a mean difference-based prediction method, a<sb>x, y</sb> and b<sb>x, y</sb> , and outputs it to the illuminance-compensated reference block generator 140 . </p><p>In the illuminance-compensated reference block generator 140 , the input illuminance compensation parameter a<sb>x, y</sb> and b<sb>x, y </sb>is used to generate an illuminance-compensated reference block including a predictor corresponding to f(i,j) according to Equation 6, and output the generated illuminance-compensated reference block to a motion predictor (not shown). .</p><p>FIG. 3 is a flowchart for explaining a method of compensating for illumination performed by the apparatus of FIG. 1 . </p><p>In operation 310, the pixel values of the restored neighboring pixels of the current block and the pixel values of the restored neighboring pixels of the reference block are received. In the case of single-view coding, the reference block is a reference block in the reconstructed previous frame. In addition, in the case of selective multi-view coding, it is a block in one reference frame among reconstructed frames of an adjacent view.</p><p>In operation 320, illuminance compensation is performed on the reference block based on the input pixel values of the restored neighboring pixels of the current block and the pixel values of the restored neighboring pixels of the reference block. In operation 320, the illuminance compensation parameter is generated based on the correlation between the pixel values of the reconstructed neighboring pixels of the current block of the current frame and the pixel values of the reconstructed neighboring pixels of the reference block as shown in Equations 7 to 10 of the current frame. Also, an illumination-compensated reference block is generated according to Equation (6) by using the generated illumination compensation parameter.</p><p>4 is a block diagram illustrating a video encoding apparatus to which the illumination compensation method according to the present invention is applied.</p><p>The video encoder according to the present invention includes a transform and quantization unit 410, an inverse transform and inverse quantization unit 420, a frame storage unit 430, an illuminance compensator 440, an ME/MC unit 450, and a first adder. 460 , a second adder 462 , and an entropy encoder 470 .</p><p>The transform and quantization unit 410 transforms input image data in order to remove spatial redundancy of the image data. In addition, transform coefficient values obtained by transform encoding are quantized according to a predetermined quantization step to obtain N×M data, which is two-dimensional data composed of quantized transform coefficient values. An example of the image transformation used may be a DCT (Discrete Cosine Transform). Quantization is performed according to a predetermined quantization step.</p><p>The inverse transform and inverse quantization unit 420 inversely quantizes the image data quantized by the transform and quantization unit 410 and inversely transforms the inversely quantized image data, for example, inverse DCT. </p><p>The second adder 462 generates a reconstructed image by adding the predicted image output from the ME/MC unit 450 to the data reconstructed from the inverse transform and inverse quantization unit 420 .</p><p>The frame storage unit 430 stores the image restored by the second adder 462 in units of frames. </p><p>The illuminance compensator 440 receives the restored peripheral pixel values of the current block and the restored peripheral pixel values of the reference block input from the frame storage 430 , generates an illuminance-compensated prediction block, and compensates the generated illuminance. The predicted block is output to the ME/MC unit 450 . Optionally, when the input image is an MVC-based image, the neighboring pixel values of the reference block are reconstructed neighboring pixel values of the reference block located within the frame of the neighboring view, in which case the neighboring pixel values of the reference block are stored in the MVC-based reference frame It is input from a part (not shown).</p><p>Since the illuminance compensator 440 performs the same function as the illuminance compensator 100 of FIG. 1 according to the present invention, a detailed description of the illuminance compensator 440 will be omitted for simplicity of description.</p><p>The ME/MC unit 450 estimates a motion vector (MV) per macroblock based on the input image data of the current frame and the illuminance-compensated reference block output from the illuminance compensator 440 . Also, a motion-compensated prediction region P, for example, a 16×16 region selected by motion estimation, is generated based on the estimated motion vector and output to the first adder 460 .</p><p>That is, the ME/MC unit 450 uses the illuminance-compensated reference block obtained by the illuminance compensator 440, and uses the illuminance-compensated SAD value corresponding to the current block according to Equation 11 below, that is, the IC-SAD ( Illumination compensated SAD) is obtained. In addition, a final motion vector is searched for by comparing IC SAD values.</p><p><maths num="11"><df><img file="KR100856411B1_D0013.tif" /></df></maths></p><p>In Equation 11, a<sb>x, y</sb> and b<sb>x, y</sb> The value of is changed according to the motion vector (x,y), and is a constant for each motion vector.</p><p>Optionally, the ME/MC unit 450 further includes a comparator (not shown).</p><p>The comparison unit uses an encoding method based on the input image data of the current frame and the illuminance-compensated reference block output from the illuminance compensator 440, and the input image data of the current frame and the illuminance-compensated output from the frame storage unit 430. The efficiencies of the encoding schemes based on the reference blocks that are not referenced are compared, and a scheme with high encoding efficiency is selected. In this case, the ME/MC unit generates a motion-compensated prediction region P obtained according to the selected motion prediction and compensation scheme, and outputs it to the first adder 460 .</p><p>Also, it is possible to selectively transmit 1-bit flag information indicating whether illumination is compensated in units of macroblocks to the decoder through syntax. In addition, it is also possible to selectively evaluate the performance of the illuminance compensation method in units of GOPs or slices to transmit flag information indicating whether illuminance is compensated in units of GOPs or slices.</p><p>The first adder 460 inputs difference information between the original image and the predictor output from the ME/MC unit 40 to the transform and quantization unit 410 in a predetermined block unit. If the final motion vector for the current block determined through the motion vector search process is (p, q), the first adder 460 calculates the residual (p, q) of the current block as shown in Equation 12 below. and output it to the transform and quantization unit 410 .</p><p><maths num="12"><df><img file="KR100856411B1_D0014.tif" /></df></maths></p><p>In Equation 12, a<sb>p, q </sb>and b<sb>p, q</sb> is a variable value depending on the motion vector (p, q), and is a constant for each motion vector.</p><p>The entropy encoding unit 470 receives information about the quantized transform coefficients output from the transform and quantization unit 410 and the motion vector output from the motion prediction and compensation unit, and entropy-encodes the resulting encoded bitstream. print out</p><p>5 is a flowchart illustrating an encoding method performed in the video encoding apparatus to which the illumination compensation method according to the present invention of FIG. 4 is applied.</p><p>In step 510, transform and quantization are performed.</p><p>In operation 520, inverse transform and inverse quantization are performed on the transformed and quantized data to generate a reconstructed image.</p><p>In operation 530, illuminance compensation is performed on the reference block based on the pixel values of the reconstructed neighboring pixels of the current block to be encoded and the pixel values of the reconstructed neighboring pixels of the reference block. In this embodiment, the reference block is a block in the previous frame. However, optionally, when the input image is an MVC-based image, the reference block is a block within a frame of an adjacent view.</p><p>In operation 540, a prediction block is generated by performing motion prediction and compensation based on the reference block on which the illumination compensation has been performed.</p><p>In operation 550, a residual image is generated based on the original image and the generated prediction block. Transformation and quantization are performed on the generated residual image, and then the transformed and quantized residual image is entropy-encoded together with a motion vector obtained during motion prediction.</p><p>6 is a block diagram illustrating a video decoding apparatus to which the illumination compensation method according to the present invention is applied.</p><p>The video decoding apparatus shown in FIG. 6 includes an entropy decoding unit 610 , an inverse quantization and inverse transformation unit 620 , a frame storage unit 630 , an illuminance compensator 640 , an ME/MC unit 650 , and an adder unit. (660).</p><p>The entropy decoding unit 610 entropy-decodes the encoded input stream to extract image data, motion vectors, and the like. The entropy-decoded image data is input to the inverse quantization and inverse transform unit 620 , and motion vector information is input to the ME/MC unit 650 .</p><p>The inverse transform and inverse quantization unit 620 performs inverse transform and inverse quantization on the image data extracted by the entropy decoder 610 .</p><p>The frame storage unit 630 stores the image data inversely quantized and inversely transformed by the inverse transform and inverse quantization unit 620 in units of frames. </p><p>The illuminance compensator 640 receives the restored peripheral pixel values of the current block and the restored peripheral pixel values of the reference block input from the frame storage 630 , generates an illuminance-compensated reference block, and compensates the generated illuminance. The predicted block is output to the ME/MC unit 650 . Since the illuminance compensator 640 performs the same function as the illuminance compensator of FIG. 1 according to the present invention, a detailed description of the illuminance compensator 640 will be omitted for simplicity of description.</p><p>Optionally, when the video decoding apparatus of FIG. 6 is an MVC-based decoding apparatus, the neighboring pixel values of the reference block are reconstructed neighboring pixel values of the reference block located in the frame of the neighboring view, in this case the neighboring pixel values of the reference block is input from an MVC-based frame storage unit (not shown). </p><p>The ME/MC unit 650 estimates a motion vector MV per macroblock based on the input image data of the current frame and the illuminance-compensated reference block output from the illuminance compensator 640 . In addition, a motion-compensated prediction region P, for example, a 16×16 region selected by motion estimation, is generated based on the estimated motion vector and output to the adder 660 .</p><p>That is, the ME/MC unit 650 uses the illuminance-compensated reference block obtained by the illuminance compensator 640 to obtain the illuminance-compensated SAD value corresponding to the current block, ie, the IC-SAD value, according to Equation (11). save In addition, a final motion vector is searched for by comparing IC SAD values.</p><p>Also, optionally, the ME/MC unit 650 further includes a comparator (not shown). The comparison unit uses an encoding method based on the input image data of the current frame and the illuminance-compensated reference block output from the illuminance compensator 640, and the input image data of the current frame and the illuminance-compensated output from the frame storage unit 630. The efficiencies of coding schemes based on non-referred reference blocks are compared, and a scheme based on motion prediction and compensation with high coding efficiency is selected. At this time, the ME/MC unit generates a motion-compensated prediction area P obtained according to the selected motion prediction and compensation method, and outputs it to the adder 660 .</p><p>In addition, optionally, motion prediction based on an illumination-compensated reference block, which is determined based on flag information indicating whether illumination is compensated, extracted from the entropy decoder 610, and a compensated prediction block or a non-illuminance-compensated reference block based on The motion prediction and the compensated prediction block are output to the adder 660 .</p><p>The adder 660 adds the image reconstructed by the inverse transform and inverse quantization unit 620 and the predictor output from the ME/MC unit 650 to the display unit (not shown) and the frame storage unit 630 . print out</p><p>According to the present embodiment, in the adder 660, residue`(p,q), which is a reconstructed residual signal obtained by reconstructing the residual signal residue(p,q) input from the inverse quantization and inverse transform unit 620, and ME/MC By adding the pixel values of the prediction block input from the unit 650, the final reconstructed pixel f`(x,y) in the current block is calculated according to Equation 13 below.</p><p><maths num="13"><df><img file="KR100856411B1_D0015.tif" /></df></maths></p><p>In Equation 13, a<sb>p, q </sb>and b<sb>p, q</sb> is a variable value depending on the motion vector (p, q), and is a constant for each motion vector.</p><p>7 is a flowchart illustrating a video decoding method performed by the video decoding apparatus according to the present invention shown in FIG. 6 .</p><p>In operation 710, image data and motion vectors are extracted by entropy-decoding the encoded input stream.</p><p>In step 720, inverse transform and inverse quantization are performed on the image data extracted in step 710.</p><p>In operation 730, image data restored by inverse quantization and inverse transformation is stored in units of frames.</p><p>In operation 740, illuminance compensation is performed on the reference block based on the pixel values of the reconstructed neighboring pixels of the current block to be decoded and the pixel values of the reconstructed neighboring pixels of the reference block. In this case, when the input stream is MVC-based, the neighboring pixel values of the reference block are the restored neighboring pixel values of the reference block in the frame of the neighboring view.</p><p>In operation 750, a prediction block is generated by performing motion prediction and compensation based on the reference block on which the illumination compensation has been performed and the extracted motion vector.</p><p>In operation 760, a reconstructed image is generated using the prediction block generated in operation 750. For example, a reconstructed image is generated by adding a prediction block to an image on which inverse transformation and inverse quantization have been performed.</p><p>Hereinafter, an illumination compensation method and a motion compensation method in the 8x8 mode will be described with reference to FIG. 8 .</p><p>In the case of processing in the 8x8 mode, all four 8x8 subblocks exist in a block having a size of 16x16, and each has a different motion vector. In this case, an example of the processing sequence is shown in FIG. In the 8x8 mode, only the processing unit for performing illumination compensation is changed to the 8x8 sub-block, and the processing method is the same as in the 16x16 mode. that is, a<sb>x, y</sb> and b<sb>x, y</sb> is obtained through the linear regression method, ADP method, and DAP method. Also, the motion compensation process in the video encoding apparatus and the video decoding apparatus is the same as in the case of the 16x16 mode, except that the motion compensation process is performed in units of 8x8 blocks.</p><p>Hereinafter, an illumination compensation method and a motion compensation method in a 4x4 mode will be described with reference to FIG. 9 .</p><p>When processing in the 4x4 mode, there are all 16 4x4 sub-blocks in a block having a size of 16x16, and each has a different motion vector. In this case, an example of the processing sequence is shown in FIG. In the 4x4 mode, only the processing unit for performing illumination compensation is changed to the 4x4 sub-block, and the processing method is the same as in the 16x16 mode. i.e. a<sb>x, y</sb> and b<sb>x, y</sb> is obtained through the linear regression method, ADP method, and DAP method. Also, the motion compensation process in the video encoding apparatus and the video decoding apparatus is the same as in the case of the 16x16 mode, except that the 4x4 block unit is used.</p><p>Hereinafter, an illumination compensation method and a motion compensation method in the adaptive mode will be described.</p><p>In the case of the adaptive mode, the video encoding apparatus may perform illuminance compensation and encoding according to the 16x16 mode, the 8x8 mode, and the 4x4 mode in units of macroblocks, and may select one of them. In this case, the residual block according to the mode is encoded and transmitted. In this case, the information indicating the corresponding mode is transmitted to the video decoding apparatus. Also, optionally, when illumination compensation and motion estimation are possible for each 16x16 mode, 8x8 mode, and 4x4 mode in the decoding apparatus, it is possible even without transmitting mode information.</p><p>In this embodiment, only the 16x16 mode, the 8x8 mode, and the 4x4 mode have been dealt with, but the illuminance compensation method according to the present invention can be applied to other arbitrary block-based motion compensation schemes.</p><p>The present invention can also be implemented as computer-readable codes on a computer-readable recording medium. The computer-readable recording medium includes all types of recording devices in which data readable by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, etc. include In addition, the computer-readable recording medium is distributed in a computer system connected to a network, so that the computer-readable code can be stored and executed in a distributed manner.</p><p>So far, with respect to the present invention, the preferred embodiments have been looked at. Those of ordinary skill in the art to which the present invention pertains will understand that the present invention can be implemented in a modified form without departing from the essential characteristics of the present invention. Therefore, the disclosed embodiments are to be considered in an illustrative rather than a restrictive sense. The scope of the present invention is indicated in the claims rather than the foregoing description, and all differences within the scope equivalent thereto should be construed as being included in the present invention.</p>

<p>As described above, in the illuminance compensation method according to the present invention and the video encoding and decoding method to which the illuminance compensation method according to the present invention is applied, there is an effect that encoding efficiency can be increased because parameters for illuminance compensation do not need to be transmitted. .</p>