KR20060122671A - Method for scalably encoding and decoding video signal - Google Patents

Abstract

A method for scalably encoding and decoding a video signal is provided to perform the prediction of a video at an FGS(Fine Grained Scalability) enhanced layer. An FGS encoder calculates an FGS base layer with a quantization coefficient through DCT(Discrete Cosine Transform) conversion and quantization with a predetermined step size related to the image data of X. The FGS encoder restores the image data of the FGS base layer from the FGS base layer with the quantization coefficient and calculates the FGS similarity enhanced layer with a quantization coefficient through the DCT conversion related to the difference of the image data and the restored FGS base layer, and the quantization with a small step size.

Description

Scalable method for encoding and decoding video signals {Method for scalably encoding and decoding video signal}

1 illustrates a conventional method for improving coding efficiency in connection with FGS,

2 illustrates an embodiment according to the present invention for a method of encoding an FGS enhanced layer through a closed-loop structure at a macroblock level,

3 illustrates an example of a method of scanning and transmitting a quantization coefficient of an FGS enhanced layer.

4 illustrates an embodiment of a slice header area according to the present invention;

5 illustrates an embodiment of a macroblock header region according to the present invention.

The present invention relates to a scalable encoding and decoding method of an image signal, and more particularly, to a method of performing a prediction operation on an FGS enhanced layer through a closed-loop structure based on a macroblock. .

For digital video signals transmitted and received wirelessly by mobile phones and laptops and mobile TVs and handheld PCs, which are widely used in the future, it is difficult to allocate wide bands as in TV signals. Therefore, the standard to be used for the image compression method for such a mobile portable device should be higher the compression efficiency of the video signal.

In addition, such a mobile portable device is inevitably varied in its ability to process or present. Therefore, the compressed image has to be prepared in such a variety that it is different from each other by various variables such as transmission frames per second, resolution, bits per pixel, etc. for the same image source. This means that it must be provided, which is a burden on the content provider.

For this reason, the content provider has high-speed bitrate compressed video data for one video source, decodes the compressed video when requested by the mobile device, and then fits the video capability of the requested device. This is provided by re-encoding the video data. However, this method requires a transcoding (decoding + scaling + encoding) process, and thus a time delay occurs in providing a video requested by the mobile device. Transcoding also requires complex hardware devices and algorithms as the target encoding varies.

Scalable video codec (SVC) has been proposed to solve such disadvantages. This method encodes a video signal, and encodes at the highest quality, but decodes a partial sequence of the resulting picture (frame) sequence (sequence of frames selected intermittently throughout the sequence) to some extent. It's a way to ensure that. Motion Compensated Temporal Filter (or MCTF) is an example of an encoding scheme proposed for use with the scalable video codec.

Scalability is a new concept introduced in MPEG-2 to improve error tolerance and adaptability to bit rates. The scalability includes a base layer consisting of a screen having a low resolution or a small size, and an enhanced layer or enhancement layer consisting of a screen having a higher resolution or a larger size. The base layer is a bit stream encoded to be independently decodable, and the enhanced layer is a bit stream generally used to improve a bit stream in the base layer, for example, original data and data encoded in the base layer. This is a more detailed encoding of the difference between. Scalability includes spatial scalability, temporal scalability, SNR scalability and the like.

SNR scalability is a method of improving image quality, and transmits transform coefficients corresponding to each pixel, for example, discrete cosine transform (DCT) coefficients, divided into base layer and enhanced layer according to the resolution of the bit representation. do.

Data generated through motion estimation, prediction, and the like for the video signal is transformed and quantized, for example, into DCT transform coefficients. In this case, the transform coefficient is quantized according to a step size set to correspond to a predetermined quality (or a predetermined bit rate) in the quantization process, and the generated quantization coefficient becomes an SNR base layer.

The quantization process obtains higher compression efficiency by expressing transform coefficients as finite representative values according to the quantization step size. Higher compression is possible through the quantization process, but once the quantized value cannot be restored to the original video signal. There is a loss when rebuilding.

In order to compensate for the loss (error) that occurs in the encoding process (DCT transform and quantization), the DCT and the quantization process are performed for the difference between the original data and the data reconstructed through inverse quantization and inverse DCT of the SNR base layer. Perform again to generate level 1 of the SNR enhanced layer. In this case, the step size in the quantization process is set to correspond to a quality one step higher than the predetermined quality corresponding to the SNR base layer, that is, smaller than that in the quantization process of the SNR base layer. This is because the quantization process is performed on the difference value of the data.

As described above, DCT conversion of the difference value between the original data and the restored data is repeated, and the quantization step is repeated according to the quantization step size set in the above-described manner, thereby compensating for errors occurring in the encoding process such as the DCT transformation and the quantization process. Several levels of SNR enhanced layers are created in turn. In addition, as the quality of the SNR enhanced layer being decoded may gradually increase in quality, the SNR scalability may be improved by fine grained scalability (GFS), progressive refinement, or continuous refinement (FG). successive refinement). The SNR base layer and the SNR enhanced layer are also referred to simply as the base layer and the FGS layer.

Conventionally, as a method for improving coding efficiency in relation to FGS, as shown in FIGS. 1A to 1B, a base layer of an entire frame (SNR base layer) or an FGS layer of an entire frame (SNR). It is limited to a method of encoding a base layer (SNR base layer) using reconstructed image data of an enhanced layer.

Since the FGS layer is related to a coefficient domain in which image data is transformed, a prediction operation on the FGS layer has little meaning in the conventional FGS structure, and thus a method for improving coding efficiency is not appropriate.

Therefore, unlike the conventional FGS encoding structure that encodes the difference between the original data and the base layer into the FGS layer, the structure that predicts the FGS layer using reconstructed image data of the base layer of the same frame is It is proposed as a method for improving coding efficiency with respect to the FGS layer. In addition, in order to use temporal redundancy in the FGS layer, a closed-loop that is appropriately weighted to the reconstructed image data of the base layer of the same frame as the FGS layer of the previous frame and used for prediction of the FGS layer. The structure using the closed-loop structure remains based on the slice level.

The present invention has been made to solve this problem, and an object of the present invention is to provide a method for improving coding efficiency with respect to encoding of an FGS layer.

A specific object of the present invention is to provide a method for predicting an FGS enhanced layer using the FGS enhanced layer of another frame and the same frame at the macroblock level and a method for defining the weight at the macroblock level. .

In order to achieve the above object, a method of encoding a video signal according to an embodiment of the present invention is based on a macroblock, by using a first layer for a first frame and a second layer for a second frame. Encoding a second layer for the first frame; And recording information indicating that the second layer is encoded based on the macroblock, wherein the second layer is encoded to compensate for an error occurring in the encoding process of the first layer, and the second layer is encoded. The frame may be a frame which is a reference for motion prediction with respect to the first frame.

The information is characterized in that the information is recorded in the header area of the slice including the second layer for the first frame.

The first image data reconstructed from the first layer for the first frame and the second image data reconstructed from the second layer for the second frame are used for encoding the second layer for the first frame. It is done. In this case, the second image data may be reconstructed using both the first layer and the second layer for the second frame.

The first image data and the second image data are each weighted and used, wherein a first weight that weights either the first image data or the second image data is a macroblock of a second layer with respect to the first frame. The second weight recorded in the header area of the apparatus and weighting other image data other than the one is inferred from the first weight.

In addition, the first weight is recorded according to the condition of the first layer for the first frame, wherein the condition of the first layer is characterized in that the presence or absence of data of the sub block in the macroblock of the first layer. .

The image data obtained by weighting and adding the similar image data for the second layer for the first frame and the first image data and the second image data, respectively, generated using the first layer for the first frame. The second layer for the first frame is encoded using the difference.

According to the present invention, a method of decoding an encoded video bit stream is based on a macroblock, using a first layer for a first frame and a second layer for the first frame using a second layer for a second frame. Decoding a second layer, wherein the second layer is encoded to compensate for an error occurring in the encoding process of the first layer, and the second frame is a frame that is a reference of motion prediction for the first frame. It is characterized by that.

The image data for the second layer of the first frame is weighted with first image data reconstructed from the first layer for the first frame and second image data reconstructed from the second layer for the second frame. And the image data obtained by reconstructing only the second layer of the first frame.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The As described above, in the conventional FGS encoding method, to obtain the FGS base layer (X _B ^Q) having the quantization coefficients via a DCT transform and quantization with a predetermined step size of the original image data (X _O ^D) of the X _O FGS enhanced layer has a ^D and the X _B image data (X _B ^R) of the base layer restored from ^Q quantized coefficients after the quantization step size smaller set than the DCT transform and the predetermined step size for the difference of ( X _E ^Q )

However, in the FGS encoding method of the closed-loop structure, the quantization coefficient of the FGS enhanced layer is not directly calculated from the difference between X _O ^D and X _B ^R.

The FGS encoder of the closed-loop structure obtains the X _B ^Q using a conventional FGS encoding method and uses the same to obtain an FGS pseudo enhanced layer (X _PE ^Q ) having a quantization coefficient. Thereafter, the X _PE ^Q is reconstructed to obtain image data (X _PE ^R ), and this is added to X _B ^R obtained by reconstructing the X _B ^Q to obtain image data for an FGS-like enhanced layer.

Also, the FGS encoder of the closed-loop structure may include the reconstructed image data (X _B ^R ) of the FGS base layer of the same frame (current frame) and the frame that is the reference (reference) of the current frame, for example, the previous frame. Predicted (reference) image data for the FGS enhanced layer is generated by weighting and adding image data for the FGS enhanced layer to a predetermined weight.

Thereafter, the FGS encoder having the closed-loop structure includes an FGS enhanced having a quantization coefficient through a DCT transform and quantization on a difference between the image data of the FGS-like enhanced layer and the prediction (reference) image data of the FGS enhanced layer. Find the layer.

Meanwhile, the closed-loop FGS decoder obtains image data of the FGS enhanced layer of the reference frame from the quantization coefficients of the FGS base layer and the FGS enhanced layer of the already received reference frame, and receives the FGS base layer of the current frame already received. Reconstruct the video data of the FGS base layer of the current frame from. The closed-loop FGS decoder adds the image data of the FGS enhanced layer of the reference frame and the FGS base layer of the current frame by the predetermined weight, thereby predicting (reference) the FGS enhanced layer of the current frame. ) Generate video data.

The closed-loop FGS decoder adds the image data reconstructed from the quantization coefficients of the FGS enhanced layer of the current frame received thereafter and the generated prediction (reference) image data to add an image of the FGS enhanced layer of the current frame. You can create data.

Therefore, even when the quantization coefficient of the FGS enhanced layer is not received due to degradation of the transmission environment such as a reduction in transmission band and truncation, the closed-loop FGS decoder improves image quality than the image data of the FGS base layer. The predicted image data can be obtained.

2 illustrates an embodiment according to the present invention for a method of encoding an FGS layer (enhanced layer) through a closed-loop structure at the macroblock level.

For convenience, an arbitrary macroblock in the current frame is referred to as X, a block of a reference frame corresponding to X, that is, a reference block used when performing an inter-frame motion prediction operation on the X.

According to the present invention, the FGS encoder performs quantization coefficients through DCT transform and quantization of a predetermined step size with respect to the image data of X (which may be image data encoded as a difference value of an image through a motion prediction operation) (X _O ^D ). Obtain an FGS base layer (X _B ^Q ) with. In addition, the FGS encoder restores the image data (X _B ^R ) of the FGS base layer from the X _B ^Q , and compares the DCT transform and the predetermined step size with respect to the difference between X _O ^D and the restored X _B ^R. The FGS pseudo-enhanced layer X _PE ^Q having a quantization coefficient is obtained through smaller step size quantization.

Thereafter, the FGS encoder restores image data (X _PE ^r ) from the X _PE ^Q and adds the reconstructed X _PE ^r to the X _B ^R to obtain image data (X _PE ^R ) for an FGS-like enhanced layer. Get Here, the X _PE ^r corresponds to image data for quantization coefficients coded only in the FGS-like enhanced layer, and the X _PE ^R corresponds to image for quantization coefficients coded in the FGS base layer and the FGS-like enhanced layer, respectively. It is the sum of the data.

When the image data X _O ^D of X is image data encoded with an image difference value, a process of obtaining original image data by performing an inverse operation of the motion prediction operation on the obtained X _PE ^R may be further required. .

Meanwhile, the FGS encoder obtains an FGS base layer and an FGS enhanced layer having quantization coefficients through FGS coding on a reference frame including the block Y, for example, a previous frame. The FGS encoder may reconstruct image data of the FGS enhanced layer of the reference frame using the quantization coefficients of the FGS base layer and the FGS enhanced layer of the reference frame, and from there, the FGS enhancement for the block Y. Image data Y _E ^R can be obtained.

The FGS encoder generates the prediction (reference) image data (X _E ^P ) for the FGS enhanced layer of the macroblock X by weighting and adding the Y _E ^R and the X _B ^R to a predetermined weight. In this case, the weights applied to the Y _E ^R and the X _B ^R are set to have a complementary relationship such that the sum of the two weights is 1, and the weight for the Y _E ^R is a value between 0 and 1, for example. If expressed and set to any value from 0 to 15 or 0 to 31, the weight for X _B ^R can be easily derived from it.

The FGS encoder then performs DCT on the difference (X _E ^r ) between the image data (X _PE ^R ) for the FGS-enhanced layer and the prediction (reference) image data (X _E ^P ) for the FGS enhanced layer. The FGS enhanced layer X _E ^Q having quantization coefficients is obtained through transformation and quantization. In this case, the quantization step applied to the quantization can be arbitrarily adjusted, but it is preferable to use the quantization step applied when obtaining the quantization coefficient of the FGS-like enhanced layer.

Accordingly, the current frame is represented by the FGS base layer X _B ^Q and the FGS enhanced layer X _E ^Q obtained differently from the conventional FGS encoding method according to the FGS encoding method according to the present invention.

On the other hand, the quantization coefficients of the FGS enhanced layer are not sequentially transmitted in the pixel order of the macroblock, but are scanned and transmitted three times along a predetermined path, for example, a zigzag path.

i) any subblock in the macroblock that has no data in the FGS base layer, that is, the quantization coefficient of the FGS base layer is zero, and ii) the data in the FGS base layer, that is, the quantization coefficient of the FGS base layer is not zero. In a subblock including at least one pixel, pixels having a quantization coefficient of 0 in the FGS base layer are divided into pixels; The quantization coefficients of the FGS enhanced layer are scanned and transmitted according to a predetermined path.

For example, as shown in FIG. 3, a 16x16 macroblock is composed of four 8x8 subblocks, the first subblock is an FGS base layer composed of zero data, and the FGS base layer of the remaining subblocks is zero. At least one pixel having non-quantization coefficients is included. Whether or not the FGS base layer of the sub-block consists of zero data can be easily checked by a CBP (Coded Block Pattern) defined in 8x8 block units.

The FGS encoder sequentially scans and transmits the FGS enhanced layer for the corresponding pixel while advancing the first subblock having the FGS base layer of 0 along the zigzag path (Path 1). Afterwards, the FGS encoder sequentially advances the FGS enhanced layer for pixels having zero data in the FGS base layer while advancing along the zigzag path for the remaining subblocks (the subblock in which the FGS base layer is not 0). Scan to and send. In addition, the FGS encoder sequentially scans and transmits the FGS enhanced layer for a pixel whose data of the FGS base layer is not 0 while proceeding along the zigzag path again with respect to the remaining sub-blocks.

This is because the FGS enhanced layer for a subblock or pixel having no data in the FGS base layer is relatively more important than that of a subblock or pixel that is not.

Meanwhile, as described above, when the FGS encoder generates the prediction image data X _E ^P for the FGS enhanced layer of the macroblock X in the process of generating the FGS enhanced layer, the Y _E ^R And weight the X _B ^R. In this case, in order to apply a more sophisticated weight, the macroblock X is divided into subblocks, and weights are differently applied to X _B ^R (or Y _E ^R ) depending on whether data of the FGS base layer for the sub block is present. Can be.

For a subblock in which the FGS base layer is 0, the FGS base layer is not bound to the complementary relationship between the weights applied to the Y _E ^R and the X _B ^R , and the FGS base layer is applied to the Y _E ^R for the non-zero subblock. Can be weighted to a value different from the weight, i.e., any value between 0 and 1 (W _ZB ).

As shown in FIG. 5, the FGS encoder includes a weight W _ZB (or mb_lvl_base_ref_weight_for_zero_base_block_plus_1 in FIG. 5) applied to X _B ^R (or Y _E ^R ) for a sub block having an FGS base layer of 0 and the remaining sub blocks. The weight (W _NZB ) (mb_lvl_base_ref_weight_for_zero_base_coeff_plus_1) applied to X _B ^R (or Y _E ^R ) may be distinguished and recorded in the header area of the macroblock.

Since the FGS encoder needs to inform the decoder whether the FGS enhanced layer encoded through the closed-loop structure is based on slices or macroblocks, a predetermined flag is shown in FIG. 4. For example, define 'mb_weight_override_present' and write it in the slice header area.

When the FGS encoder encodes an FGS enhanced layer through a closed-loop structure based on a slice, the FGS encoder sets the 'mb_weight_override_present' to, for example, '0' and records the slice in the slice header area. Record the weights applied to X _B ^R (or Y _E ^R ) of the macroblock.

In addition, when the FGS encoder encodes an FGS enhanced layer through a closed-loop structure based on a macroblock, as illustrated in FIG. 5, the FGS encoder sets 'mb_weight_override_present' to '1', for example. Recorded in the slice header area, and the weight applied to the X _B ^R (or Y _E ^R ) associated with the macro block is recorded in the header area of each macro block.

Of course, the weights (W _ZB ) (mb_lvl_base_ref_weight_for_zero_base_block_plus_1) applied to X _B ^R for the subblocks having the FGS base layer 0 are distinguished from the weights (W _NZB ) (mb_lvl_base_ref_weight_for_zero_base_coeff_plus_1) applied to the X _B ^R for the remaining subblocks. Can be recorded in the header area of. Similarly, when encoding based on a slice, a block (or subblock) in which the FGS base layer is 0 and a weight (W _ZB , W _NZB ) (base_ref_weight_for_zero_base_block_plus_1, base_ref_weight_for_zero_base_coeff_plus_1) for the remaining blocks are distinguished from each other. Can record

The data stream encoded by the method described so far is transmitted to the decoding device by wire or wirelessly or transmitted through a recording medium, and the decoding device reconstructs the original video signal according to the method described later.

Meanwhile, the FGS decoder according to the present invention determines that the data included in the slice whose 'slice_type' value recorded in the slice header area is 'Progressive_Refinement' corresponds to the FGS enhanced layer and checks whether 'mb_weight_override_present' exists. . If the 'mb_weight_override_present' exists, it is determined that the FGS enhanced layer in the slice is encoded through the closed-loop structure.

When the value of the 'mb_weight_override_present' is '0', the FGS decoder determines that the weights used to calculate the reference image data used in the process of encoding the FGS enhanced layer are equally applied to all macroblocks in the slice. do. On the other hand, when the value of 'mb_weight_override_present' is '1', the FGS decoder determines that the FGS enhanced layer is encoded by applying the weight differently to each macroblock in the slice.

Next, when the value of 'mb_weight_override_present' is '1', a process of decoding the FGS enhanced layer for an arbitrary macroblock X will be described.

The FGS decoder is used to calculate reference image data of the FGS enhanced layer of the macroblock X from the header area of the macroblock X when the FGS enhanced layer for the current frame including the macroblock X is received. The weights W _ZB and W _NZB (for example, mb_lvl_base_ref_weight_for_zero_base_block_plus_1 and mb_lvl_base_ref_weight_for_zero_base_coeff_plus_1) are checked in FIG. 5.

The FGS decoder restores image data of the FGS enhanced layer of the reference frame from the quantization coefficients of the FGS base layer and the FGS enhanced layer of the reference frame including Y, which is a reference block for the macroblock X, Obtain image data Y _E ^R of the FGS enhanced layer for the block Y. In this case, the reference frame is generally a frame before the frame including X, and the quantization coefficients of the FGS base layer and the FGS enhanced layer of the reference frame are already received.

Further, the FGS decoder is in a state of receiving the quantized coefficients (X _B ^Q) of the FGS base layer for the macroblock X (more precisely frame including the X), the group of the FGS base layer from X _B ^Q Restore the image data (X _B ^R ).

Thereafter, the FGS decoder selects a weight to be applied to the corresponding subblock among the W _ZB or the W _NZB according to the presence or absence of data of the FGS base layer for the subblock in the macroblock X, and uses the selected weight to calculate the X _B By weighting and adding ^R and Y _E ^R , prediction (reference) image data X _E ^P for the FGS enhanced layer of the macroblock X is generated.

Of course, if the W _NZB is weighted to x (i) _B ^R for any sub-block (x (i)) whose non-zero FGS base layer is generated when the reference image data is generated, the x (i) (1-W _NZB ) is applied to y (i) _E ^R for the sub block y (i) in the block Y corresponding to. In addition, y (i) _E ^R for subblock y (i) in the block Y corresponding to any subblock x (j) in which the FGS base layer is 0, W _ZB or (1-W _NZB ) Applied by weight.

The FGS decoder may reconstruct image data X _E ^{r of} only X _E ^Q from the quantization coefficient X _E ^Q of the FGS enhanced layer for the received macroblock X, and then reconstruct the reference image data. In addition to (X _E ^P ), the image data of the FGS enhanced layer for the macroblock X is reconstructed.

If the FGS decoder does not receive the quantization coefficient (X _E ^Q ) of the FGS enhanced layer for the macroblock X, but receives the header of the macroblock X, the weight is included in the header of the macroblock X. It is possible to generate reference image data (X _E ^P ) for the FGS enhanced layer of the macroblock X using X, so that X _E ^P is the image data (X) of the FGS base layer reconstructed from the X _B ^Q. _B ^R ) can provide improved image quality.

The FGS decoder according to the present invention described above may be mounted in a mobile communication terminal or the like or in an apparatus for reproducing a recording medium.

As described above, preferred embodiments of the present invention have been disclosed for the purpose of illustration, and those skilled in the art can improve, change, and further various embodiments within the technical spirit and the technical scope of the present invention disclosed in the appended claims. Replacement or addition may be possible.

Accordingly, image prediction in the FGS enhanced layer is possible based on the macroblock, and even when data of the FGS enhanced layer is partially lost, it is possible to provide an image having higher quality than the FGS base layer, thereby improving coding efficiency. It becomes possible.

Claims (16)

Based on a macroblock, encoding a second layer for the first frame using a first layer for a first frame and a second layer for a second frame; And And recording information indicating that the second layer is encoded based on a macroblock, In this case, the second layer is encoded to compensate for an error occurring during the encoding process of the first layer, and the second frame is a frame that is a reference frame for motion prediction with respect to the first frame. Way. The method of claim 1, Wherein the information is recorded in a header area of a slice including a second layer for the first frame. The method of claim 1, That the first image data reconstructed from the first layer for the first frame and the second image data reconstructed from the second layer for the second frame are used for encoding the second layer for the first frame. Characterized by a video signal encoding method. The method of claim 3, wherein And the second image data is reconstructed using both the first layer and the second layer for the second frame. The method of claim 3, wherein And the first image data and the second image data are each weighted and used. The method of claim 5, And recording a first weight weighting one of the first image data and the second image data in a header area of a macroblock of a second layer for the first frame. And a second weighting factor for weighting other image data other than the one is inferred from the first weighting factor. The method of claim 6, And the first weight is recorded according to the condition of the first layer with respect to the first frame. The method of claim 7, wherein The condition of the first layer is a method of encoding a video signal, characterized in that the presence or absence of data of the sub-block in the macroblock of the first layer. The method of claim 5, The image data obtained by weighting and adding the similar image data for the second layer for the first frame and the first image data and the second image data, respectively, generated using the first layer for the first frame. And encoding a second layer for the first frame using the difference. Based on the macroblock, decoding the second layer for the first frame using the first layer for the first frame and the second layer for the second frame, In this case, the second layer is encoded to compensate for an error occurring in the encoding process of the first layer, and the second frame is an encoded video bit stream, wherein the frame is a reference for motion prediction for the first frame. How to decode it. The method of claim 10, And determining whether the second layer is encoded based on a macroblock from a header region of a slice including the second layer for the first frame. The method of claim 10, That the first image data reconstructed from the first layer for the first frame and the second image data reconstructed from the second layer for the second frame are used for decoding of the second layer for the first frame. Featuring a method of decoding an encoded video bit stream. The method of claim 12, And the second image data is reconstructed using both the first layer and the second layer for the second frame. The method of claim 12, And reading a first weight value for weighting either the first image data or the second image data from a header area of a macroblock of a second layer for the first frame. The one is weighted with the first weight, and other image data other than the one is weighted with a value inferred from the first weight and used for decoding the second layer with respect to the first frame. To decode the encoded video bit stream. The method of claim 14, The first weight is divided and recorded according to the presence or absence of data of a sub block in the macroblock of the first layer with respect to the first frame, and the sub block of the second image data is weighted differently according to the presence or absence of the data. To decode the encoded video bit stream. The method of claim 12, The image data of the second layer of the first frame may include image data obtained by weighting and adding the first image data and the second image data, respectively, and image data obtained by reconstructing only the second layer of the first frame. A method for decoding an encoded video bit stream, characterized in that it is generated.

Images