KR20100136391A - Method and apparatus for encoding and decoding using flexible filter structures - Google Patents

Abstract

PURPOSE: A restoring image filtering method and an encoding/decoding apparatus and a method thereof using an optimum filter are provided to improve the coding efficiency of the reconstructed image by filtering the reconstructed image to the filtering structure in which the degradation degree is minimum. CONSTITUTION: A loop filter(170) determines one filtering structure with a plurality of filtering structures based on the filtering and the filtered reconstructed images with the final filter structure. An entropy encoder(140) encodes filter structure information corresponding to the final filter structure. The loop filter comprises an output unit(171) producing the filter coefficient of the filtering structure and a decision unit deciding the final filter structure.

Description

Reconstruction image filtering method using multiple filters and encoding / decoding device and method using same {METHOD AND APPARATUS FOR ENCODING AND DECODING USING FLEXIBLE FILTER STRUCTURES}

      The present invention relates to a method for filtering a reconstructed image using a plurality of filters and an encoding / decoding device to which the same is applied. / Decryption apparatus.

The present invention is derived from a study conducted as part of the IT source technology development project of the Ministry of Information and Communication and the Ministry of Information and Communication Research and Development (Task Management Number: 2008-F-011-01, Task Name: Development of Next-Generation DTV Core Technology).

Recently, a method for filtering a reconstructed image by obtaining an adaptive filter coefficient for each image after designing an in-loop filter based on a Wiener filter before the reconstructed image is stored in a decoded picture buffer in an encoding apparatus. Is going on. In this method, the filter has a central symmetric structure in order to reduce the amount of information of the filter coefficients to be encoded. In this case, since the filter is designed using only a single filter structure, the characteristics of each image may not be reflected.

The present invention provides a method of filtering a reconstructed image using a plurality of filters, and an encoding / decoding apparatus using the same.

The encoding method according to an embodiment of the present invention comprises the steps of: filtering a reconstructed image by each of the plurality of filter structures, and determining one of the plurality of filter structures as the final filter structure by using the filtered plurality of reconstructed images. It includes.

Here, the plurality of filter structures may include at least one of an adaptive filter structure and a fixed filter structure. In this case, the adaptive filter structure may include at least one of a center, a vertical, a horizontal, and a user-defined symmetric structure, and the fixed filter structure may represent a filter structure having a predetermined filter coefficient.

The filtering may include calculating a filter coefficient corresponding to each of the plurality of filter structures when the plurality of filter structures are the adaptive filter structure, and filtering the reconstructed image using each of the calculated filter coefficients. It may include.

In this case, the encoding may include encoding a final filter coefficient and filter structure information corresponding to the determined final filter structure if the determined final filter structure is an adaptive filter structure, and determining the determined final filter structure if the determined final filter structure is a fixed filter structure. Filter structure information corresponding to the encoding may be encoded.

In addition, the encoding apparatus according to an embodiment of the present invention filters the reconstructed image by each of the plurality of filter structures and determines one of the plurality of filter structures as the final filter structure by using the filtered reconstructed images. It may include a loop filter, and an entropy encoding unit for encoding the filter structure information corresponding to the determined final filter structure.

In addition, if the plurality of filter structures are adaptive filter structures, the loop filter may include a calculating unit configured to calculate a filter coefficient corresponding to each of the plurality of filter structures, and the reconstructed image by using each of the calculated filter coefficients. The determining unit may determine one of the plurality of filter structures as the final filter structure by using the plurality of reconstructed images.

In this case, if the plurality of filter structures are fixed filter structures, the loop filter filters the reconstructed images using predetermined filter coefficients, and determines one of the plurality of filter structures as the final filter structure using the filtered reconstructed images. Can be.

In addition, the decoding method according to an embodiment of the present invention, receiving the encoded filter structure information for each frame, obtaining the final filter structure used during encoding from the received filter structure information, and the obtained final filter structure And filtering the reconstructed image.

In the receiving step, if the final filter structure is an adaptive filter structure, the encoded final filter coefficient corresponding to the final filter structure and filter structure information corresponding to the final filter structure may be received. In this case, the decoding may include decoding the received final filter coefficient and the filter structure information, and the filtering may include filtering the reconstructed image with the final filter structure using the decoded final filter coefficient.

In addition, the decoding apparatus according to an embodiment of the present invention, the entropy decoding unit for receiving the encoded filter structure information for each frame, and obtains the final filter structure used during encoding from the received filter structure information, and obtained the final filter And a loop filter for loop filtering the reconstructed image with the structure.

According to the present invention, the reconstructed image is filtered by the filter structure having the minimum degree of degradation, and the reconstructed image may be filtered and encoded in consideration of characteristics of the image, and the encoding efficiency may be improved.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, in describing the present invention, when it is determined that the detailed description of the related air technology or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be abbreviated or omitted.

1 is a block diagram showing a configuration of an encoding apparatus according to an embodiment of the present invention.

Referring to FIG. 1, the encoding apparatus 100 according to an embodiment of the present invention may include an intra prediction unit 105, a motion estimator 110, a motion compensator 120, a first difference unit 125, And a DCT and quantization unit 130, an entropy encoding unit 140, an inverse quantization and inverse DCT unit 150, a second difference unit 155, a deblocking unit 160, and a loop filter 170.

The intra prediction unit 105 predicts and encodes a macroblock to be encoded using spatially adjacent pixel values in order to encode a block of a current image.

The motion estimation unit 110 generates a motion vector of a preset current block in a current frame to be encoded using a reference frame stored in a buffer (not shown) or a storage unit (not shown). The reference frame may be a previous frame continuously inputted before the current frame or a next frame continuously inputted after the current frame.

The motion estimator 110 may estimate the motion by using various linearity estimation algorithms such as block matching algorithm (PMA), phase correlation, and HSBMA.

The motion compensator 120 generates a predicted block that is a predicted value of the current block based on the motion vector generated by the motion estimator 110. The first difference unit 125 generates a difference block by subtracting the prediction block from the current block.

The DCT and quantization unit 130 performs discrete cosine transform on the difference block generated by the first difference unit 125 and quantizes the quantized transform coefficient.

The entropy encoder 150 entropy encodes encoding information such as a quantized transform coefficient and a motion vector, filter structure information received from the loop filter 170 to be described later, and a final filter coefficient. In this case, the encoded transform coefficient, the filter structure information, and the final filter coefficient are inserted into the bitstream together with the motion vector and transmitted to the decoding apparatus 400.

The inverse quantization and inverse DCT unit 150 inverse quantizes the differential blocks quantized in the DCT and the quantization 130 for prediction of a frame to be encoded next, and inverse discrete cosine transform to restore the differential blocks before encoding. .

The second difference unit 155 reconstructs the current block before encoding by adding the reconstructed differential block and the prediction block generated by the motion compensator 120.

The deblocking unit 160 outputs a reconstructed image by deblocking filtering the current block reconstructed by the second differential unit 150. As a result, the boundary error between blocks in the reconstructed image is reduced, thereby smoothing the block boundary of the reconstructed image.

The loop filter 170 filters the reconstructed image output from the deblocking unit 160 using a plurality of filter structures. Here, the filter structure includes an adaptive filter structure and a fixed filter structure, the adaptive filter structure includes at least one of a central symmetry structure, a vertical symmetry structure, a horizontal symmetry structure, and a user-defined symmetry structure, the fixed filter structure The filter coefficient represents a filter structure in which the encoding structure and the decoding device already know the filter structure as a predetermined filter structure.

A method of filtering the reconstructed image by the loop filter 170 will be described with reference to FIG. 2 below. In FIG. 2, a method of loop filtering a reconstructed image will be described by taking an example in which the plurality of filter structures are the adaptive filter structure.

 According to FIG. 2, the loop filter 170 includes a calculator 171 and a determiner 173, and the loop filter 170 filters the reconstructed image.

The calculator 171 calculates filter coefficients for each of a plurality of preset filter structures. That is, the calculation unit 171 calculates a center symmetric filter coefficient, a vertical symmetric filter coefficient, a horizontal symmetric filter coefficient, and a user-defined symmetric filter coefficient. Here, since the content of calculating the filter coefficient based on the filter structure is a well known technique, a detailed description thereof will be omitted.

The determiner 173 filters the reconstructed image input from the deblocking unit 160 using the filter coefficient calculated by the calculator 171. The determiner 173 determines the final filter structure by using the coding rate of the filtered reconstructed image. At this time, the determination unit 173 filters the reconstructed image for each frame to determine the final filter structure for each frame.

Specifically, the determination unit 173 calculates a coding rate of the plurality of reconstructed images by using each of the plurality of filtered reconstructed images and the original image, and a filter structure of the reconstructed image having the highest coding rate among the calculated coding rates; The filter coefficients are determined by the final filter structure and the final filter coefficient, respectively.

That is, the determination unit 173 filters the reconstructed image using each of the center symmetry, the vertical symmetry, the horizontal symmetry, and the user-defined symmetric filter coefficients, and filters the center image, the vertical symmetry, the horizontal symmetry, and the user-defined symmetry structure. The coding rate of the reconstructed image is calculated, and the final filter structure is determined using the filter structure of the reconstructed image having the highest encoding rate of the reconstructed image.

Here, an example of the coding rate may be a degradation degree of the reconstructed image, and the determination unit 173 compares the resolution of the reconstructed image and the original image (that is, the current frame which is the input image before hatching). The deterioration degree of can be calculated. In addition, the determiner 173 may determine the final filter structure J by applying a rate-distortion optimization according to Equation 1 below to the filtered reconstructed image and the original image.

Where D is the distortion of the entire frame, R is the total amount of bits needed to encode the frame, and R _Filter is the amount of bits needed to encode the filter coefficients.

Also, the determination unit 173 outputs the filter structure information including the index indicating the final filter structure and the final filter coefficient to the entropy encoder 140. At this time, the determination unit 173 sets the filter structure information according to the final filter structure based on Table 1 below, and outputs the filter structure information to the entropy encoder 140.

Final filter structure index Center Symmetrical Structure 0 Vertical Symmetry Structure One Horizontal symmetry structure 2 Custom Symmetry Structure 3

That is, the determination unit 173 sets the filter structure information to 0 if the final filter structure is the center symmetric structure, the filter structure information to 1 if the final filter structure is the vertical symmetry structure, and the filter if the final filter structure is the horizontal symmetry structure. If the structure information is 2 and the final filter structure is a user-defined symmetry structure, the filter structure information can be set to 3.

The entropy encoder 140 entropy-codes the final filter coefficients and the filter structure information for each frame, inserts them into the bitstream, and transmits the final filter coefficients and the filter structure information to the decoding apparatus 400. At this time, if the final filter structure is a user-defined symmetric structure, the final filter coefficient and the filter structure information corresponding to the user-defined symmetric structure is entropy coded, such as SPS (Sequence Parameter Set), PPS (picture Parameter Set), new NAL unit type, SEI It may be included in a bitstream structure of a message or a slice header.

3 is a flowchart provided to explain an encoding method using a plurality of filter structures according to an embodiment of the present invention. In FIG. 3, a method of loop filtering a reconstructed image will be described by taking a case where the plurality of filter structures are the adaptive filter structure as an example.

First, the calculation unit 171 calculates a filter coefficient for each of the plurality of filter structures (310). Here, the filter structure includes an adaptive filter structure and a fixed filter structure, the adaptive filter structure includes at least one of a central symmetry structure, a vertical symmetry structure, a horizontal symmetry structure, and a user-defined symmetry structure, the fixed filter structure The filter coefficient represents a filter structure in which the encoding structure and the decoding device already know the filter structure as a predetermined filter structure.

Next, the determiner 173 filters the reconstructed image using the calculated plurality of filter coefficients (320). That is, the determination unit 173 filters the reconstructed image by using the central symmetric filter coefficient, filters the reconstructed image by using the vertical symmetric filter coefficient, filters the reconstructed image by using the horizontal symmetric filter coefficient, and selects a user-defined symmetric filter coefficient. Filter the reconstructed image.

The determination unit 173 calculates a coding rate of each filtered reconstructed image (330).

In detail, in operation S320, the determiner 173 calculates a coding rate of each of the reconstructed images filtered using the center, vertical, horizontal, and user-defined symmetric filter coefficients, and each of the reconstructed images filtered using the original image. Here, an example of the coding rate may include a degree of degradation of the reconstructed image, and the determination unit 173 may calculate the coding rate of the reconstructed image by using various techniques for calculating a well-known encoding rate.

Next, the determination unit 173 determines a final filter structure using the calculated coding rates of the plurality of reconstructed images (340). That is, the determination unit 173 determines the filter structure and the filter coefficient of the reconstructed image having the highest coding rate among the calculated reconstruction images, respectively, as the final filter structure and the final filter coefficient. At this time, the determination unit 173 sets the filter structure information according to the aforementioned Table 1 based on the final filter structure.

The entropy encoder 140 entropy encodes the final filter coefficients and the filter structure information for each frame (350), inserts them into the bitstream, and transmits them to the decoding apparatus 400 (360).

At this time, if the final filter structure is a user-defined symmetric structure, the entropy encoder 140 entropy-encodes the final filter coefficient and the filter structure information corresponding to the user-defined symmetric structure SPS (Sequence Parameter Set), PPS (picture Parameter Set) It may be included in a bitstream structure of a new NAL unit type, an SEI message, or a slice header.

In the above, the operation of loop filtering the reconstructed image when the plurality of filter structures is the adaptive filter structure has been described. However, when the plurality of filter structures are the fixed filter structure, the calculation unit 171 may be omitted in FIG. 2. In FIG. 3, S310 may be omitted.

Specifically, if the plurality of filter structures are fixed filter structures, the filter coefficients corresponding to each of the plurality of fixed filter structures are preset, and the loop filter 170 may calculate filter coefficients corresponding to each of the plurality of fixed filter structures. no need. Accordingly, the loop filter 170 loop-filters each reconstructed image using filter coefficients corresponding to each of a plurality of preset fixed filter structures, and calculates a coding rate of each filtered reconstructed image to determine a final filter structure. have.

Similarly, when a plurality of filter structures are mixed with an adaptive filter structure and a fixed filter structure, a filter coefficient corresponding to the adaptive filter structure is calculated to loop-filter the reconstructed image, and a filter coefficient corresponding to the preset fixed filter structure. The final filter structure may be determined by loop filtering the reconstructed image and calculating a coding rate of each filtered reconstructed image.

In this case, if the determined final filter structure is a fixed filter structure, the encoding apparatus 100 entropy encodes the filter structure information corresponding to the final filter structure and transmits the entropy encoding to the decoding apparatus 400. That is, if the final filter structure is a fixed filter structure, the encoding apparatus 100 does not transmit the final filter coefficient corresponding to the final filter structure to the decoding apparatus 400.

4 is a block diagram showing a configuration of a decoding apparatus according to an embodiment of the present invention.

Referring to FIG. 4, the decoding apparatus 400 according to an embodiment of the present invention includes an entropy decoding unit 410, an inverse quantization and inverse DCT unit 420, a motion compensator 430, an adder 435, And a deblocking unit 440 and a loop filter 450.

The entropy decoder 410 entropy decodes the bit stream to extract transform coefficients and motion vector filter structure information. If the received bit stream includes the final filter coefficient, the entropy decoder 410 entropy decodes the bit stream. To extract the final filter coefficients. That is, the entropy decoding unit 410 decodes at least one of the encoded final filter coefficients and the filter structure information received from the encoding apparatus 100.

The inverse quantization and inverse DCT unit 420 dequantizes the extracted transform coefficients and inverse discrete cosine transform to obtain a differential block.

The motion compensator 430 generates a prediction block of the current block by using the motion vector extracted by the entropy decoder 410.

The adder 435 reconstructs the current block by adding the difference block obtained by the inverse quantization and inverse DCT unit 420 and the prediction block generated by the motion compensation unit 430.

The deblocking unit 440 outputs a reconstructed image by deblocking filtering the current block reconstructed by the adder 435.

The loop filter 450 loop-filters the reconstructed image using at least one of the final filter coefficient and the filter structure information extracted by the entropy decoder 410. In this case, the loop filter 450 filters the reconstructed image by determining which of the plurality of preset filter structures is extracted filter structure information.

Specifically, for example, when the filter structure information is set to 1 indicating a central symmetric structure that is an adaptive filter structure, the loop filter 450 filters the reconstructed image into the central symmetric structure by using the final filter coefficient. Here, the central symmetry structure represents the final filter structure determined by the encoding apparatus 100.

Similarly, when the filter structure information is set as a vertical, horizontal, or user-defined symmetric structure that is an adaptive filter structure, the loop filter 450 uses the final filter coefficient extracted by the entropy decoding unit 410 to deblocking unit ( The reconstructed image input from 440 is filtered into a vertical, horizontal, or user-defined symmetric structure. Through this, the error generated by the discrete cosine transform, quantization, etc. may be attenuated to output the filtered image data close to the original image.

In this case, for example, when the filter structure information is set to an index indicating the fixed filter structure, the loop filter 450 loop-filters the reconstructed image with the fixed filter structure using a filter coefficient corresponding to the fixed filter structure. Here, the fixed filter structure represents a final filter structure determined by the encoding apparatus 100, and the filter coefficient corresponding to the fixed filter structure is a predetermined value already known to the encoding apparatus 100 and the decoding apparatus 400. In this way, the decoding apparatus 400 may loop filter the reconstructed image with the filter structure loop-filtered by the encoding apparatus 100 without receiving the final filter coefficient from the encoding apparatus 100.

5 is a flowchart provided to explain a decoding method using a plurality of filter structures according to an embodiment of the present invention. Hereinafter, the case where the filter structure information indicates the adaptive filter structure will be described as an example.

First, the entropy decoding unit 410 receives a bitstream (510) and entropy decodes the received bitstream to extract filter structure information (520). At this time, if the received bit stream includes the final filter coefficient, the entropy decoder 410 extracts the final filter coefficient by entropy decoding the bit stream.

Subsequently, when it is determined that the filter structure information is set to the central symmetric structure (YES at 530), the loop filter 440 filters the reconstructed image to the central symmetric structure using the extracted final filter coefficients (540).

In this case, if it is determined that the filter structure information is not set to the central symmetry structure (NO: 530), the loop filter 440 determines whether the filter structure information is set to the vertical symmetry structure (550).

If it is determined that the filter structure information is set to the vertical symmetry structure (YES at 550), the loop filter 440 filters the reconstructed image to the vertical symmetry structure using the extracted final filter coefficients (560).

In this case, if it is determined that the filter structure information is not set to the vertical symmetry structure (NO: 550), the loop filter 440 determines whether the filter structure information is set to the horizontal symmetry structure (570).

If it is determined that the filter structure information is set to the horizontal symmetry structure (570: YES), the loop filter 440 filters the reconstructed image to the horizontal symmetry structure using the extracted final filter coefficient (580).

In this case, if it is determined that the filter structure information is not set to the horizontal symmetry structure (570: NO), the loop filter 440 filters the reconstructed image into a user-defined symmetric structure using the extracted final filter coefficients (590).

In FIG. 5, the case in which the filter structure information indicates the adaptive filter structure has been described as an example. However, when the decoded filter structure information indicates the fixed filter structure, the loop filter 440 uses a filter coefficient corresponding to the fixed filter structure. Loop-filter the reconstructed image with a fixed filter structure. Here, the filter coefficients corresponding to the fixed filter structure are preset and known by the encoding apparatus 100 and the decoding apparatus 400.

Until now, the filter structure used in the encoding / decoding apparatus and method has been described as being a symmetric structure based on center, vertical, horizontal, and user definition, but the present invention is not necessarily limited thereto. , And at least one of a user-defined asymmetric structure.

In addition, while the vertical, horizontal, center, user-defined symmetrical and asymmetrical filter structures have been described above, other filter structures may be used in the present encoding / decoding apparatus and method to determine final filter structures and final filter coefficients.

In the above description, the indices are set to 0 to 3 based on Table 1, but the indexes are not limited thereto, and the indices may include bits, octets, bytes, and the like according to a filter structure supported by the encoding and decoding apparatus. It may be preset in units.

In addition, while the content of calculating the deterioration degree of the reconstructed image by comparing the resolution of the reconstructed image and the original image has been described, the present invention is not limited thereto. The structure can be used to determine the final filter structure of the current frame. That is, in the present encoding / decoding apparatus and method, the final filter structure can be determined without using the original image.

In addition, although the above description has been made based on the image, the image may be divided into blocks and the present invention may be implemented based on each block. That is, each block can be used to determine the same or different filter. Therefore, a plurality of filters may be used in one image.

In addition, in the above description, the final filter structure is determined after loop filtering the reconstructed image. However, the loop filter 170 may determine the final filter structure before loop filtering the reconstructed image.

That is, the determiner 173 may determine the final filter structure by using the statistical characteristics of the reconstructed image or the distribution of pixel values.

In this way, the decoding apparatus 400 may infer and obtain the final filter structure of the reconstructed image, which is loop filtered by the encoding apparatus, from the images that are already decoded. Here, since inferring the final filter structure from the decoded images is a well known technique, a detailed description thereof will be omitted.

In addition, in the above description, the filter structure applied to the reconstructed image having the lowest deterioration has been described as the final filter structure. However, the calculation unit 171 calculates each calculation complexity for loop filtering each of the reconstructed images with a plurality of filter structures. The calculator 173 may determine the filter structure having the lowest computational complexity among the calculated computational complexity as the final filter structure.

As described above, the present invention has been described by way of limited embodiments and drawings, but the present invention is not limited to the above embodiments, and those skilled in the art to which the present invention pertains various modifications and variations from such descriptions. This is possible.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the claims below but also by the equivalents of the claims.

1 is a block diagram showing a configuration of an encoding apparatus according to an embodiment of the present invention.

2 is a block diagram showing the configuration of a loop filter according to an embodiment of the present invention.

3 is a flowchart provided to explain an encoding method using a plurality of filter structures according to an embodiment of the present invention.

4 is a block diagram showing a configuration of a decoding apparatus according to an embodiment of the present invention.

5 is a flowchart provided to explain a decoding method using a plurality of filter structures according to an embodiment of the present invention.

<Explanation of symbols for the main parts of the drawings>

105: intra prediction unit

110: motion estimation unit

120: motion compensation unit

125, 155: first difference part, second difference part

130: DCT and quantization unit

140: entropy encoder

150: dequantization and inverse DCT

160: deblocking unit

170: loop filter

171: calculating unit

173: Decision

Claims (20)

Filtering the reconstructed image by each of the plurality of filter structures; And Determining one of the plurality of filter structures as a final filter structure using the filtered plurality of reconstructed images Encoding method comprising a. The method of claim 1, The plurality of filter structures include at least one of an adaptive filter structure and a fixed filter structure, The adaptive filter structure includes at least one of a central, vertical, horizontal, and user defined symmetry structure, And said fixed filter structure represents a filter structure in which a filter coefficient is preset. The method of claim 2, The user-defined symmetry structure, represents a user-defined filter symmetry structure, And the user-defined symmetric structure is included in a bitstream structure of a sequence parameter set (SPS), a picture parameter set (PPS), a new NAL unit type, an SEI message, or a slice header. The method of claim 1, The filtering step, Calculating a filter coefficient corresponding to each of the plurality of filter structures if the plurality of filter structures are the adaptive filter structures; And Filtering the reconstructed image using each of the calculated filter coefficients. Encoding method comprising a. The method of claim 1, The determining step, The final filter structure is determined for each frame, Encoding filter structure information corresponding to the determined final filter structure for each frame Encoding method comprising a. The method of claim 5, The encoding step, If the determined final filter structure is an adaptive filter structure, the final filter coefficients and filter structure information corresponding to the determined final filter structure are encoded. And if the determined final filter structure is a fixed filter structure, encoding filter structure information corresponding to the determined final filter structure. The method of claim 1, The determining step, And a coding rate of each of the filtered plurality of reconstructed images, and determining a filter structure of the reconstructed image having the high encoded rate as the final filter structure. The method of claim 1, The filtering step, And encoding the looped image of the reconstructed image. A loop filter for filtering the reconstructed image by each of the plurality of filter structures, and determining one of the plurality of filter structures as the final filter structure by using the filtered plurality of reconstructed images; And An entropy encoder that encodes filter structure information corresponding to the determined final filter structure. Encoding apparatus comprising a. 10. The method of claim 9, The loop filter, A calculation unit for calculating a filter coefficient corresponding to each of the plurality of filter structures if the plurality of filter structures are an adaptive filter structure; And A determination unit for filtering the reconstructed image using each of the calculated filter coefficients, and determining one of the plurality of filter structures as a final filter structure using the filtered plurality of reconstructed images. Encoding apparatus comprising a. 10. The method of claim 9, The loop filter, If the plurality of filter structures is a fixed filter structure, the encoding apparatus for filtering the reconstructed image by each of the predetermined filter coefficients, and determines any one of the plurality of filter structures as the final filter structure using the filtered plurality of reconstructed image. . 10. The method of claim 9, The entropy coding unit, If the determined final filter structure is an adaptive filter structure, the final filter coefficients and filter structure information corresponding to the determined final filter structure are encoded. And if the determined final filter structure is a fixed filter structure, encoding filter structure information corresponding to the determined final filter structure. 10. The method of claim 9, The plurality of filter structures include at least one of an adaptive filter structure and a fixed filter structure, The adaptive filter structure includes at least one of a central, vertical, horizontal, and user defined symmetry structure, And the fixed filter structure is a filter structure in which a filter coefficient is preset. Receiving filter structure information encoded for each frame; Obtaining a final filter structure used in encoding from the received filter structure information; And Filtering the reconstructed image by the obtained final filter structure. Decryption method comprising a. The method of claim 14, The filter structure information is an index indicating a final filter structure used for filtering the reconstructed image, The final filter structure comprises at least one of an adaptive filter structure and a fixed filter structure, The adaptive filter structure includes at least one of a central, vertical, horizontal, and user defined symmetry structure, The fixed filter structure is a decoding method showing a filter structure with a predetermined filter coefficient. The method of claim 15, Decoding the received filter structure information More, The filtering step, And decoding the reconstructed image using the decoded filter structure information. The method of claim 16, The receiving step, If the final filter structure is the adaptive filter structure, the coded final filter coefficient corresponding to the final filter structure and filter structure information corresponding to the final filter structure are received. The decoding step, Decoding the received final filter coefficient and the filter structure information; The filtering step, And decoding the reconstructed image by the final filter structure using the decoded final filter coefficients. An entropy decoder which receives the encoded filter structure information for each frame; And A loop filter for obtaining a final filter structure used for encoding from the received filter structure information and loop filtering the reconstructed image with the obtained final filter structure Decoding apparatus comprising a. The method of claim 18, The entropy decoding unit, If the final filter structure is an adaptive filter structure, the encoded final filter coefficient corresponding to the final filter structure and the filter structure information corresponding to the final filter structure are received and decoded. The loop filter, And if the final filter structure is an adaptive filter structure based on the decoded filter structure information, loop-filtering the reconstructed image with the final filter structure using the final filter coefficient. The method of claim 18, The entropy decoding unit, If the final filter structure is a fixed filter structure, receiving and decoding filter structure information corresponding to the final filter structure, The loop filter, And if the final filter structure is a fixed filter structure, loop filtering the reconstructed image by a filter coefficient corresponding to the preset fixed filter structure.

Images