KR100710338B1 - Multi-format video decoder and deblocking filtering method - Google Patents

Abstract

The present invention relates to a multi-format video decoder and a deblocking filtering method for reconstructing a video compressed by various compression coding methods. In particular, the present invention applies a deblocking filter used to reconstruct a video of the H.264 standard to a process of reconstructing a video of the MPEG-2 standard, thereby improving the image quality of the reconstructed video by removing block artifacts of the MPEG-2 video. It is effective to let.

Deblocking Filter, MPEG-2, H.264, Boundary Conditions

Description

Multi-format video decoder and deblocking filtering method

1 is a block diagram illustrating an embodiment of a multi-format video decoder according to the present invention;

2A and 2B illustrate examples of inter-block vertical boundary pixels and horizontal boundary pixels.

3 is a view showing an example of determination of boundary conditions of a deblocking filter according to the present invention.

4 illustrates an example of mapping MPEG-2 quantization coefficients to H.264 quantization coefficients according to the present invention.

Explanation of symbols for main parts of the drawings

100: frame memory 110: decoding control unit

111: entropy decoder 112: IS / IQ

113: inverse transform unit 114: intra prediction unit

115: motion compensation unit 116: switching unit

117: adder 118: deblocking control unit

119: deblocking filter 120121: selection unit

The present invention relates to a video decoder for reconstructing a compressed video, and more particularly, to a method for deblocking filtering of video data compressed and encoded by MPEG-2 compression algorithm.

Video compression methods have been developed in various ways depending on the application. The basic framework of the video compression method is to divide the continuous video frames constituting the video into blocks of a certain size and remove spatial and temporal redundancy according to each block position. In other words, in order to remove spatial redundancy, a method of converting an image frame into a frequency domain in units of blocks divided within the same image frame and quantizing the same to reduce the transmission burden on high frequency components, and to remove temporal redundancy. It uses motion estimation and compensation technique between video frames. Widely used video compression is represented by the Moving Picture Expert Group (MPEG), which is classified into several standards according to the needs of the times and applications.

MPEG-1 has been widely used as a method of compressing a video by an early sequential scanning method and storing it in a storage medium such as a VCD (Video Compact Disk) or compressing a small video.

MPEG-2 used a variety of techniques by adding interlaced video processing to compress high quality video. The MPEG-2 is most widely used in various fields such as storage devices such as DVD (Digital Versatile Disk), DTV broadcasting such as satellite, cable, and terrestrial wave, personal video recorder (PVR), and video transmission on a network. One of the video standards used.

MPEG-4 is a compression standard used for video streaming service or wireless video service in a network environment by applying new techniques such as moving picture, still picture, graphic picture, and object model.

In order to support more complex and diverse applications and to facilitate the creation and distribution of video contents, the need for a new video compression standard with higher compression efficiency has been raised after the MPEG-4 video compression standard. In order to meet these demands, new compression methods such as H.264 and VC-1 (Video Coding-1) have been proposed.

H.264 is also called MPEG-4 Part 10-AVC (Advanced Video Coding), and it can be said that H.26L and MPEG-4 part 2 Visual have evolved. H.264 is an efficient video compression technique that reduces the amount of compressed bits by half while maintaining the same image quality as compared to MPEG-2.

As such, video compression methods are developed in various ways, and there is a constant demand for new methods. In addition, as the mass storage of digital storage media, the simplification and the high quality and diversification of digital TV services, and the interconnection of various media through the network, the necessity of restoring various types of compressed videos in one receiver is emerging.

Taking a digital TV receiver as an example, it is essential and basic to restore a MPEG-2 video, which is a basic broadcast screen, and display it on the screen. In this case, in order to provide various services to viewers through data broadcasting of digital TV, it is often necessary to restore a still image or to restore a video by another type of compression technique. In addition, the receiver may be connected to the Internet to perform web browsing or to download and restore a video.

Various video compression standards commonly use the above-described frequency transform, quantization techniques, and basic coding techniques of motion estimation and motion compensation. Therefore, one video decoder can be adaptively designed to cope with various video compression standards.

On the other hand, in order to compressively encode a video, image frames are divided into a predetermined size and processed. As a side effect, deterioration in image quality of blocking artifacts and ringing artifacts occurs.

The blocking artifact is a phenomenon in which a block shape is visualized due to a difference in quantization values between neighboring blocks at divided block boundaries. That is, the blocking artifacts are generated by encoding using different quantization coefficients in units of blocks during encoding. The ring artifact is a phenomenon in which a substantial boundary of an object is crushed as a high frequency component is removed by quantization.

Accordingly, an object of the present invention is a multi-format video decoder and deblocking filtering that improves the quality of a reconstructed image by removing block artifacts of a video encoded by various methods in a multi-format video decoder capable of decoding various video compression standards. In providing a method.

Another object of the present invention is to apply a deblocking filter used to reconstruct a video of the H.264 standard to the process of reconstructing a video of the MPEG-2 standard, thereby removing block artifacts of the video encoded and encoded by the MPEG-2 method. A multi-format video decoder and a deblocking filtering method are provided.

In order to achieve the above object, the multi-format video decoder according to the present invention includes a reconstruction unit for reconstructing the video signal received by compression-encoding by at least one compression encoding method; A deblocking filter configured to deblock the filtered image and output the deblocked filter for display; A frame memory for storing a reference image to be used for restoring the image signal of the restoration unit; And a decoding controller configured to select one of an output image of the decompression unit and an output image of the deblocking filter according to a compression encoding scheme of the decompressed image, and store the decoded image as a reference image in the frame memory. Characterized in that comprises a.

The decoding control unit selects and stores the output image of the reconstruction unit in the frame memory when the compression encoding scheme is MPEG-2.

The multi-format video decoder further includes a deblocking controller, wherein the deblocking controller is configured to determine the filtering strength of the deblocking filter according to a boundary condition and a compression encoding scheme of a unit block consisting of nxn pixels in a received video signal. It features.

The boundary condition of the unit block may vary depending on whether or not the compression coding scheme is MPEG-2.

When the compression coding scheme is MPEG-2, the deblocking control unit determines whether at least a macroblock is boundary among the MPEG-2 encoding information, a macroblock type between adjacent unit blocks, an encoding type of an 8x8 block, and a coded block pattern (CBP). The boundary strength of the unit block is determined using the value of), and the larger the boundary strength value of the unit block is, the stronger the deblocking filtering is.

The deblocking filtering method according to the present invention comprises the steps of: reconstructing an image signal received by compression encoding by at least one compression encoding method; Restoring the compressed and input video signal by using a corresponding decoding method; Deblocking filtering the restored image and outputting the filtered image for display; And storing the reconstructed image as it is, or deblocking filtering and storing the reconstructed image according to an encoding scheme of the reconstructed image when the reconstructed image is stored as a reference image.

Other objects, features and advantages of the present invention will become apparent from the following detailed description of embodiments taken in conjunction with the accompanying drawings.

Hereinafter, with reference to the accompanying drawings, preferred embodiments of the present invention that can specifically realize the above object will be described. At this time, the configuration and operation of the present invention shown in the drawings and described by it will be described as at least one embodiment, by which the technical spirit of the present invention and its core configuration and operation is not limited. .

The present invention will be described as an embodiment to apply to a multi-format video decoder that adaptively reconstructs video data compressed by various methods in accordance with H.264, VC-1, MPEG-2, and the like.

In particular, the present invention applies a deblocking filtering process required to restore a video of the H.264 standard or the VC-1 standard to a process of restoring the video of the MPEG-2 standard, and thus, block artifacts of the video compressed and encoded by the MPEG-2 method. To remove it.

This is because the MPEG-2 standard does not include a deblocking filter. That is, in the case of a video decoder capable of reconstructing video of various standards, the deblocking filter used in the H.264 standard does not operate in the process of reconstructing an MPEG-2 video. According to the present invention, the idle deblocking filter is used to reconstruct an MPEG-2 video to improve the quality of the reconstructed video.

1 is a detailed block diagram illustrating an embodiment of a multi-format video decoder capable of reconstructing a video of H.264 or MPEG-2 standard according to the present invention, and includes a frame memory 100, a decoding control unit 110, and entropy. Decoder 111, Inverse Scan / Inverse Quantizer 112, Inverse Transformer 113, Intra Prediction Unit 114, Motion Compensator (or Inter Predictor) 115, Switcher 116, Adder 117, the deblocking control unit 116, the deblocking filter 119, the first selector 120, and the second selector 121.

That is, a VES (Video Elementary Stream) bitstream corresponding to a program selected by a system decoder or a transport demux (not shown) is input to the entropy decoder 111.

The bitstream is an encoded format of video data and is encoded in an entropy encoding format. Accordingly, the entropy decoder 111 decodes each syntax element encoded with a variable length code (VLC) or a context adaptive binary arithmetic code (CABAC) and provides the decoded block to a corresponding block.

Each syntax information includes picture header information, prediction information, and residual information.

The picture header information indicates information such as a picture structure, a picture coding type, a frame size, a reference frame, and the like. The information is transmitted to the decoding control unit 110 to provide other blocks. It is used to control.

The prediction information is an intra prediction mode, which is information used for compensating a prediction value in a block unit, within a same frame, or a reference frame, which is inter frame prediction information, which compensates a prediction value in a block unit from another frame. It includes an index (Reference frame index) and a motion vector. The intra prediction mode is provided to the intra predictor 114, and the reference frame index and the motion vector are provided to the motion compensator 115.

The difference information is expressed in the form of a coefficient obtained by converting a difference value compensated for an intra or inter prediction value into a frequency domain, and is provided to the inverse scan / inverse quantization unit 112.

The inverse scan / inverse quantization unit 112 performs inverse quantization while scanning the input transform coefficients that are zigzag-scanned in an inverse scan, that is, a raster scan method, and outputs the inverse quantization unit 113 to the inverse transform unit 113. The inverse transform unit 113 inversely transforms an inverse quantized transform coefficient (IDCT) and outputs the inverse discrete cosine transform (IDCT) to the adder 117. The adder 117 reconstructs the complete block by adding the IDCT signal and the prediction value output through the switching unit 116 and then outputs the deblocking filter 119 and the first and second selection units 120 and 121.

Meanwhile, when the macro block type to be decoded is an intra mode, the intra prediction unit 114 receives an intra prediction mode signal from the entropy decoder 111 and compensates the prediction value of the current block from adjacent pixel values in the same frame. Output to the switching unit 116.

When the macroblock type to be decoded is the inter mode, the motion compensator 115 receives a reference frame index and a motion vector from the entropy decoder 111 to calculate a corresponding position of the reference frame to compensate for the prediction value. Output to the switching unit 116.

The switching unit 116 selects an output of the intra predictor 114 according to the predicted pixel value, or selects an output of the motion compensator 115 and outputs the output to the adder 117.

The switching of the switching unit 116 is performed by the decoding control unit 110.

The present invention can adaptively decode a video of the H.264 standard and the MPEG-2 standard, and adaptive decoding of main decoding function blocks according to the video standard is performed by the decoding control unit 110.

The decoded pixel, which is completely reconstructed by the adder 117, is output for display and stored in the frame memory 100 when the image frame currently being reconstructed is later used as a reference frame and is not used as a reference frame. If not stored in the frame memory 100 is output only for display.

In this case, the image stored in the frame memory 100 may be stored through the deblocking filter 119 or may be directly stored without passing through the deblocking filter 119. To this end, the first selector 120 selects the output of the adder 117 under the control of the decoding controller 110, or selects the output of the deblocking filter 119 and outputs the output to the frame memory 100.

In addition, the image output from the adder 117 for display may be output through the deblocking filter 119 or may be directly output without passing through the deblocking filter 119. To this end, the second selector 121 selects the output of the adder 117 under the control of the decoding controller 110 or selects the output of the deblocking filter 119 and outputs the output to the display. According to an embodiment of the present invention, the reconstructed image output for display irrespective of the compression encoding scheme is unconditionally passed through the deblocking filter 119 and the second selector 121.

The decoding controller 110 controls the first and second selectors 120 and 121 according to whether the reconstructed image is the H.264 standard or the MPEG-2 standard.

That is, in case of H.264 standard video reconstruction, the first and second selectors 120 and 121 select the output of the deblocking filter 119. In other words, in the case of H.264 standard video reconstruction, the image reconstructed by the adder 117 always passes through the deblocking filter 119 regardless of whether it is used as a reference frame.

In case of MPEG-2 standard video reconstruction, the first selector 120 selects the output of the adder 117 and outputs the output to the frame memory 100, and the second selector 121 deblocks the output. The output of filter 119 is selected and then output for display.

This removes the block artifacts of the reconstructed image to be displayed.

That is, in the case of video reconstruction of the H.264 standard, the image reconstructed by the adder 117 passes through the deblocking filter 119 regardless of whether the image reconstructed is used as a reference frame.

On the contrary, in case of MPEG-2 standard video reconstruction, when the image reconstructed by the adder 117 is used as a reference frame, the image is stored in the frame memory 100 without passing through the deblocking filter 119 and the display is simultaneously displayed. In order to pass through the deblocking filter 119. That is, when decoding the MPEG-2 video, the deblocking filter 119 is used as a post processing means for the decoded image frame, and the image has passed through the deblocking filter 119. The frame is output for display. In this case, the reference frame stored in the frame memory 100 is an image frame that does not pass through the deblocking filter 119.

In the case of H.264, the deblocking filter is included in the encoder of the transmitting side, and the difference between the deblocking filtered signal and the original video signal is transmitted by compression encoding, so that the output of the deblocking filter 119 is output to the frame memory ( 100, but in the case of MPEG-2, since the deblocking filter is not included in the transmitter, the reference frame stored in the frame memory 100 should not pass through the deblocking filter 119.

Meanwhile, the deblocking control unit 118 determines whether the deblocking filter 119 employs a filter and a filter strength using the macroblock information decoded by the entropy decoder 111.

In one embodiment, the deblocking control unit 118 adjusts whether the deblocking filter 119 employs the filter and the strength of the filter according to the boundary condition of the unit block. Here, the size of the unit block is 4 × 4 pixels. At this time, the boundary condition of the unit block depends on whether it is H, 264 or MPEG-2.

That is, the deblocking control unit 118 checks boundary strength based on encoding information of a block boundary to apply a deblocking filter to distinguish block artifacts caused by block-based quantization and an actual boundary of an object. And filter strength.

2A and 2B show pixels of a block boundary with respect to a horizontal boundary and a vertical boundary.

3 illustrates an example of boundary strength (bS) for determining the filtering strength of a unit block.

The bS of the deblocking filter 119 employed in the H.264 video standard is determined by comparing macroblock types, differences in reference frames, presence of difference values, motion vectors, and the like among adjacent unit blocks.

For example, when bS is 4, filtering is most strongly applied. After passing through the deblocking filter 119, pixel values of the boundary pixels p0, p1, p2, q0, q1, and q2 of FIG. 2 are changed. Can be. The boundary condition for bS = 4 is a case where a boundary between unit blocks is a macroblock boundary, two macroblocks are of the same frame / field coding type, and one of the blocks is a macroblock encoded by an intra type.

When bS has a value of 1, 2, and 3, after passing through the deblocking filter 119, pixel values of the boundary pixels p0, p1, q0, and q1 of FIG. 2 may be changed. In this case, the boundary condition where bS = 3 is a case where a boundary between unit blocks is a macroblock boundary, two macroblocks are different frame / field encoding types, and one of the blocks is a macroblock encoded by an intra type. The boundary condition where bS = 2 is when one or all of the unit blocks are differentially coded. The boundary condition to be bS = 1 is when the motion vector difference between unit blocks is 1 or more or motion compensation is performed from different reference frames. In other cases than the above, 0 is substituted in bS so as not to go through the deblocking filter 119.

On the other hand, the bS of the deblocking filter 119 employed in the MPEG-2 moving picture standard also has a macroblock boundary in the MPEG-2 coded information and a macroblock type, 8x8 block between adjacent unit blocks. Is determined using encoding information such as a coding type and a coded block pattern (CBP). That is, the encoding information for determining the boundary condition of the deblocking filter 119 is applied in place of the MPEG-2 encoding information in consideration of the decision variable of H.264.

In this case, the boundary condition where bS = 4 is when the boundary between unit blocks is a macroblock boundary and one or both macroblock types of two macroblocks are encoded as an intra macroblock. The boundary condition where bS = 3 is not applicable. This is because MPEG-2 does not adopt frame / field coding for each macroblock. The boundary condition where bS = 2 is a case where CBP information of one or all of the boundary blocks is set to one.

In the case of MPEG-2, the boundary condition for bS = 1 is that when the motion vector is constant in the macroblock, the boundary between the unit blocks is the macroblock boundary and the following condition is satisfied.

First, the difference between the motion vectors of two adjacent macro blocks is greater than or equal to one.

Second, when the picture structure currently being decoded is a frame, a motion compensation type is different between adjacent blocks, or one of them has a field motion compensation type.

Third, when a picture structure currently being decoded is a field, a motion compensation type is different between adjacent blocks, or one of them has a 16x8 motion compensation type.

The deblocking control unit 118 determines whether the deblocking filter 119 is to be applied and is used as a factor for limiting the value of the filter output according to the boundary strength bS inferred from a quantization parameter. Parameters.

That is, factors influencing the filtering result value according to the boundary condition bS of the deblocking filter depend on the quantization coefficients.

At this time, in the case of H.264, each factor is selected according to the quantization coefficient, and the range of the quantization coefficient is 0 to 51 as shown in FIG. On the other hand, the range of the quantization coefficient of MPEG-2 has a value of 0-31. Therefore, in order to use the H.264 deblocking filter for MPEG-2 video, it is necessary to change the quantization coefficient. That is, the quantization coefficients of MPEG-2 should be mapped to the quantization coefficients of H.264.

4 shows an example of changing a quantization coefficient between video standards. Referring to FIG. 4, it can be seen that the quantization coefficients of MPEG-2 are mapped in close proximity to intermediate values of the quantization coefficients of H.264.

This is just one embodiment, and the range of change between each value may vary depending on the application.

On the other hand, the terms used in the present invention (terminology) are terms defined in consideration of the functions in the present invention may vary according to the intention or practice of those skilled in the art, the definitions are the overall contents of the present invention It should be based on.

The present invention is not limited to the above-described embodiments, and as can be seen in the appended claims, modifications can be made by those skilled in the art to which the invention pertains, and such modifications are within the scope of the present invention.

According to the multi-format video decoder and the deblocking filtering method according to the present invention described above, the MPEG-standard is applied by applying a deblocking filter used to reconstruct an H.264 standard video to a MPEG-2 standard video reconstruction. 2 The block artifacts of the video are removed to improve the quality of the reconstructed video.

Those skilled in the art will appreciate that various changes and modifications can be made without departing from the spirit of the present invention.

Therefore, the technical scope of the present invention should not be limited to the contents described in the embodiments, but should be defined by the claims.

Claims (13)

A reconstruction unit for reconstructing an image signal received by compression encoding by at least one compression encoding method; A deblocking filter configured to deblock the filtered image and output the deblocked filter for display; A frame memory for storing a reference image to be used for restoring the image signal of the restoration unit; And When the reconstructed unit stores the reconstructed image as a reference image, the decoding controller selects one of an output image of the decompression unit and an output image of the deblocking filter according to the compression encoding scheme of the reconstructed image and stores the decoded image in the frame memory. And a multi-format video decoder. The method of claim 1, wherein the decoding control unit And if the compression encoding method is MPEG-2, selects an output image of the reconstruction unit and stores the output image in a frame memory. The method of claim 1, And a deblocking control unit configured to determine a filtering strength of the deblocking filter according to a boundary condition of a unit block consisting of nxn pixels and a compression encoding scheme in a received video signal. The method of claim 3, wherein The boundary condition of the unit block is different depending on whether or not the compression coding scheme is MPEG-2. The method of claim 4, wherein the deblocking control unit If the compression coding method is MPEG-2, the MPEG-2 coded information is determined using at least a macro block boundary, a macroblock type, an 8x8 block encoding type, and a coded block pattern (CBP) value between adjacent unit blocks. The boundary strength of the unit block is determined, and the larger the boundary strength value of the unit block, the more strongly the deblocking filtering is applied. The method of claim 5, The boundary condition where the boundary strength (bS) = 4 is when the boundary between unit blocks is a macroblock boundary and the macroblock type of one or both macroblocks is encoded as an intra macroblock, The boundary condition to be bS = 2 is when the CBP information of one or both of the boundary blocks is set to 1, The boundary condition to be bS = 1 is that if the boundary between the unit blocks is a macroblock boundary and the difference between the motion vectors of at least two adjacent macroblocks is greater than or equal to 1, and the picture structure being decoded is a frame, When the motion compensation type is different or one of them has a field motion compensation type, when the picture structure being decoded is a field, the motion compensation type is different between adjacent blocks or one of the two has a 16x8 motion compensation type. Multi-format video decoder. Restoring an image signal that is received by compression encoding by at least one compression encoding method; Restoring the compressed and input video signal by using a corresponding decoding method; Deblocking filtering the restored image and outputting the filtered image for display; And And storing the reconstructed image as it is, or deblocking filtering and storing the reconstructed image according to an encoding method of the reconstructed image when the reconstructed image is stored as a reference image. 8. The method of claim 7, wherein said storing step If the encoding scheme is MPEG-2, the deblocking filtering method stores the reconstructed image as a reference image without deblocking filtering. The method of claim 7, wherein And determining the filtering intensity of the deblocking filtering step according to the boundary condition of the unit block consisting of nxn pixels and the compression encoding scheme in the received image signal. The method of claim 9, The boundary condition of the unit block is different depending on whether the compression coding scheme is MPEG-2 or not. The method of claim 9, If the compression coding method is MPEG-2, the MPEG-2 coded information is determined using at least a macro block boundary, a macroblock type, an 8x8 block encoding type, and a coded block pattern (CBP) value between adjacent unit blocks. The deblocking filtering method determines the boundary strength of the unit block, and the deblocking filtering is more strongly applied as the boundary strength value of the unit block is larger. The method of claim 11, wherein And mapping the quantization coefficients of MPEG-2 to the H.264 quantization coefficients to reconstruct and deblock the filtered image encoded by the MPEG-2 scheme. The method of claim 12, wherein said mapping step And a finer mapping of the difference between the quantization coefficients of MPEG-2 in the vicinity of an intermediate value of the quantization coefficients of H.264.

Images