JP2008109642A - Video decoding apparatus and method - Google Patents

Abstract

It is possible to prevent a post filter process having a strong removal strength from being performed on a moving region.
A video decoding apparatus that decodes a bitstream encoded by switching between field encoding and frame encoding in units of areas constituting an interlaced picture, and a post filter for removing noise A filter processing unit 302 that performs processing, confirms the encoding type of the target region, and is included in a picture temporally continuous to the current picture, and the encoding type of the region located at the same position as the target region, or , Depending on whether the coding type confirmation unit 301 that confirms the coding type of a region spatially adjacent to the target region and the confirmed coding type is frame coding or field coding, A noise removal strength determination unit 305 that determines a noise removal strength, and the filter processing unit 302 applies the noise removal strength to the target region. The post-filter processing in accordance with to do.
[Selection] Figure 8

Description

  The present invention relates to a video decoding apparatus and method for decoding a compressed bit stream, and more particularly to a post filter that reduces image quality fluctuations in the time direction.

  Applications that distribute video via an IP (Internet Protocol) network and applications that record video on storage media such as a DVD (Digital Versatile Disc) / BD (Blu-ray Disc) are widely used. In these applications, in order to fit within a predetermined transmission band or media capacity, information is compressed and encoded using video encoding technology.

  Video coding is done by removing the spatial and temporal redundancy contained in the video. Spatial redundancy is removed by intra-picture prediction encoding that uses a correlation between pixels and encodes a difference between a pixel value predicted from a decoded adjacent pixel value and a target pixel value. On the other hand, temporal redundancy is removed by inter-picture predictive coding that uses the correlation between pictures and encodes the difference between a predicted picture generated by predicting motion from a decoded picture and the target picture. The

  As shown in FIG. 1, the encoded video is composed of a plurality of picture groups called GOP (Group Of Pictures). The first I picture of the GOP is encoded by intra-picture prediction encoding, and the picture following the I picture is encoded by inter-picture prediction encoding.

  Note that when the target video is in the interlace format, either field coding or frame coding can be selected. As shown in FIG. 2, field coding is a method of coding a top field and a bottom field as independent pictures. Frame coding is a method of coding in frame units in which the top field and the bottom field are combined.

  In general, frame coding is suitable for a scene where the frame correlation is stronger than the field correlation, that is, a scene with little motion. Field coding is suitable for scenes where field correlation is stronger than frame correlation, that is, scenes with motion. Therefore, when encoding the target video, frame encoding is easily selected in an area with little motion, and field encoding is easily selected in an area with much motion.

  Also, switching between field coding and frame coding can be performed not in units of pictures but in units of blocks called macroblocks. In this case, as shown in FIG. 3, two upper and lower macroblocks are handled as a pair. As for coding efficiency, frame coding is suitable for macroblocks with little motion, and field coding is suitable for macroblocks with motion, as in the case of picture units. In order to increase the coding efficiency, it is necessary to control these switching well.

  However, video with good coding efficiency is not always excellent. As a main example, there is coding noise called pulsing noise (Non-Patent Document 1). This pulsing noise is caused by the image quality of the I picture.

  As described above, an I picture is encoded in a closed picture. For this reason, since the I picture has poor image quality continuity between pictures, image quality fluctuations in the time direction occur in the I picture that is the head of the GOP, resulting in visual degradation. In particular, in the case where encoding is performed with all I pictures, which may occur in applications requiring random accessibility such as video search / editing, image quality fluctuations occur in all pictures. For this reason, flicker noise that is more noticeably deteriorated than pulsing noise occurs (Patent Document 1, Patent Document 2, and Non-Patent Document 2).

  As described above, it is necessary to establish an encoding method not only for encoding efficiency but also for subjective quality. However, in the case of the above pulsing noise and flicker noise, a method for completely reducing the coding side has not been established at present.

  Therefore, it is necessary to reduce this encoding noise with a post filter on the decoding side. By executing the post-filter process, the image quality continuity in the time direction can be increased, and the above-described noise can be reduced.

  Specifically, the frame index is t, the output picture is o (t), the decoded picture is p (t), and the noise removal strength is W (0 is weak and 1 is strong). The smoothing filter represented by the following formula 1 is applied.

  o (t) = (1−W) × p (t) + W × o (t−1) (Formula 1)

  At this time, if the noise removal strength W is increased in a portion where there is a motion, an afterimage is generated. Therefore, it is necessary to determine the noise removal strength W according to the presence or absence of motion.

  However, in order to detect the presence or absence of motion, it is necessary to calculate differences between pictures and the like, and the amount of processing becomes a problem. Therefore, when the input is a compressed bit stream, a method using a motion vector obtained in the decoding process has been proposed (Patent Document 3). However, the I picture that is the main cause of image quality fluctuation does not have a motion vector. That is, a method using a motion vector cannot be applied to an I picture.

Here, a method for detecting a motion of a picture by referring to an encoding type has been proposed (Patent Document 4). As described above, field coding is likely to be selected as the coding type for the region with motion, and frame coding is likely to be chosen as the coding type for the region with little motion. That is, by detecting the coding type, it is possible to detect a region with motion and a region with little motion. Therefore, using this method, it is also conceivable to selectively remove noise from an area with motion and an area with little motion for an I picture having no motion vector.
JP 2005-323315 A Japanese Patent Laid-Open No. 2005-20771 JP 2002-171424 A JP 2004-104171 A PCSJ2005 draft "Method of reducing I picture flicker by quantization control with detent" PCSJ2005 draft "H.264 / MPEG-4 AVC intra-coded frame flicker reduction method"

  However, the technique described in Patent Document 4 has a problem that the relationship between the encoding type and the presence or absence of motion is not absolute. In other words, frame coding is not always selected for coding because it is a non-motion area. Similarly, field coding is not necessarily chosen just because it is a region with motion. Therefore, even if the post-filter process with increased noise removal strength is executed just because the encoding type is frame encoding, there is a possibility that the post-filter process may be executed up to an area with motion. In this case, an afterimage is generated in a moving area, and the subjective image quality is deteriorated.

  Accordingly, the present invention has been made in view of the above-described problems, and an image that prevents post-filter processing with strong removal strength from being performed on an area with movement by improving the accuracy of determination of the presence or absence of movement. It is an object to provide a decoding apparatus and method.

  The video decoding apparatus according to the present invention includes a field encoding that encodes a top field and a bottom field independently in units of regions that constitute an interlaced picture, and a frame unit that combines the top field and the bottom field. A video decoding apparatus for decoding a coded bitstream by switching between frame encoding to be performed, a filter processing means for performing post-filter processing for removing noise of a decoded picture, and the post-filter The coding type of the region located in the same position as the target region, which is included in a picture temporally continuous with the current picture including the target region, and confirming the coding type of the target region to be processed Or an encoding type confirmation that confirms the encoding type of an area spatially adjacent to the target area. And a noise removal strength determination for determining a noise removal strength of the post-filter processing according to whether the coding type confirmed by the coding type confirmation device is frame coding or field coding Means for performing post-filtering on the target area in accordance with the noise removal strength.

  As a result, it is possible to accurately determine the presence or absence of movement of the target region by checking the coding type not only for the target region to be subjected to the noise removal process but also for other regions. Therefore, it is possible to prevent the post filter process having a strong removal strength from being performed on a moving region.

  Further, the coding type confirmation unit confirms the coding type of a region located in the same position as the target region, which is included in a plurality of pictures temporally continuous from the current picture, and the noise removal strength determination unit Is a picture that is temporally continuous from the current picture when the coding type of the target area is frame coding, and the coding type confirmed by the coding type confirmation unit is frame coding As the number of pictures including the region is larger, the noise removal strength of the post filter processing may be increased.

  Thereby, the strength of the noise removal filter can be increased as the number of pictures including a region whose coding type is frame coding continues. That is, the strength of the noise removal filter can be set to be strong for an area that is more likely to have no movement. Therefore, it is possible to remove the conspicuous noise in an area where there is almost no movement while protecting the movement of the input image as much as possible.

  Further, the coding type confirmation unit confirms coding types of a plurality of regions spatially adjacent to the target region, and the noise removal strength determination unit determines that the coding type of the target region is frame coding. In some cases, the noise removal strength of the post-filter processing may be increased as the number of regions in which the coding type confirmed by the coding type confirmation unit is frame coding increases.

  Thereby, the strength of the noise removal filter can be increased as there are more blocks whose encoding type is frame encoding. That is, the strength of the noise removal filter can be set to be strong for an area that is more likely to have no movement. Therefore, it is possible to remove the conspicuous noise in an area where there is almost no movement while protecting the movement of the input image as much as possible.

  The noise removal strength determining means sets the noise removal strength to 0 when the quantization parameter of the target region is smaller than a predetermined threshold, and the filter processing means sets the noise removal strength to 0. In some cases, the post filter processing may not be performed on the target region.

  Thereby, when the quantization parameter is equal to or smaller than a certain threshold value, it is not necessary to confirm the temporal continuity and the spatial correlation, and the processing amount can be reduced. Further, since the quantization parameter is a parameter obtained by decoding the bit stream, a new processing amount for detecting the quantization parameter is not required.

  Further, the noise removal strength determining means determines the noise removal strength so that when the quantization parameter is equal to or greater than the threshold value, the noise removal strength of the post filter processing increases as the quantization parameter increases. May be.

  Thereby, the value of the noise removal strength can be adaptively determined according to the value of the quantization parameter. Since the conspicuousness of noise varies depending on the bit rate, it is possible to prevent image quality from deteriorating by performing a filtering process having an unnecessary strength.

  The video decoding apparatus and method of the present invention suppress the occurrence of afterimages by accurately determining the presence or absence of motion and preventing post-filter processing with strong removal strength on a region with motion. Can do. Furthermore, the post filter processing can be performed with a low processing amount, and can also be applied to an I picture having no motion vector.

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

  FIG. 4 is a diagram illustrating functional blocks of the video decoding apparatus 100 according to the present embodiment.

  The video decoding apparatus 100 according to the present embodiment decodes a compressed bit stream, removes encoding noise, and outputs the result as an output frame. In this embodiment, decoding processing, noise removal strength determination processing, and post filter processing are performed in units of blocks as shown in FIG.

  The video decoding apparatus 100 includes, as components, a video decoding unit 200 that decodes a bit stream and a noise removal unit 300 that removes coding noise.

  FIG. 5 is a diagram illustrating functional blocks of the video decoding unit 200 configuring the video decoding device 100.

  The video decoding unit 200 decodes the compressed bit stream, and transfers the decoded frame, the encoding type, and the quantization parameter to the noise removing unit 300.

  Here, the encoding type is one of encoding parameters indicating whether an area to be encoded is encoded by a field or a frame. Video compression standard In H.264 / AVC, as shown in FIG. 6, it is described in a parameter called mb_field_decoding_flag and can be specified in units of macroblocks.

  The quantization parameter is H.264. In the case of H.264 / AVC, the value is 0 to 51, and the closer to 0, the higher the bit rate. If the region is encoded at a high bit rate, the noise is not noticeable, so that smoothing processing in the time direction is unnecessary.

  The video decoding unit 200 includes, as components, a variable length decoding unit 201, an inverse quantization unit 202, an inverse orthogonal transform unit 203, a decoded frame storage memory 204, a motion compensation unit 205, and an adder 206. With.

  The variable length decoding unit 201 is a processing unit that performs variable length decoding of a bitstream. The variable length decoding unit 201 decodes the bit stream into encoding parameters such as quantization parameters, motion vectors, and encoding types, orthogonal transform coefficients, and the like. Parameters related to motion compensation such as motion vectors are transferred to the motion compensation unit 205. Parameters relating to quantization, such as quantization parameters and orthogonal transform coefficients, are transferred to the inverse quantization unit 202. As for the coding type, as shown in FIG. 7, the data is collectively transferred to the noise removing unit 300 in a form associated with the macroblock address. Similarly, the quantization parameter is transferred to the noise removing unit 300.

  The inverse quantization unit 202 is a processing unit that inversely quantizes the orthogonal transform coefficient. The inverse quantization unit 202 performs inverse quantization on the orthogonal transform coefficient using the quantization parameter transferred from the variable length decoding unit 201 and the quantized orthogonal transform coefficient. Then, the inversely quantized orthogonal transform coefficient is transferred to the inverse orthogonal transform unit 203.

  The inverse orthogonal transform unit 203 is a processing unit that performs inverse orthogonal transform on the inversely quantized orthogonal transform coefficient. The inverse orthogonal transform unit 203 transfers the prediction error obtained by performing inverse orthogonal transform on the orthogonal transform coefficient transferred from the inverse quantization unit 202 to the adder 206.

  The decoded frame storage memory 204 is a storage unit that temporarily stores decoded frames.

  The motion compensation unit 205 is a processing unit that performs motion compensation using a reference frame. The motion compensation unit 205 obtains a decoded frame from the decoded frame storage memory 204, performs motion compensation using the motion vector and prediction mode transferred from the variable length decoding unit 201, and generates a prediction frame To do. The generated prediction frame is transferred to the adder 206.

  The adder 206 adds the prediction frame and the prediction error. A decoded frame obtained by adding the prediction error transferred from the inverse orthogonal transform unit 203 and the predicted frame transferred from the motion compensation unit 205 is transferred to the noise removing unit 300.

  FIG. 8 is a diagram illustrating functional blocks of the noise removal unit 300 that configures the video decoding device 100.

  The noise removing unit 300 reduces the coding noise of the decoded frame transferred from the video decoding unit 200 in accordance with the coding type, and then outputs it as an output frame.

  The noise removal unit 300 includes a coding type confirmation unit 301, a filter processing unit 302, an output frame storage memory 303, a coding type storage memory 304, and a noise removal strength determination unit 305 as components.

  The coding type confirmation unit 301 confirms the coding type of the target area to be subjected to post-filter processing, and includes the target area included in a picture temporally continuous with the current picture including the target area. This is a processing unit for confirming the coding type of a region located at the same position or the coding type of a region spatially adjacent to the target region. In the present embodiment, the encoding type of the target block transferred from the variable length decoding unit 201 is confirmed. Further, it is stored in the coding type storage memory 304 and is included in a temporally continuous frame from the current picture, and the coding type of the block at the same position as the target block is confirmed. Further, the coding type of a block spatially adjacent to the target block is confirmed. Then, the information regarding the coding type of each frame is transferred to the noise removal strength determination unit 305. Also, the encoding type of the target block is stored in the encoding type storage memory 304. Here, the target block is a block to be subjected to post filter processing.

  The filter processing unit 302 is a processing unit that performs post filter processing to remove noise. As the noise removal strength W increases, the filter processing unit 302 increases smoothing and improves image quality continuity in the time direction. Decoding is performed using the previous output frame stored in the output frame storage memory 303, the decoded frame transferred from the video decoding unit 200, and the noise removal strength transferred from the noise removal strength determination unit 305. A filter that increases the continuity in the time direction is applied to the target block of the frame. Then, a copy of the decoded frame after filtering is created, one of which is transferred to the output frame storage memory 303, and the other is transferred as an output frame to an external device such as a display.

  The output frame storage memory 303 is a storage unit that stores the output frame transferred from the filter processing unit 302.

  The encoding type storage memory 304 is a storage unit that temporarily stores the encoding type transferred from the video decoding unit 200.

  The noise removal strength determination unit 305 determines the noise removal strength of the post filter processing according to whether the coding type confirmed by the coding type confirmation unit 301 is frame coding or field coding. Is a processing unit. In the present embodiment, the noise removal strength W is determined depending on the number of coding types confirmed by the coding type confirmation unit 301 being frame coding. The noise removal strength W is determined by the following equation 2.

  W = Wq × (a × Wt + (1−a) × Ws) (Expression 2)

  Here, Wq is an intensity determined from the quantization parameter of the target block. Wt is the strength determined from the temporal continuity of the frame-coded block. Ws is an intensity determined from the number of blocks that are frame-coded among the blocks adjacent to the target block. a is a parameter that determines the specific gravity of Wt and Ws. Each value ranges from 0 to 1. When the noise removal strength W is 0, the filter processing unit 302 does not perform post filter processing on the target block.

FIG. 9 is a diagram showing the relationship between the quantization step Qs and the intensity Wq, with the horizontal axis representing the quantization step Qs corresponding to the quantization parameter Qp and the vertical axis representing the intensity Wq. When the quantization step Qs is close to 0, that is, when encoding is performed at a high bit rate, noise is not conspicuous, so that smoothing processing in the time direction is unnecessary. For this reason, as shown in FIG. 9, when the quantization parameter Qp is equal to or smaller than a certain threshold Th ₁ , that is, when the corresponding quantization step Qs is equal to or smaller than the threshold Th ₂ , the intensity Wq = 0 is set. Thus, post-filter processing can be avoided.

When the quantization parameter Qp is larger than the threshold Th ₁ , that is, when the corresponding quantization step Qs is larger than the threshold Th ₂ , the intensity Wq is increased as the quantization step Qs increases. Thereby, the noise removal strength W can be determined in consideration of the bit rate of encoding, and unnecessary post filter processing can be prevented from being performed.

  FIG. 10 is a diagram illustrating an example of calculation of Wt. Wt is expressed by Equation 3 below.

  Wt = Ct ÷ Nt (Formula 3)

  Here, Ct is a number in which pictures including blocks whose encoding type is frame encoding are temporally continuous from the current picture. Nt is the number of blocks that the encoding type confirmation unit 301 confirms. In the example of FIG. 10, since Ct = 1 and Nt = 3, Wt = 0.33. Note that if the encoding type of the target block is field encoding, Wt = 0.

  As described above, the strength Wt increases as the number of pictures including blocks whose encoding type is frame encoding continues in time. The fact that many pictures including blocks that are frame-coded are continuous in time indicates that there is a high probability that the coding type of the target block is also frame-coded. As a result, the post filter process can be more reliably executed on a portion with little movement.

  FIG. 11 is a diagram illustrating an example of calculating Ws. Ws is expressed by Equation 4 below.

  Ws = Cs ÷ Ns (Formula 4)

  Here, Cs is the number of blocks whose coding type is frame coding among the blocks confirmed by the coding type confirmation unit 301. Ns is the number of blocks that the encoding type confirmation unit 301 confirms. In the example of FIG. 11, since Cs = 3 and Ns = 4, Ws = 0.75. Note that when the encoding type of the target block is field encoding, Ws = 0.

  As described above, the intensity Ws increases as the number of blocks whose encoding type is frame encoding spatially increases and is adjacent to the target block. The fact that there are many spatially adjacent blocks whose encoding type is frame encoding indicates that there is a high probability that the encoding type of the target block is also frame encoding. As a result, the post filter process can be more reliably executed on a portion with little movement.

  The blocks to be confirmed may be not only the four blocks adjacent to the target block but also the eight blocks around the target block.

  The coefficient a shown in Expression 2 is a weighting coefficient that determines whether priority is given to temporal continuity or spatial adjacency. That is, by setting a = 0, it is possible to determine the noise removal strength taking into account only temporal continuity. Conversely, by setting a = 1, it is possible to determine the noise removal strength taking into consideration only the spatial adjacency. Thus, it is not always necessary to confirm both temporal continuity and spatial adjacency, and only one of them may be confirmed.

  This completes the description of the functional blocks of the video decoding apparatus 100 of the present invention.

  Hereinafter, the operation of the video decoding apparatus 100 of the present invention will be described using the flowchart shown in FIG.

  The flowchart shown in FIG. 12 is stored as a control program in a storage device (for example, a ROM (Read Only Memory) or a flash memory) that is not shown in the video decoding device 100, and is also not shown in the CPU. (Central Processing Unit).

  Also, the flowchart shown in FIG. 12 particularly corresponds to the operation related to the noise removing unit 300, and the description related to the video decoding unit 200 is omitted.

  The following operations are performed for each macroblock.

The noise removal strength determination unit 305 confirms the quantization parameter Qp (S11). That is, it is determined whether greater than the threshold Th ₁ with the quantization parameter Qp.

When the quantization parameter Qp is larger than the threshold Th ₁ (Yes in S11), the noise removal strength determination unit 305 sets the strength Wq using the graph shown in FIG. 9 (S12). When the quantization parameter Qp is equal to or less than the threshold Th ₁ (No in S11), Wq = 0 is set (S13).

  Next, the coding type confirmation unit 301 confirms the coding type of the target block (S14). That is, it is determined whether or not the encoding type of the target block is frame encoding. In the case of frame encoding (Yes in S14), the noise removal strength determination unit 305 calculates the values of Wt and Ws using the above-described Expression 3 and Expression 4 (S15). When it is not frame coding (No in S14), the noise removal strength determination unit 305 sets Wt = Ws = 0 (S16).

  Details of the Wt and Ws setting process (S15) are shown in FIGS. 12B and 12C. Note that either the Wt setting process (FIG. 12B) or the Ws setting process (FIG. 12C) may precede. Or it is good also as a structure which performs only any one.

  FIG. 12B is a flowchart showing an operation when determining the strength Wt.

  First, the coding type confirmation unit 301 confirms the coding type of a block located at the same position as the target block in a temporally continuous picture from the current picture including the target block (S21). Next, the noise removal strength determination unit 305 counts the number of temporally consecutive pictures including blocks whose encoding type is frame encoding (S22). And intensity | strength Wt is determined using Formula 3 (S23).

  FIG. 12C is a flowchart illustrating an operation when determining the strength Ws.

  First, the coding type confirmation unit 301 confirms the coding type of a block spatially adjacent to the target block (S31). Next, the noise removal strength determination unit 305 counts the number of blocks whose encoding type is frame encoding (S32). And intensity | strength Ws is determined using Formula 4 (S33).

  As described above, when Wq, Wt, and Ws are determined, the noise removal strength determination unit 305 sets the noise removal strength W using Equation 2 (S17). Finally, the filter processing unit 302 generates an output block o (t) using the current decoding block i (t), the immediately preceding output block o (t−1), and the noise removal strength W. (S18).

  FIG. 13 is a diagram illustrating an example of processing related to determination of the strength Wt among the above operations. In the example shown in the figure, the encoding types of the current picture and the previous picture are confirmed (Nt = 1). At this time, only the block corresponding to the address 116 has frame coding as the coding type of both the current picture and the previous picture. That is, Ct = 1. Therefore, for the block corresponding to the address 116, the strength Wt = 1 (indicated as “strong” in FIG. 13).

  Similarly, FIG. 14 is a diagram spatially showing an example of processing related to the determination of the strength Wt in the above operation. For a region where blocks whose encoding type is frame encoding are temporally continuous, the filter processing unit 302 enhances smoothing in the time direction. For other regions, smoothing is performed weakly in the time direction, or smoothing is not performed in the time direction.

  FIG. 15 is a diagram illustrating an example of processing related to determination of the strength Ws among the above operations. In the example shown in the figure, encoding types of a target block to be subjected to noise removal and four blocks adjacent to the target block are confirmed (Nt = 4). Here, it is assumed that the strength Ws is “strong” when the coding type of the target block and the coding type of two or more blocks among the surrounding four blocks are frame coding.

  FIG. 16 is a diagram spatially illustrating the example of FIG. In the case where the encoding type of the target block is frame encoding and the encoding type of two or more blocks among the four blocks adjacent to the target block is frame encoding, the filter processing unit 302 moves in the time direction. Perform strong smoothing. For other regions, smoothing is performed weakly in the time direction, or smoothing is not performed in the time direction.

  As described above, according to the video decoding apparatus and method of the present embodiment, when blocks whose encoding type is frame encoding are temporally continuous or spatially adjacent, The noise removal strength can be increased. As a result, it is possible to smooth the portion where it is certain that there is no movement in the time direction. That is, it is possible to prevent an afterimage from being generated in an image by performing a smoothing process on a region having movement.

  The video decoding apparatus and method according to the present invention have been described based on the embodiment. However, the present invention is not limited to this embodiment. Unless it deviates from the meaning of this invention, the form which carried out the various deformation | transformation which those skilled in the art can think to this embodiment, and the structure constructed | assembled combining the component in different embodiment is also contained in the scope of the present invention. .

  For example, the decoding process, the noise removal strength determination process, and the post filter process are performed in units of blocks as shown in FIG. 3, but may be performed in units of larger blocks or frames.

  In the video decoding apparatus and method according to the present embodiment, the quantization parameter Qp is used, but it is not always necessary. For example, in the case where it is clearly determined that the encoding is performed at a low bit rate, by setting a value corresponding to Wq as a fixed value, it is possible to reduce processing relating to calculation of Wq.

  In addition, the steps of the video decoding method of the present invention can be realized as a program for causing a computer to execute, or can be realized as a recording medium such as a computer-readable CD-ROM on which the program is recorded, and information and data indicating the program Alternatively, it can be realized as a signal. These programs, information, data, and signals may be distributed via a communication network such as the Internet.

  Further, some or all of the components constituting the video decoding device may be configured by a single system LSI (Large Scale Integration). The system LSI is an ultra-multifunctional LSI manufactured by integrating a plurality of components on a single chip, and specifically includes a microprocessor, a ROM, a RAM (Random Access Memory), and the like. Computer system. A computer program is stored in the RAM. The system LSI achieves its functions by the microprocessor operating according to the computer program.

  The video decoding apparatus and method of the present invention have a function of decoding a compression-encoded bitstream and further reducing encoding noise, such as BS (Broadcasting Satellite) broadcasting, terrestrial digital broadcasting, etc. Alternatively, it is useful for a recorder or the like that accumulates and plays back surveillance camera images transmitted via an IP network.

It is a figure which shows GOP structure. It is a figure which shows the picture structure in a field encoding. It is a figure which shows the macroblock which can switch field coding and frame coding. It is a functional block diagram regarding the video decoding apparatus of this invention. It is a functional block diagram regarding the video decoding part of the video decoding apparatus of this invention. H. 2 is a diagram illustrating an encoding parameter indicating an H.264 / AVC encoding type. FIG. It is a figure which shows the relationship between a macroblock address and an encoding type. It is a functional block diagram regarding the noise removal part of the video decoding apparatus of this invention. It is a figure which shows the relationship between quantization step Qs and intensity | strength Wq. It is a figure which shows an example of calculation of intensity | strength Wt. It is a figure which shows an example of calculation of intensity | strength Ws. It is a flowchart which shows operation | movement of the video decoding apparatus of this invention. It is a flowchart which shows the operation | movement of the process regarding determination of the intensity | strength Wt. It is a flowchart which shows the operation | movement of the process regarding determination of the intensity | strength Ws. It is a figure which shows an example of the process regarding determination of the intensity | strength Wt. It is a figure which shows spatially the process regarding the determination of intensity | strength Wt. It is a figure which shows an example of the process regarding determination of the intensity | strength Ws. It is a figure which shows the process regarding the determination of intensity | strength Ws spatially.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Video decoding apparatus 200 Video decoding part 201 Variable length decoding part 202 Inverse quantization part 203 Inverse orthogonal transformation part 204 Decoding frame preservation | save memory 205 Motion compensation part 206 Adder 300 Noise removal part 301 Encoding type confirmation part 302 Filter processing unit 303 Output frame storage memory 304 Encoding type storage memory 305 Noise removal strength determination unit

Claims (13)

Switching between field coding that encodes the top field and bottom field independently and frame coding that encodes the top field and bottom field in units of areas that make up an interlaced picture. A video decoding device for decoding an encoded bitstream,
Filter processing means for performing post-filter processing for removing noise of the decoded picture;
Check the encoding type of the target area to be subjected to the post-filter processing, and include the area located in the same position as the target area, which is included in a picture temporally continuous with the current picture including the target area. An encoding type or an encoding type confirmation means for confirming an encoding type of an area spatially adjacent to the target area;
Noise removal strength determination means for determining the noise removal strength of the post-filter processing according to whether the coding type confirmed by the coding type confirmation means is frame coding or field coding; Prepared,
The filter processing means includes
A video decoding apparatus, wherein post-filter processing corresponding to the noise removal strength is performed on the target region. The encoding type confirmation unit confirms an encoding type of an area located at the same position as the target area, included in a plurality of pictures temporally continuous from the current picture,
When the coding type of the target region is frame coding, the noise removal strength determination unit is a picture temporally continuous from the current picture, and the code confirmed by the coding type confirmation unit The video decoding apparatus according to claim 1, wherein the noise removal strength of the post filter processing is increased as the number of pictures including a region whose encoding type is frame encoding is larger. The noise removal strength determining means is the noise removal strength proportional to the number of pictures including a region that is temporally continuous from the current picture and whose coding type confirmed by the coding type confirmation means is frame coding. The video decoding device according to claim 2, wherein the video decoding device is determined. The encoding type confirmation unit confirms the encoding types of a plurality of regions spatially adjacent to the target region,
When the coding type of the target area is frame coding, the noise removal strength determination means has a larger number of areas where the coding type confirmed by the coding type confirmation means is frame coding. The video decoding apparatus according to claim 1, wherein a noise removal strength of the post filter processing is increased. The noise removal strength determining means determines the noise removal strength proportional to the number of regions whose coding type confirmed by the coding type confirmation means is frame coding. Video decoding device. The noise removal strength determining means sets the noise removal strength to 0 when the quantization parameter of the target region is smaller than a predetermined threshold value,
The video decoding device according to claim 1, wherein the filter processing unit does not perform the post-filter processing on the target region when the noise removal strength is zero. The noise removal strength determination means determines the noise removal strength so that the noise removal strength of the post-filter processing increases as the quantization parameter increases when the quantization parameter is equal to or greater than the threshold. The video decoding apparatus according to claim 6, wherein: The noise removal strength determining means determines the noise removal strength proportional to a quantization step corresponding to the quantization parameter when the quantization parameter is equal to or greater than the threshold. Video decoding device. The video decoding apparatus according to claim 1, wherein the filter processing unit smoothes the target region more strongly in the time direction as the noise removal strength increases as the post-filter processing. The video decoding apparatus according to claim 1, wherein the filter processing unit performs the post-filter processing using a region obtained by combining two adjacent macroblocks as the target region. Switching between field coding that encodes the top field and bottom field independently and frame coding that encodes the top field and bottom field in units of areas that make up an interlaced picture. A video decoding method for decoding an encoded bitstream, comprising:
A filtering step for performing post-filtering to remove noise in the decoded picture;
Check the encoding type of the target area to be subjected to the post-filter processing, and include the area located in the same position as the target area, which is included in a picture temporally continuous with the current picture including the target area. An encoding type or an encoding type confirmation step for confirming an encoding type of an area spatially adjacent to the target area;
A noise removal strength determining step for determining a noise removal strength of the post filter processing according to whether the coding type confirmed in the coding type confirmation step is frame coding or field coding. Including
In the filtering step,
A video decoding method, wherein post-filter processing corresponding to the noise removal strength is performed on the target region. Switch between field coding, which encodes the top field and bottom field independently, and frame coding, which encodes the top field and bottom field, in units of regions that make up an interlaced picture. A program for a video decoding device for decoding an encoded bitstream,
A filtering step for performing post-filtering to remove noise in the decoded picture;
Check the encoding type of the target area to be subjected to the post-filter processing, and include the area located in the same position as the target area, which is included in a picture temporally continuous with the current picture including the target area. An encoding type or an encoding type confirmation step for confirming an encoding type of an area spatially adjacent to the target area;
A noise removal strength determining step for determining a noise removal strength of the post filter processing according to whether the coding type confirmed in the coding type confirmation step is frame coding or field coding. Let the computer run,
In the filtering step,
A program that causes a computer to execute post-filter processing corresponding to the noise removal strength on the target region. An integrated circuit for removing noise in a decoded picture,
Filter processing means for performing post-filter processing for removing noise of the decoded picture;
Check the encoding type of the target area to be subjected to the post-filter processing, and include the area located in the same position as the target area, which is included in a picture temporally continuous with the current picture including the target area. An encoding type or an encoding type confirmation means for confirming an encoding type of an area spatially adjacent to the target area;
Noise removal strength determination means for determining the noise removal strength of the post-filter processing according to whether the coding type confirmed by the coding type confirmation means is frame coding or field coding; Prepared,
The filter processing means includes
An integrated circuit, wherein post-filter processing corresponding to the noise removal strength is performed on the target region.

Images