CN104243985B - Sample Adaptive Compensation in HEVC - Google Patents

Abstract

Improvements to sample adaptive compensation in HEVC. First, additional boundary compensation is performed for boundary pixels of each coding unit in the largest coding unit. Second, add a new mode to edge compensation (EO) types 2 and 3: the pixel value difference between the current pixel and the left adjacent pixel is 1, and the difference between the right adjacent pixel and the right adjacent pixel is more than 3, and the current pixel and the left adjacent pixel The pixel value of is different by 2, and the difference from the adjacent pixel on the right is more than 6. Finally, the cases where the difference between the current pixel and the neighboring pixels is less than or equal to 1 in types 2 and 3 are eliminated.

Description

Sample Adaptive Compensation in HEVC

joint research

This application is jointly researched by North China University of Technology and Information Institute of Beijing Jiaotong University, and supported by the following funds: National Natural Science Foundation of China (No.60903066, No.60972085), Beijing Natural Science Foundation of China (No.4102049), New Teacher of the Ministry of Education Fund (No.20090009120006), and Beijing Municipal Higher Education Intensification Plan for Talents (PHR201008187).

technical field

The present invention relates to the field of image processing, and more particularly to optimized sample adaptive compensation in HEVC.

Background technique

In April 2010, the two major international video coding standard organizations VCEG and MPEG established a video compression joint group JCT-VC (Joint collaborative Team on Video Coding) to jointly develop the High Efficiency Video Coding HEVC (High efficiency video coding) standard, which is also known as H .265. The main goal of the HEVC standard is to achieve a substantial improvement in coding efficiency compared with the previous generation standard H.264/AVC, especially for high-resolution video sequences. Its goal is to reduce the bit rate to 50% of the H.264 standard at the same video quality (PSNR).

At the current stage, HEVC still uses the hybrid coding framework adopted by H.264. Inter- and intra-frame predictive coding: De-correlation between temporal and spatial domains. Transform coding: Transform coding is performed on the residual to remove spatial correlation. Entropy coding: remove statistical redundancy. HEVC will focus on researching new coding tools or technologies within the framework of hybrid coding to improve video compression efficiency.

At present, many new coding features that have been proposed in the discussion organized by JCT-VC may be added to the HEVC standard. The specific documents of each discussion can be obtained from http://wftp3.itu.int .

The HEVC (High Efficiency Video Coding) standard was officially released as an international standard in January 2013. It is called H.265 in ITU-T, and it is called the second part of MPEG-H in ISO/IEC. At the same time, in order to enable the HEVC standard to support more application scenarios, follow-up work has been carried out, including support for high resolution and color formats, scalable coding, and 3-D/stereoscopic/multi-view coding.

However, HEVC is still evolving. Sample adaptive compensation is a new technology in HEVC, which adds fresh blood to HEVC. However, because this technology is new, there are still some deficiencies and needs to be supplemented, as described below.

In the present invention, we propose two optimization algorithms for sample adaptive compensation. From the experimental results, we can see that these two algorithms have good effects.

In the description of the present invention, the following documents are cited, which are hereby incorporated as part of the disclosure content of the present invention.

[1].G.J.Sullivan, J.R.Ohm, W.J.Han and T.Wiegand.Overview of the HighEfficiency Video Coding(HEVC)Standard, IEEE Transactions on Circuits and Systems for Video Technology, Vol.22, Issue 12, 2012, pp.1649- 1668.

[2].J.Zhu, D.Zhou, G.He, S.Goto.A combined SAO and de-blocking filterarchitecture for HEVC video decoder, in Proceedings of the Conference on '20th IEEE International Conference on Image Processing (ICIP)' , Melbourne, VIC, 15-18Sept.2013, pp.1967-1971.

[3].P.N.Subramanya, R.Adireddy, D.Anand.SAO in CTU decoding loop for HEVC video decoder, in Proceedings of the Conference on'International Conference on Signal Processing and Communication (ICSC)', Noida, 12-14Dec.2013, pp.507-511.

[4].G.B.Praveen, G.B.Praveen.Analysis and approximation of SAOestimation for CTU-level HEVC encoder, in Proceedings of the Conference on'Visual Communications and Image Processing (VCIP)', Kuching, 17-20Nov.2013, pp.1 -5.

[5]. W. Kim, J. Sole, Marta. Offset Scaling in SAO for High Bit-depth Video Coding, in Proceedings of the Conference on '13th Joint Collaborative Team on Video Coding (JCT-VC) Meeting', Incheon, KR, 18-26Apr.2013, JCTVC-M0335.

[6]. G. Laroche, T. Poirier, C. Gisquet, E. P. Onno. On Inter-Layer SAO for Base mode, in Proceedings of the Conference on '12th Joint Collaborative Team on Video Coding (JCT-VC) Meeting', Geneva, CH, 14-23Jan.2013, JCTVC-L107.

Contents of the invention

In order to solve the above technical problems, this patent improves the sample adaptive compensation in HEVC. First, additional boundary compensation is performed for boundary pixels of each coding unit in the largest coding unit. Second, add a new mode to edge compensation (EO) types 2 and 3: the pixel value difference between the current pixel and the left adjacent pixel is 1, and the difference between the right adjacent pixel and the right adjacent pixel is more than 3, and the current pixel and the left adjacent pixel The pixel value of is different by 2, and the difference from the adjacent pixel on the right is more than 6. Finally, the cases where the difference between the current pixel and the neighboring pixels is less than or equal to 1 in types 2 and 3 are eliminated.

According to one aspect, a method for determining sample-adaptive compensation parameters at a decoder in HEVC, the method being implemented in an encoder, the method comprising:

Reconstruct the largest coding unit (CU) from the encoded image;

Sample adaptive compensation is performed on the reconstructed largest CU based on the following sample adaptive compensation types: pixel band compensation (Band offset, BO) type, edge compensation (Edge offset, EO) type 1, EO type 2, EO type 3, and EO type 4, thereby obtaining compensation results for said BO type, and obtaining 4 compensation results for EO type 1, EO type 2, EO type 3 and EO type 4;

Based on the compensation result of the BO type and each of the four compensation results of EO type 1, EO type 2, EO type 3, and EO type 4 and the rate-distortion cost value of the reconstructed maximum CU relative to the original image, Select the optimal sample adaptive compensation type from BO type, EO type 0, EO type 1, EO type 2, EO type 3 and EO type 4, wherein the EO type 0 means no compensation and thus corresponds to the maximum CU of said reconstruction;

determining a first compensation value corresponding to the selected optimal sample adaptive compensation type, wherein the first compensation value corresponding to the EO type 0 is 0;

determining a second boundary compensation value based on the selected optimal sample adaptive compensation type,

encoding the selected optimal sample adaptive compensation type, the first compensation value and the second boundary compensation value into the encoding information associated with the largest CU for transmission;

It is characterized by:

Pixel type 2 includes the case where the pixel value of the current pixel is smaller than the pixel value of the left adjacent pixel by 1 and greater than the pixel value of the right adjacent pixel by more than 3, and the pixel value of the current pixel is larger than that of the left adjacent pixel When the pixel value is smaller than 2 and larger than the pixel value of the adjacent pixel on the right by 6 or more,

Pixel type 3 includes the case where the pixel value of the current pixel is greater than the pixel value of the right adjacent pixel by 1, smaller than the pixel value of the left adjacent pixel by more than 3, and the pixel value of the current pixel is greater than that of the right adjacent pixel When the pixel value is greater than 2 and less than the pixel value of the adjacent pixel on the left by more than 6,

In pixel type 2 and pixel type 3, if the difference between the pixel value of the current pixel and the pixel value of two adjacent pixels is less than or equal to 1, the current pixel will not be compensated,

The determining a second boundary compensation value includes: calculating a plurality of rate-distortion cost values for a plurality of boundary compensation values of boundary pixels of each transform unit in the largest CU based on the optimal sample adaptive compensation type, and A boundary compensation value is selected based on the plurality of rate-distortion cost values.

In a further aspect, the values of the syntax elements sao_type_idx and sao_eo_class in the largest CU are set based on the optimal sample adaptive compensation type, wherein the values of sao_type_idx are 0, 1, and 2 respectively indicating no sample adaptive compensation, using Pixel band compensation (BO) and edge compensation (EO), the value of sao_eo_class 0-3 means EO type 1, EO type 2, EO type 3 and EO type 4.

In a further aspect, the EO types 1-4 use horizontal lines, vertical lines, 135 degree diagonal lines and 45 degree diagonal lines, respectively.

In a further aspect, in the BO type, the entire pixel range is uniformly divided into 32 pixel bands, 4 of which are selected, and positive compensation or negative compensation for the selected 4 pixel bands is determined, Wherein the rate-distortion cost is used to determine the positive compensation or negative compensation for the selected 4 pixel bands, the optimal sample adaptive compensation type is the BO type, and the first compensation value includes the determined Positive or negative compensation for the selected 4 pixel bands.

In a further aspect, in the EO types 1-4, the first compensation value includes compensation values determined based on compensation functions respectively corresponding to the EO types 1-4.

In a further aspect, in the EO types 1-4, the first compensation value includes compensation values determined based on compensation functions respectively corresponding to the EO types 1-4.

According to another aspect, an apparatus for determining sample-adaptive compensation parameters at a decoder in HEVC, said apparatus being implemented in an encoder, said apparatus comprising:

A unit for reconstructing a largest coding unit (CU) from an encoded image;

It is used to perform sample adaptive compensation on the reconstructed largest CU based on the following four sample adaptive compensation types: pixel band compensation (Band offset, BO) type, edge compensation (Edge offset, EO) type 1, EO type 2, EO type 3 and EO type 4, thereby obtaining compensation results for said BO type, and obtaining 4 units of compensation results for EO type 1, EO type 2, EO type 3 and EO type 4;

For each of the 4 compensation results based on the BO type and the 4 compensation results of EO type 1, EO type 2, EO type 3, and EO type 4 and the rate-distortion generation of the reconstructed maximum CU relative to the original image Value, select the optimal sample adaptive compensation type from BO type, EO type 0, EO type 1, EO type 2, EO type 3 and EO type 4, wherein the EO type 0 means no compensation and thus a unit corresponding to the reconstructed largest CU;

A unit for determining a first compensation value corresponding to the selected optimal sample adaptive compensation type, wherein the first compensation value corresponding to the EO type 0 is 0;

means for determining a second boundary compensation value based on the selected optimal sample adaptive compensation type,

A unit for encoding the selected optimal sample adaptive compensation type, the first compensation value and the second boundary compensation value into the encoding information associated with the largest CU for transmission;

It is characterized by:

Pixel type 2 includes the case where the pixel value of the current pixel is smaller than the pixel value of the left adjacent pixel by 1 and greater than the pixel value of the right adjacent pixel by more than 3, and the pixel value of the current pixel is larger than that of the left adjacent pixel When the pixel value is smaller than 2 and larger than the pixel value of the adjacent pixel on the right by 6 or more,

Pixel type 3 includes the case where the pixel value of the current pixel is greater than the pixel value of the right adjacent pixel by 1, smaller than the pixel value of the left adjacent pixel by more than 3, and the pixel value of the current pixel is greater than that of the right adjacent pixel When the pixel value is greater than 2 and less than the pixel value of the adjacent pixel on the left by more than 6,

In pixel type 2 and pixel type 3, if the difference between the pixel value of the current pixel and the pixel value of two adjacent pixels is less than or equal to 1, the current pixel will not be compensated,

The unit for determining the second boundary compensation value includes: calculating a plurality of boundary compensation values for boundary pixels of each transform unit in the largest CU based on the optimal sample adaptive compensation type. rate-distortion cost values, and select a unit of boundary compensation value based on the multiple rate-distortion cost values.

According to another aspect, the invention proposes a computer program product comprising a computer readable medium comprising a program code which, when executed by a processor, performs the method as described above.

According to another aspect, the present invention proposes an HEVC-based codec for performing the video coding method as described above.

Description of drawings

Figure 1 shows a high-level flowchart of sample adaptive compensation according to one embodiment of the present invention.

Fig. 2 shows a schematic diagram of pixel band compensation according to an embodiment of the present invention.

Fig. 3 shows the determination of pixel edge compensation type according to an embodiment of the present invention.

Fig. 4 shows a schematic diagram of types of similar edge compensation according to an embodiment of the present invention.

FIG. 5 shows edge units inside a LCU according to an embodiment of the present invention (the dotted line part is a transform unit boundary pixel).

Fig. 6 shows a schematic diagram of the second and third types of sample adaptive compensation according to an embodiment of the present invention.

Fig. 7 shows a schematic diagram of pixel types without sample adaptive compensation according to an embodiment of the present invention.

FIG. 8 illustrates a request for no sample adaptive compensation according to one embodiment of the present invention.

Figures 9(a) and (b) illustrate a Type 2 supplementary pixel pattern according to one embodiment of the present invention.

Figures 9(c) and (d) illustrate supplementary pixel patterns showing Type 3 according to one embodiment of the present invention.

Figures 9(e) and (f) show pixel patterns that need to be culled from types 2 and 3 according to one embodiment of the present invention.

Fig. 10 shows a schematic diagram of a method according to an embodiment of the present invention.

Fig. 11(a) shows a flowchart of a method according to an embodiment of the present invention.

Fig. 11(b) shows a flow chart of the apparatus according to one embodiment of the present invention.

Figure 12 shows the basic schematic diagram of the HEVC encoder.

Detailed ways

Various aspects are now described with reference to the figures. In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of one or more aspects. It may be evident, however, that such aspects can be practiced without these specific details.

As used in this application, the terms "component", "module", "system" and the like are intended to refer to a computer-related entity such as, but not limited to, hardware, firmware, a combination of hardware and software, software, Or software in execution. For example, a component may be, but is not limited to being limited to, a process running on a processor, a processor, an object, an executable, a thread of execution, a program, and/or a computer. For example, both an application running on a computing device and the computing device can be components. One or more components can reside within a process and/or thread of execution and a component can be localized on one computer and/or distributed between two or more computers. In addition, these components can execute from various computer readable media having various data structures stored thereon. Components can communicate by means of local and/or remote processes, such as from signals having one or more data packets, for example, from interacting with another component in a local system, a distributed system, and/or with another component in a distributed system by means of a signal. Data of a component on a network such as the Internet that interacts with other systems by means of signals.

FIG. 12 shows a schematic structural diagram of a video encoder implemented by High Efficiency Video Coding (HEVC). The encoder architecture of HEVC is roughly the same as the encoder architecture used by H.264, mainly for further research and improvement on the algorithms used in each module, especially for high-resolution video sequences. The goal of the improvement is Under the same video quality (PSNR), the bit rate is reduced to 50% of the H.264 standard.

Since the encoder architecture of HEVC is roughly the same as the encoder architecture used by H.264, so as not to confuse the present invention, the overall architecture in FIG. 9 is not described in this application.

In HEVC, the input video is first divided into small blocks called coding tree units (coding tree units, CTU). Those skilled in the art can understand that a CTU is equivalent to the concept of a previous standard macroblock (macroblock). A coding unit (CU) is a square (pixel) unit with a prediction mode (intra, inter, or skip). The prediction unit division method based on CTU and CU is shown in Fig. 1 .

1.1 Introduction to Sample Adaptive Compensation

Sample adaptive compensation is a new technology emerging in HEVC, which adds fresh blood to HEVC.

A new loop filter is used in HEVC, including a deblocking filter (Deblocking Filter, DF) and a sample adaptive offset (Sample Adaptive Offset, SAO). The sample adaptive compensation is located after the deblocking filter, and the purpose is to reduce the average distortion of the region. The sample adaptive compensation first uses the selected classification method to classify and count the samples in the largest coding unit, so as to obtain a Compensation value, and then add the compensation value to the corresponding kind of pixel value.

Adaptive sample compensation is a module located after the deblocking filter in the encoding process. This module completes a process of modifying the decoded sample value. The modification process is to compensate a certain value for the sample under certain circumstances [1] [2]. The sample adaptive compensation filter is applied to a certain area. In each LCU, a syntax element of sao_type_idx is used to select the filter mode of the unit, where "sao" means samples adaptive offset. When sao_type_idx is 0, it means that the largest coding unit does not apply sample adaptive compensation. When the value is 1 and 2, it means that pixel band compensation (BO, band offset) and edge compensation (EO, edge offset) are used [3] [4]. The flowchart of sample adaptive compensation is shown in Fig.1.

In Pixel Band Compensation mode, the selected compensation value is directly dependent on the pixel's amplitude. In this mode, the range of all pixels is uniformly divided into 32 parts, that is, divided into 32 pixel bands, as shown in Figure 2, the pixel band mode will select 4 consecutive pixel bands, and The sample values in the four pixel bands are positively (directly) compensated or negatively (directly) compensated. The reason why continuous 4-pixel bands are selected for compensation is that banding artifacts usually appear in smooth areas, and pixel values in a maximum coding unit tend to be concentrated in fewer pixel bands. Selecting 4 compensation values at the same time matches the operation of the edge compensation mode, because the edge compensation also selects 4 compensation values.

In edge compensation, the value of the syntax element sao_eo_class from 0 to 3 indicates that horizontal, vertical, 135-degree diagonal lines and 45-degree diagonal lines are used in edge compensation, as shown in Figure 3 . All pixel values in the largest coding unit will be classified as one of the five types shown in Table 1 and Figure 4, where c represents the current pixel value, and a and b represent two adjacent pixel values respectively. The classification is based on the decoded pixel sample values, so no additional signal is needed to convey the classification result [5]. Relying on the edge index classification on the pixel samples, through a lookup table transmitted to the decoder, we can add corresponding compensation values to the corresponding sample values. For types 1 and 2, the compensation value is always positive (towards), while the compensation value of types 3 and 4 is always negative (towards), so edge filtering produces a smooth effect.

In Fig. 4, types 1 and 2 perform positive compensation, and types 3 and 4 perform negative compensation.

Table 1 Pixel boundary compensation classification decision

type condition 1 c<a&&c<b 2 (c<a&&c==b)||(c==a&&c<b) 3 (c>a&&c==b)||(c==a&&c>b) 4 c＞a&&c＞b 0 Not above

In summary, the pixel band compensation and edge compensation of the sample adaptive compensation will respectively transmit the magnitude of 4 compensation values for each maximum coding unit, and for the pixel band compensation, the sign of the compensation (positive or negative) needs to be transmitted of [6]. The selection of compensation values and corresponding syntax element values, such as sao_type_idx and sao_eo_class, are selected by optimizing the effect of rate distortion. Meanwhile, the sample adaptive compensation parameters of the current block may also be obtained by inheriting the sample adaptive compensation parameters of the left and upper LCUs. In general, sample-adaptive compensation is a nonlinear filtering operation that allows further improvement of the reconstructed signal, and it can both enhance the signal expression in smooth regions and improve the edge parts.

2. Sample Adaptive Compensation Optimization Based on Transform Unit Boundary

The size of the transformation unit (TransformUnit, TU) in HEVC is from 4x4 to 32x32, while the size of the transformation unit in H.264 is not larger than 8x8. Because the use of discrete cosine transform (Discrete Cosine Transform, DCT) will cause quantization distortion of transform coefficients, a larger transform unit will cause more obvious ring artifacts and more serious distortion at the edge of the transform unit. However, the sample adaptive compensation achieves the reduction of the average distortion degree of the region, and the deblocking filter does not reduce the distortion degree, so the sample distortion at the edge of the TU has not been specially processed.

optimization

Through our research, it is found that even for the transformation unit after the deblocking filter, the distortion on the boundary of the transformation unit is much larger than that inside the transformation unit. In order to reduce distortion on TU boundaries in a targeted manner, we add a new mode for sample adaptive compensation, TU edge mode. This mode aims to give an extra compensation to TU boundaries, resulting in lower TU boundary sample distortion and lower bitrate. The transformation unit inside the largest coding unit is shown in Figure 5, where the dotted line part is the pixel sample on the boundary of the transformation unit.

After statistics, we found that the distortion at the boundary of the transformation unit is about 20% larger than that inside the transformation unit. In order to further obtain the compensation difference between the boundary of the transformation unit and the interior of the transformation unit, we pass the samples at the boundary of the transformation unit and the samples inside the transformation unit through samples Adaptive Compensation. The results show that for edge compensation, the average distortion degree of TU boundary is 0.45 larger, and for pixel band compensation, the average distortion degree of TU boundary is 0.4 larger.

Therefore, we propose the following novel compensation method:

a) When counting the samples, count the pixel samples inside the transformation unit and the samples on the boundary of the transformation unit respectively.

b) Calculate the optimal compensation mode and compensation value according to the original sample adaptive compensation, and calculate the rate-distortion cost. As shown in the type index 0 in Table 2, that is, no additional compensation is performed on the boundary of the transformation unit.

c) Perform additional compensation on the boundary of the transformation unit, and traverse the strategies for performing additional compensation on the boundary of the transformation unit with type indexes 1, 2, and 3 in Table 2, that is, add the internal compensation value of the transformation unit to the value in Table 2 and then compensate to the interior of the transformation unit. Calculate the rate-distortion penalty imposed by these three types.

d) Compare the above four rate-distortion cost values, select the type with the smallest cost as the compensation type of the current block, and transmit the index value.

Table 2 Additional compensation for transform units

The experimental results show that this algorithm reduces the BD-rate by -0.13 compared with the original algorithm, and the performance of the sample adaptive compensation algorithm has been improved by 26%. It can be seen that this algorithm has a certain effect on the optimization of sample adaptive compensation .

3. Compensation mode optimization for sample adaptive compensation

In the process of research, we found that sample adaptive compensation has very strict requirements on sample statistics, and there are a large number of regular pixels that need to be compensated but have not received corresponding compensation. As shown in Figure 6, in the second and third types of sample adaptive compensation, we can see that only when the current pixel must have exactly the same pixel value as an adjacent pixel will it be judged whether the current pixel belongs to the second, Three types of sample adaptive compensation. The problem caused by such a compensation strategy is that there are a large number of pixels as shown in Figure 7, that is, pixels with a large gradient but cannot be judged as samples for adaptive compensation, but no compensation is performed. In Figure 7, pixel c and The difference between pixel a or b is only small, but the gradient is very large. This situation is completely in line with the original design intention of sample adaptive compensation to make the edge part smoother, and the pixel difference between the pixel in Figure 7 and the standard type in Figure 6 is only 1, 2 pixel values, so we propose this algorithm based on this idea.

At the same time, we found that there is an unreasonable situation in the sample adaptive compensation. This situation is: the current pixel is judged as the second or third type, but even if the current pixel value differs from the adjacent pixel by only one or very few pixel values, that is, the pixel gradient is very small, as shown in Figure 8, the The position is also compensated according to the second or third type. This is obviously unreasonable, because the current pixel is already smooth, and there is no room for further smoothing, which violates the original intention of edge compensation in sample adaptive compensation. Therefore, we also take this as one of the contents of our research.

Therefore, we propose that adding the four modes a, b, c, and d shown in Figure 9 into the sample adaptive compensation will bring many pixel samples that need to be supplemented but not compensated into the sample adaptive compensation, and remove The e and f modes are more conducive to obtaining more accurate compensation values.

Specifically, the situation shown in Figure 9(a) is that the pixel value difference between the current pixel c and the adjacent pixel a is 1, and the difference between the pixel value of the adjacent pixel b on the other side is more than 3; Figure 9(b ) is that the pixel value difference between the current pixel c and the adjacent pixel a is 2, and the pixel value difference between the pixel b on the other side is more than 6; the situation shown in Figure 9(c) is that the current pixel c The pixel value of the adjacent pixel b differs by 1, and the pixel value of the adjacent pixel a on the other side differs by more than 3; the situation shown in Figure 9(d) is that the pixel value of the current pixel c is different from the adjacent pixel b 2, and the pixel value of pixel a on the other side differs by more than 6.

We include the situation in Figure 9(a) and (b) into Type 2, and the situation in Figure 9(c) and (d) into Type 3, and then we exclude the situation in Figure 9(e) and (f) Types 2 and 3 compensation of sample adaptive compensation.

Then perform statistics and rate-distortion optimization according to the standard steps of sample adaptive compensation, select the optimal compensation mode for the current largest coding unit, and perform sample adaptive compensation on the current largest coding unit.

Fig. 11(a) shows a flowchart of a method for determining sample adaptive compensation parameters at the decoder side in HEVC according to an embodiment of the present invention, and the method is implemented in an encoder.

Combined with the schematic diagram shown in Figure 10, it can be seen that:

In step 1101, in an encoder, a largest coding unit (CU) is reconstructed from an encoded image.

In step 1102, sample adaptive compensation is performed on the reconstructed largest CU based on the following four types of sample adaptive compensation: pixel band offset (Band offset, BO) type, edge compensation (Edge offset, EO) type 1, EO Type 2, EO Type 3, and EO Type 4, thereby obtaining compensation results for the BO type, and obtaining 4 compensation results for EO Type 1, EO Type 2, EO Type 3, and EO Type 4.

In step 1103, based on the compensation result of the BO type and each of the four compensation results of EO type 1, EO type 2, EO type 3 and EO type 4 and the maximum CU of the reconstruction relative to the original image Rate-distortion cost value, select the optimal sample adaptive compensation type from BO type, EO type 0, EO type 1, EO type 2, EO type 3 and EO type 4, wherein the EO type 0 means no compensation and thus corresponds to the largest CU of the reconstruction.

In step 1104, a first compensation value corresponding to the selected optimal sample adaptive compensation type is determined, wherein the first compensation value corresponding to the EO type 0 is 0.

In step 1105, a second boundary compensation value is determined based on the selected optimal sample adaptive compensation type.

In step 1106, the selected optimal sample adaptive compensation type, the first compensation value and the second boundary compensation value are encoded into the encoding information associated with the largest CU for transmission.

In one embodiment of the present invention, pixel 2 includes the case that the pixel value of the current pixel is 1 smaller than the pixel value of the adjacent pixel on the left and greater than the pixel value of the adjacent pixel on the right by more than 3, and the pixel value of the current pixel When the pixel value is 2 smaller than the pixel value of the adjacent pixel on the left and greater than the pixel value of the adjacent pixel on the right by 6 or more.

In one embodiment of the present invention, pixel 3 includes the case where the pixel value of the current pixel is greater than the pixel value of the adjacent pixel on the right by 1, and the pixel value of the adjacent pixel on the left is smaller than the pixel value of more than 3, and the pixel value of the current pixel When the value is 2 greater than the pixel value of the adjacent pixel on the right and 6 or more smaller than the pixel value of the adjacent pixel on the left.

In one embodiment of the present invention, in pixel type 2 and pixel type 3, if the difference between the pixel value of the current pixel and the pixel values of two adjacent pixels is less than or equal to 1, the current pixel does not perform compensation.

In step 1105, determining the second boundary compensation value further includes: based on the optimal sample adaptive compensation type, calculating multiple rate-distortion rates for multiple boundary compensation values of boundary pixels of each transform unit in the largest CU cost, and select a boundary compensation value based on the plurality of rate-distortion cost values.

In a preferred embodiment of the present invention, the values of the syntax elements sao_type_idx and sao_eo_class in the largest CU are set based on the optimal sample adaptive compensation type, wherein the values of sao_type_idx are 0, 1, and 2 respectively indicating that no sample self-adaptive compensation is performed. Adaptive compensation, using pixel band compensation (BO) and edge compensation (EO), the value 0-3 of sao_eo_class means EO type 1, EO type 2, EO type 3 and EO type 4.

In a preferred embodiment of the present invention, the EO types 1-4 respectively use horizontal lines, vertical lines, 135-degree diagonal lines and 45-degree diagonal lines.

In a preferred embodiment of the present invention, in the BO type, the full pixel range is uniformly divided into 32 pixel bands, 4 of which are selected, and the positive compensation for the selected 4 pixel bands is determined or negative compensation, wherein a rate-distortion cost is used to determine positive or negative compensation for the selected 4 pixel bands, the optimal sample adaptive compensation type is the BO type, and the first compensation value includes Positive or negative compensation is determined for the selected 4 pixel bands.

In a preferred embodiment of the present invention, in the EO types 1-4, the first compensation value includes compensation values determined based on compensation functions respectively corresponding to the EO types 1-4.

In a preferred embodiment of the present invention, in the EO types 1-4, the first compensation value includes compensation values determined based on compensation functions respectively corresponding to the EO types 1-4.

Fig. 11(b) shows a block diagram of an apparatus for determining sample-adaptive compensation parameters at the decoder side in HEVC according to an embodiment of the present invention, and the apparatus is implemented in an encoder. Each unit in FIG. 11( b ) corresponds to the corresponding method steps in FIG. 11( a ), so details are not repeated here.

The method disclosed in the present invention can be realized by software, hardware, firmware and so on.

When implemented in hardware, video encoders can be implemented using general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gates, or transistor logic devices, discrete hardware components, or any combination thereof designed to perform the functions described herein, may be implemented or performed. A general-purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, eg, a DSP and a microprocessor, multiple microprocessors, one or more microprocessors with a DSP core, or any other such architecture. Additionally, at least one processor may comprise one or more modules operable to perform one or more of the steps and/or operations described above.

When a video encoder is implemented with a hardware circuit such as an ASIC, FPGA, it may include various circuit blocks configured to perform various functions. Those skilled in the art can design and implement these circuits in various ways according to various constraints imposed on the entire system, so as to realize various functions disclosed in the present invention.

While the foregoing disclosures discuss exemplary aspects and/or embodiments, it should be noted that many changes and/or changes may be made therein without departing from the scope of the described aspects and/or embodiments as defined by the claims. Revise. Also, although elements of the described aspects and/or embodiments are described or claimed in the singular, the plural is contemplated unless limitation to the singular is explicitly stated. In addition, all or part of any aspect and/or embodiment can be used in combination with all or part of any other aspect and/or embodiment, unless a difference is indicated.

Claims (8)

1. A method for determining sample adaptive compensation parameters at a decoder end in HEVC, said method being implemented in an encoder, said method comprising: Reconstruct the largest coding unit (CU) from the encoded image; The reconstructed largest CU is sample-adaptively compensated based on the following sample-adaptive compensation types: Pixel Band Offset (BO) Type, Edge Offset (EO) Type 1, EO Type 2, EO Type 3, and EO Type 4, so that Compensation results are obtained for the BO type, and 4 compensation results are obtained for EO type 1, EO type 2, EO type 3 and EO type 4; Based on the compensation result of the BO type and each of the four compensation results of EO type 1, EO type 2, EO type 3, and EO type 4 and the rate-distortion cost value of the reconstructed maximum CU relative to the original image, Select the optimal sample adaptive compensation type from BO type, EO type 0, EO type 1, EO type 2, EO type 3 and EO type 4, wherein the EO type 0 means no compensation and thus corresponds to the maximum CU of said reconstruction; determining a first compensation value corresponding to the selected optimal sample adaptive compensation type, wherein the first compensation value corresponding to the EO type 0 is 0; determining a second boundary compensation value based on the selected optimal sample adaptive compensation type, encoding the selected optimal sample adaptive compensation type, the first compensation value and the second boundary compensation value into the encoding information associated with the largest CU for transmission; It is characterized by: Pixel type 2 includes the case where the pixel value of the current pixel is smaller than the pixel value of the left adjacent pixel by 1 and greater than the pixel value of the right adjacent pixel by more than 3, and the pixel value of the current pixel is larger than that of the left adjacent pixel When the pixel value is smaller than 2 and larger than the pixel value of the adjacent pixel on the right by 6 or more, Pixel type 3 includes the case where the pixel value of the current pixel is greater than the pixel value of the right adjacent pixel by 1, smaller than the pixel value of the left adjacent pixel by more than 3, and the pixel value of the current pixel is greater than that of the right adjacent pixel When the pixel value is greater than 2 and less than the pixel value of the adjacent pixel on the left by more than 6, In pixel type 2 and pixel type 3, if the difference between the pixel value of the current pixel and the pixel value of two adjacent pixels is less than or equal to 1, the current pixel will not be compensated, The determining a second boundary compensation value includes: calculating a plurality of rate-distortion cost values for a plurality of boundary compensation values of boundary pixels of each transform unit in the largest CU based on the optimal sample adaptive compensation type, and selecting a boundary compensation value based on the plurality of rate-distortion cost values, Wherein, the values of the syntax elements sao_type_idx and sao_eo_class in the largest CU are set based on the optimal sample adaptive compensation type, wherein the values of sao_type_idx 0, 1, and 2 respectively indicate that no sample adaptive compensation is performed, and pixel band compensation ( BO) and edge compensation (EO), the value 0-3 of sao_eo_class means EO type 1, EO type 2, EO type 3 and EO type 4. 2. The method of claim 1, wherein the EO types 1-4 use horizontal lines, vertical lines, 135-degree diagonal lines, and 45-degree diagonal lines, respectively. 3. The method according to claim 1, wherein, in the BO type, the entire pixel range is uniformly divided into 32 pixel bands, 4 of which are selected, and the selected 4 pixel bands are determined to be positive or negative compensation of , wherein the rate-distortion cost is used to determine the positive or negative compensation for the selected 4 pixel bands, the optimal sample adaptive compensation type is the BO type, and the first The compensation value includes positive or negative compensation determined for the selected 4 pixel bands. 4. The method of claim 1, wherein, in the EO types 1-4, the first compensation value comprises compensation values determined based on compensation functions respectively corresponding to the EO types 1-4. 5. A device for determining sample adaptive compensation parameters at a decoder in HEVC, said device being implemented in an encoder, said device comprising: A unit for reconstructing a largest coding unit (CU) from an encoded image; Used to perform sample adaptive compensation on the reconstructed largest CU based on the following 4 sample adaptive compensation types: Pixel Band Offset (BO) Type, Edge Offset (EO) Type 1, EO Type 2, EO Type 3, and EO Type 4, thereby obtaining compensation results for the BO type, and obtaining 4 units of compensation results for EO type 1, EO type 2, EO type 3 and EO type 4; For each of the 4 compensation results based on the BO type and the 4 compensation results of EO type 1, EO type 2, EO type 3, and EO type 4 and the rate-distortion generation of the reconstructed maximum CU relative to the original image Value, select the optimal sample adaptive compensation type from BO type, EO type 0, EO type 1, EO type 2, EO type 3 and EO type 4, wherein the EO type 0 means no compensation and thus a unit corresponding to the reconstructed largest CU; A unit for determining a first compensation value corresponding to the selected optimal sample adaptive compensation type, wherein the first compensation value corresponding to the EO type 0 is 0; means for determining a second boundary compensation value based on the selected optimal sample adaptive compensation type, A unit for encoding the selected optimal sample adaptive compensation type, the first compensation value and the second boundary compensation value into the encoding information associated with the largest CU for transmission; It is characterized by: Pixel type 2 includes the case where the pixel value of the current pixel is smaller than the pixel value of the left adjacent pixel by 1 and greater than the pixel value of the right adjacent pixel by more than 3, and the pixel value of the current pixel is larger than that of the left adjacent pixel When the pixel value is smaller than 2 and larger than the pixel value of the adjacent pixel on the right by 6 or more, Pixel type 3 includes the case where the pixel value of the current pixel is greater than the pixel value of the right adjacent pixel by 1, smaller than the pixel value of the left adjacent pixel by more than 3, and the pixel value of the current pixel is greater than that of the right adjacent pixel When the pixel value is greater than 2 and less than the pixel value of the adjacent pixel on the left by more than 6, In pixel type 2 and pixel type 3, if the difference between the pixel value of the current pixel and the pixel value of two adjacent pixels is less than or equal to 1, the current pixel will not be compensated, The unit for determining the second boundary compensation value includes: calculating a plurality of boundary compensation values for boundary pixels of each transform unit in the largest CU based on the optimal sample adaptive compensation type. Rate-distortion cost values, and select a unit of boundary compensation value based on the plurality of rate-distortion cost values, Wherein, the values of the syntax elements sao_type_idx and sao_eo_class in the largest CU are set based on the optimal sample adaptive compensation type, wherein the values of sao_type_idx 0, 1, and 2 respectively indicate that no sample adaptive compensation is performed, and pixel band compensation ( BO) and edge compensation (EO), the value 0-3 of sao_eo_class means EO type 1, EO type 2, EO type 3 and EO type 4. 6. The device according to claim 5, wherein, in the BO type, the entire pixel range is uniformly divided into 32 pixel bands, 4 of which are selected, and the selected 4 pixel bands are determined to be positive or negative compensation of , wherein the rate-distortion cost is used to determine the positive or negative compensation for the selected 4 pixel bands, the optimal sample adaptive compensation type is the BO type, and the first The compensation value includes positive or negative compensation determined for the selected 4 pixel bands. 7. A computer readable medium containing program code which, when executed by a processor, performs the method of claim 1. 8. A codec based on HEVC, which is used to execute the video coding method according to any one of claims 1-4.