WO2017004883A1 - 基于时域信息的自适应视频预处理方法 - Google Patents
基于时域信息的自适应视频预处理方法 Download PDFInfo
- Publication number
- WO2017004883A1 WO2017004883A1 PCT/CN2015/089105 CN2015089105W WO2017004883A1 WO 2017004883 A1 WO2017004883 A1 WO 2017004883A1 CN 2015089105 W CN2015089105 W CN 2015089105W WO 2017004883 A1 WO2017004883 A1 WO 2017004883A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- pixel
- value
- domain information
- bilateral filtering
- adaptive video
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/154—Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/85—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression
Definitions
- the present invention relates to the field of video preprocessing, and in particular to an adaptive video preprocessing method based on domain information.
- the video preprocessing technique is a series of processing operations performed before the video is encoded, and the main purpose is to reduce the code rate and improve the quality.
- video preprocessing techniques mainly include traditional methods of filtering, interpolation and deinterlacing. Pre-processing methods based on the combination of the coding of the region of interest and JND can effectively improve the quality and efficiency of video coding.
- the technical problem to be solved by the present invention is to provide an adaptive video preprocessing method based on domain information, to maximize the correlation between the front and back frames, and to reduce the code rate and improve the subjective quality.
- An adaptive video preprocessing method based on domain information the steps of which are as follows:
- Step Sl reading in one frame of video image data
- Step S2 obtaining weighting coefficients of each pixel based on the domain information; [0009] Step S3, calculating a bilateral filter value of each pixel;
- Step S4 According to the weighting coefficient and the bilateral filtering value, the value of the processed pixel is taken from the weighted sum of the original pixel and the bilateral filtering value of the pixel, that is, the video pre-processing value is obtained.
- step S1 when step S1 is performed, the first frame data is not processed.
- step S2 is performed to calculate a weighting coefficient of each pixel based on the running estimated residual, and the weighting coefficient w t (p) is calculated according to the following information:
- p is the current pixel position
- ⁇ is a pixel region of a predetermined size centered at p
- Residue n (p) is Thresh, which is to cause the function to be smooth.
- step S3 the calculation formula of the bilateral filtering value ⁇ ⁇ ) is as follows:
- Q f is the window size of the bilateral filtering
- I n (p) is the original pixel value
- It is the spatial domain kernel and the value domain kernel, which are calculated by the following formulas:
- step S4 the calculation formula of the pixel value after the image processing is as follows:
- I n (p) is the original pixel value
- I f (p) is the bilateral filtering value of the pixel
- Thresh 2 is the threshold value, when the motion of the pixel in the region is estimated When the residual is below the threshold, the pixel value remains unchanged.
- the present invention calculates the weighting coefficient based on the ⁇ domain information, especially the weighting coefficient based on the motion estimation residual, thereby realizing the quantitative calculation of the influence of the ⁇ domain information, Then, the original image is processed based on the bilateral filtering value and the weighting coefficient, and the obtained image processed value is a weighted sum of the bilateral filtering value and the original pixel value, so that the motion estimation is accurate and the relative area is kept unchanged.
- the boundary preservation effect of noise reduction is realized by bilateral filtering, which can preserve the correlation between the front and back frames of the video to the greatest extent, and obtain the obvious rate reduction. And the improvement of subjective quality.
- FIG. 1 is a frame diagram of an adaptive video preprocessing method based on domain information according to the present invention.
- FIG. 2 is a graph showing changes in rate yield and average return of a test sequence at various resolutions of the adaptive video preprocessing method based on the domain information of the present invention.
- FIG. 3 is a schematic diagram of an adaptive video preprocessing method based on the domain information of the present invention at 174 kbps, a direct coded sequence frame, and a preprocessed coded sequence frame.
- FIG. 4 is a schematic diagram showing two regions of the 37th frame sequence of the PartyScen e at a code rate of 646 kbps according to the adaptive video preprocessing method based on the domain information of the present invention.
- the present invention provides an adaptive video preprocessing method based on domain information, and the main steps are as follows:
- Step S1 reading in one frame of video image data
- Step S2 obtaining weighting coefficients of each pixel based on the domain information
- Step S3 calculating a bilateral filtering value of each pixel
- Step S4 Obtain a video pre-processing value according to the weighting coefficient and the bilateral filtering value.
- the first frame data may not be processed by default.
- the processing method is adopted for the pixel points or regions to be processed, which not only removes redundant information, but also makes the motion estimation more accurate.
- the boundary retention characteristic of the processing mode can maintain important features of the image.
- the present invention uses a weighting coefficient, so that the value of the processed pixel is taken from the original pixel and the weighted sum of the bilaterally filtered values of the pixel, combined with the motion estimation information, the weighting coefficients are as follows: [0042]
- p is the current pixel position
- ⁇ is the 3x3 pixel region centered at p
- Residue n(p) is the residual of the motion estimation of the pixel point in the ⁇ region
- Thresh is for The parameter that makes the function smoother may range from 400 to 600, preferably 500. It can be seen that when the pixel region is stationary, or the motion estimation is more accurate, the sum of the squares of the motion residuals of the region will be smaller, which in turn will result in a larger weighting coefficient W t (p), and conversely, when the motion estimation residual is If the value is too large, the weighting factor W t (p) will become smaller.
- I f (p) bilateral filtering is as follows:
- Q f is the window size of the bilateral filtering. In the specific implementation, it can take the horizontal and vertical direction ⁇ 3 pixel position size, that is, 7x7 size window, and w d (p, q, o d ) and w r (p, q,o ⁇ (space domain kernel and value domain kernel, respectively:
- I n (p) is the original pixel value
- I f (p) is the value after the pixel is bilaterally filtered.
- the threshold Thresh 2 may have a value ranging from 40 to 60, preferably 50.
- the experimental effect of the preprocessing method of the present invention is as follows: Compared with the previous encoders such as H.264, HEVC adopts a more accurate and efficient prediction method for motion estimation, and thus in this experiment, we calculate the residual ⁇ The motion estimation method of HEVC described above is employed.
- the original video sequence and the pre-processed video sequence are encoded on the HM 12.0 encoder, respectively, and their code rate and subjective quality are compared.
- the test sequences are 416x240p, 832x480p, 1280x720p, 1920xl080p, respectively.
- the QP for each test sequence is chosen from 31 to 40.
- an average rate loss of 6% without subjective quality loss can be obtained (as shown in FIG. 2), and subjective quality can be improved in some video sequences, as shown in the figure. 3 and FIG. 4, wherein the upper image in FIG. 3 is a directly encoded image, and the lower image is a pre-processed encoded image; in the lower six small images in FIG. 4, the left a group is pre-processed.
- the encoded image, while the middle b group is the directly encoded image, and the right c group is the global bilateral filtered encoded image.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Image Processing (AREA)
Abstract
本发明公开一种基于时域信息的自适应视频预处理方法,步骤如下:步骤S1、读入一帧视频图像数据;步骤S2、获得各像素基于时域信息的加权系数;步骤S3、计算各像素的双边滤波值;步骤S4、根据加权系数和双边滤波值,让处理后的像素点的值取自于原像素及该像素双边滤波值的加权和,获得视频预处理值。本发明基于各像素的双边滤波值和加权系数处理原始图像,所获得的值为双边滤波值和原始像素值的加权和,实现了在运动估计准确以及相对静止区域保持其时域信息不变,在运动估计剧烈以及残差估计相对不准确的区域通过双边滤波实现降噪的边界保存效果,可最大程度地保留视频前后帧之间的时域相关性,并同时获得明显的码率下降和主观质量的提升。
Description
说明书 发明名称:基于时域信息的自适应视频预处理方法 技术领域
[0001] 本发明涉及视频预处理领域, 特别是涉及到基于吋域信息的自适应视频预处理 方法。
背景技术
[0002] 随着信息技术的发展, 尤其是高清视频和移动互联网的发展, 多媒体视频的数 量以及需求呈现快速的膨胀。 新一代视频编码标准 HEVC的发展较上一代的标准 已经降低了 50%的码率, 发展下一代新的视频编码标准并带到实际的应用中仍然 需要较长的发展吋间, 而通过结合视频预处理的技术, 可以明显的提升视频编 码的效率和提升其主观质量。
[0003] 视频预处理技术是在视频进行编码前, 对其进行的一系列处理操作, 主要目的 是为了降低码率和提升质量。 经过多年的研究发展, 视频预处理技术主要包括 传统的滤波、 插值和去隔行等方法。 基于感兴趣区域的编码和 JND相结合的预处 理方法都可以有效的提升视频编码的质量和效率。
[0004] 在常见的视频预处理技术中, 空域信息的处理很容易会破坏前后帧之间的吋域 相关性, 从而会降低视频预处理的效果。
技术问题
[0005] 本发明所要解决的技术问题是, 提供一种基于吋域信息的自适应视频预处理方 法, 以最大程度地保留前后帧之间的吋域相关性, 并降低码率, 提升主观质量
问题的解决方案
技术解决方案
[0006] 为解决上述技术问题, 本发明采用如下技术方案: 一种基于吋域信息的自适应 视频预处理方法, 其步骤如下:
[0007] 步骤 Sl、 读入一帧视频图像数据;
[0008] 步骤 S2、 获得各像素基于吋域信息的加权系数;
[0009] 步骤 S3、 计算各像素的双边滤波值;
[0010] 步骤 S4、 根据加权系数和双边滤波值, 让处理后的像素点的值取自于原像素以 及该像素双边滤波值的加权和, 即获得视频预处理值。
[0011〗 进一步地, 进行步骤 S1时, 对第一帧数据不进行处理。
[0012] 进一步地, 进行步骤 S2吋, 计算获得各像素基于运行估计残差的加权系数, 结 合运动估计的信息, 加权系数 w t(p)的计算公式如下:
'令 : * ¾
[0014] 在公式 (1) 中, p为当前像素位置, Ω为以 p为中心的预定大小的像素区域, 而 Residue n(p)为 Thresh ,是促使函数平滑的
[0015] 进一步地, 在步骤 S3中, 双边滤波值 Ι ρ)的计算公式如下:
[0017〗 在公式 (2) 中, Q f为双边滤波的窗口大小, I n(p)是原像素值, 而 w d(p, q,a d) 和 w r(p, q,o 分别是空间域核和值域核, 分别由以下公式计算而得:
(5)
[0022] 在公式 (5) 中, I n(p)是原像素值, I f(p)是该像素的双边滤波值, Thresh 2为闺 值, 当该区域中的像素点的运动估计的残差和低于阈值吋, 则保留其像素值不 变。
发明的有益效果
有益效果
[0023] 通过采用上述技术方案, 本发明具有以下技术效果: 本发明通过计算基于吋域 信息的加权系数, 特别是基于运动估计残差的加权系数, 从而将吋域信息的影 响实现量化计算, 再基于双边滤波值和加权系数处理原始图像, 所获得的图像 处理后的值为双边滤波值和原始像素值的加权和, 实现了在运动估计准确以及 相对静止区域保持其吋域信息不变, 在运动估计剧烈以及残差估计相对不准确 的区域通过双边滤波实现降噪的边界保存效果, 可以最大程度地保留视频前后 帧之间的吋域相关性, 并同吋获得明显的码率的下降和主观质量的提升。
对附图的简要说明
附图说明
[0024] 图 1是本发明基于吋域信息的自适应视频预处理方法的框架图。
[0025] 图 2是本发明基于吋域信息的自适应视频预处理方法的各个分辨率下测试序列 的码率收益以及平均收益的变化曲线图。
[0026] 图 3是本发明基于吋域信息的自适应视频预处理方法在 174kbps吋直接编码序列 帧以及预处理后编码序列帧的示意图。
[0027] 图 4是本发明基于吋域信息的自适应视频预处理方法在码率 646kbps下 PartyScen e第 37帧序列两个区域对比示意图。
本发明的实施方式
[0028] 需要说明的是, 在不冲突的情况下, 本申请中的实施例及实施例中的特征可以
相互结合, 下面结合附图和具体实施例对本发明作进一步详细说明。
[0029] 如图 1所示, 本发明提供一种基于吋域信息的自适应视频预处理方法, 其主要 步骤如下:
[0030] 步骤 Sl、 读入一帧视频图像数据;
[0031] 步骤 S2、 获得各像素基于吋域信息的加权系数;
[0032] 步骤 S3、 计算各像素的双边滤波值;
[0033] 步骤 S4、 根据加权系数和双边滤波值获得视频预处理值。
[0034] 以下对各步骤的具体实施过程进行详细说明。
[0035] 步骤 S1
[0036] 为保证吋域信息的准确性, 可默认第一帧数据不进行处理。
[0037] 步骤 S2
[0038] 在进行编码操作吋, 若以前的运动残差在进行变换量化之后会获得清零的效果 , 而在滤波操作之后可能不能够进行量化变换到 0, 从而会减弱降低码率的效果
[0039] 在视频编码的运动估计过程中, 当运动估计的残差非常小吋以及属于静止区域 吋, 证明运动估计的准确度比较高, 残差则会很小, 在进行最终的变换量化后 , 码率信息会很少, 同吋质量损失会很小, 因而在进行预处理的过程中, 对这 类的像素点或者像素区域进行保留或者进行轻微的滤波操作; 而对于其它运动 复杂的像素点或者区域, 可以进行强度更大的滤波操作, 去除视觉冗余信息, 保留或者强化关键区域, 进而可以达到降低码率提升质量的效果。
[0040] 由于双边滤波的边界保留特性, 因此在对要处理的像素点或者区域采用该处理 方式, 不但可以去除冗余信息, 让运动估计更加准确, 另一方面, 该处理方式 的边界保留特性, 可以维持图像的重要特征。
[0041] 为了让原始图像在进行预处理操作之后, 同吋对静止区域以及运动估计相对准 确区域保留更多的原始信息, 同吋对非静止区域以及运动估计误差相对偏大的 区域去除冗余信息但是保留关键特征信息, 本发明采用一个加权系数, 让处理 后的像素点的值取自于原像素以及该像素双边滤波值的加权和, 结合运动估计 的信息, 加权系数如下:
[0042]
½€ 纖咖 :¾ :, w
( 1)
[0043] 在公式 (1) 中, p为当前像素位置, Ω为以 p为中心的 3x3像素区域, 而 Residue n(p)为 Ω区域中像素点的运动估计的残差, Thresh ,是为了让该函数更加平滑的参 数, 其取值范围可以是 400~600, 优选为 500。 可以看出, 当该像素区域静止, 或者运动估计更加准确吋, 该区域的运动残差值平方和会更小, 进而会导致加 权系数 W t(p)更大, 反之, 当运动估计残差值偏大吋, 则加权系数 W t(p)就会变小
[0044] 步骤 S3
[0045] 进行本步骤吋, I f(p)双边滤波的定义如下:
在公式 (2) 中, Q f为双边滤波的窗口大小, 在具体实施吋, 可以取水平和垂 直方向 ±3像素位置大小, 即 7x7的大小窗口, 而 w d(p, q,o d)和 w r(p, q,o ^(分别是空 间域核和值域核:
[0050] 步骤 S4
(5)
[0053] 在公式 (5) 中, I n(p)是原像素值, I f(p)是该像素进行双边滤波之后的值。 当 运动残差越小以及相对静止区域, 预处理后的像素值更加趋于原像素值, 反之 , 预处理后的像素值更加趋近双边滤波的像素值。 通过设定阈值 Thresh 2 , 阈值 Thresh 2的取值范围可以是 40~60, 优选为 50。 当该区域的残差和低于阈值 吋, 认定该像素位置运动估计准确, 应该保留其像素值不变, 同吋, 也降低了
[0054] 本发明的预处理方法的实验效果如下: 相对于以前 H.264等编码器, HEVC在运 动估计上采取了更加准确和高效的预测方法, 因而在该实验中, 我们计算残差 吋, 采用了上面所描述的 HEVC的运动估计方法。 原视频序列以及进行预处理后 的视频序列分别在 HM 12.0编码器上进行编码, 比较其码率以及主观质量。 其中 在该实验中, 参数选择分别为 Thresh 1=500, Thresh 2=50, o d=3 , ^=15。 测试序 列分别有 416x240p, 832x480p, 1280x720p, 1920xl080p。 每个测试序列的 QP选 取从 31到 40之间。
[0055] 通过采用本发明的预处理方法, 可以获得平均 6%的无主观质量损失的码率下 降 (如图 2所示) , 而在一些视频序列中还能获得主观质量的提升, 如图 3和图 4 所示, 其中, 图 3中的上图为直接编码后的图像, 下图为预处理后编码的图像; 图 4中的下方六张小图中, 左侧的 a组为预处理后编码的图像, 而中间的 b组为直 接编码的图像, 右侧的 c组为全局双边滤波后编码的图像。
[0056] 尽管已经示出和描述了本发明的实施例, 对于本领域的普通技术人员而言, 可 以理解在不脱离本发明的原理和精神的情况下可以对这些实施例进行多种变化 、 修改、 替换和变型, 本发明的范围由所附权利要求及其等同范围限定。
Claims
[权利要求 1] 一种基于吋域信息的自适应视频预处理方法, 其特征在于, 其步骤如 下:
步骤 Sl、 读入一帧视频图像数据;
步骤 S2、 获得各像素基于吋域信息的加权系数;
步骤 S3、 计算各像素的双边滤波值;
步骤 S4、 根据加权系数和双边滤波值, 让处理后的像素点的值取自于 原像素以及该像素双边滤波值的加权和, 即获得视频预处理值。
[权利要求 2] 如权利要求 1所述的基于吋域信息的自适应视频预处理方法, 其特征 在于, 进行步骤 S1吋, 对第一帧数据不进行处理。
[权利要求 3] 如权利要求 1或 2所述的基于吋域信息的自适应视频预处理方法, 其特 征在于, 进行步骤 S2吋, 计算获得各像素基于运行估计残差的加权系 数, 结合运动估计的信息, 加权系数 w t(p)的计算公式如下:
在公式 (1) 中, p为当前像素位置, Ω为以 p为中心的预定大小的像 素区域, 而 Residue n(p)为 Ω区域中像素点的运动估计的残差, Thresh 是促使函数平滑的参数。
[权利要求 4] 如权利要求 3所述的基于吋域信息的自适应视频预处理方法, 其特征 在于, 在步骤 S3中, 双边滤波值 I f(p)的计算公式如下:
在公式 (2) 中, Q f为双边滤波的窗口大小, I n(p)是原像素值, 而 w d (p, q,o d)和 w r(p, q,o 分别是空间域核和值域核, 分别由以下公式计算
而得:
[权利要求 5] 如权利要求 4所述的基于吋域信息的自适应视频预处理方法, 其特征 在于, 在步骤 S4中, 图像处理之后的像素值的计算公式如下:
* :'' l4 ) + |i (0):' %C¾ c f lM
(5)
在公式 (5) 中, In(p)是原像素值, If
(p)是该像素的双边滤波值, Thresh 2为阈值, 当该区域中的像素点的 运动估计的残差和低于阈值吋, 则保留其像素值不变。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510397733.8 | 2015-07-08 | ||
| CN201510397733.8A CN104994397B (zh) | 2015-07-08 | 2015-07-08 | 基于时域信息的自适应视频预处理方法 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2017004883A1 true WO2017004883A1 (zh) | 2017-01-12 |
Family
ID=54306135
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2015/089105 Ceased WO2017004883A1 (zh) | 2015-07-08 | 2015-09-08 | 基于时域信息的自适应视频预处理方法 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN104994397B (zh) |
| WO (1) | WO2017004883A1 (zh) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105959663B (zh) * | 2016-05-24 | 2018-09-21 | 厦门美图之家科技有限公司 | 视频帧间信号连续性的优化处理方法、系统及拍摄终端 |
| CN109845266B (zh) * | 2016-10-14 | 2021-08-10 | 联发科技股份有限公司 | 用以移除涟漪效应的平滑化滤波方法和装置 |
| WO2021223144A1 (zh) * | 2020-05-07 | 2021-11-11 | 深圳市大疆创新科技有限公司 | 图像处理方法和装置 |
| CN114330662B (zh) * | 2021-12-29 | 2025-10-10 | 中国电信股份有限公司 | 运维指标监控模型训练方法、装置及数据滤波方法 |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1100260A1 (en) * | 1999-11-12 | 2001-05-16 | STMicroelectronics S.r.l. | Spatio-temporal filtering method for noise reduction during pre-processing of picture sequences in video encoders |
| WO2007071172A1 (en) * | 2005-12-21 | 2007-06-28 | Hong Kong Applied Science and Technology Research Institute Co. Ltd | Auto-regressive method and filter for denoising images and videos |
| US7394856B2 (en) * | 2003-09-19 | 2008-07-01 | Seiko Epson Corporation | Adaptive video prefilter |
| CN101448077A (zh) * | 2008-12-26 | 2009-06-03 | 四川虹微技术有限公司 | 一种自适应视频图像3d降噪方法 |
| CN103024248A (zh) * | 2013-01-05 | 2013-04-03 | 上海富瀚微电子有限公司 | 运动自适应的视频图像降噪方法及其装置 |
| CN103533214A (zh) * | 2013-10-01 | 2014-01-22 | 中国人民解放军国防科学技术大学 | 一种基于卡尔曼滤波和双边滤波的视频实时去噪方法 |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100454970C (zh) * | 2006-09-30 | 2009-01-21 | 四川长虹电器股份有限公司 | 空间域像素数据处理方法 |
| CN102281439B (zh) * | 2011-06-16 | 2013-07-24 | 杭州米加科技有限公司 | 流媒体视频图像预处理方法 |
| CN103632356B (zh) * | 2012-08-29 | 2017-08-18 | 华为技术有限公司 | 提高图像空间分辨率的方法及装置 |
| US9300846B2 (en) * | 2012-09-10 | 2016-03-29 | Apple Inc. | Signal shaping for improved mobile video communication |
| CN103108109B (zh) * | 2013-01-31 | 2016-05-11 | 深圳英飞拓科技股份有限公司 | 一种数字视频降噪系统与方法 |
-
2015
- 2015-07-08 CN CN201510397733.8A patent/CN104994397B/zh active Active
- 2015-09-08 WO PCT/CN2015/089105 patent/WO2017004883A1/zh not_active Ceased
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1100260A1 (en) * | 1999-11-12 | 2001-05-16 | STMicroelectronics S.r.l. | Spatio-temporal filtering method for noise reduction during pre-processing of picture sequences in video encoders |
| US7394856B2 (en) * | 2003-09-19 | 2008-07-01 | Seiko Epson Corporation | Adaptive video prefilter |
| WO2007071172A1 (en) * | 2005-12-21 | 2007-06-28 | Hong Kong Applied Science and Technology Research Institute Co. Ltd | Auto-regressive method and filter for denoising images and videos |
| CN101448077A (zh) * | 2008-12-26 | 2009-06-03 | 四川虹微技术有限公司 | 一种自适应视频图像3d降噪方法 |
| CN103024248A (zh) * | 2013-01-05 | 2013-04-03 | 上海富瀚微电子有限公司 | 运动自适应的视频图像降噪方法及其装置 |
| CN103533214A (zh) * | 2013-10-01 | 2014-01-22 | 中国人民解放军国防科学技术大学 | 一种基于卡尔曼滤波和双边滤波的视频实时去噪方法 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104994397A (zh) | 2015-10-21 |
| CN104994397B (zh) | 2018-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wu et al. | Learned block-based hybrid image compression | |
| CN107027029B (zh) | 基于帧率变换的高性能视频编码改进方法 | |
| CN105847797B (zh) | 图像解码方法 | |
| CN101478678B (zh) | 基于感兴趣区域运动补偿时域滤波方法 | |
| CN107820085B (zh) | 一种基于深度学习的提高视频压缩编码效率的方法 | |
| CN110087087A (zh) | Vvc帧间编码单元预测模式提前决策及块划分提前终止方法 | |
| CN106170093B (zh) | 一种帧内预测性能提升编码方法 | |
| JP2010218549A5 (zh) | ||
| Liu et al. | Enlarged motion-aware and frequency-aware network for compressed video artifact reduction | |
| CN101379827A (zh) | 基于边缘的时空滤波方法和装置 | |
| TW200535717A (en) | Directional video filters for locally adaptive spatial noise reduction | |
| US20250392747A1 (en) | Inter coding using deep learning in video compression | |
| WO2017004889A1 (zh) | 基于jnd因子的超像素高斯滤波预处理方法 | |
| Meng et al. | Enhancing quality for VVC compressed videos by jointly exploiting spatial details and temporal structure | |
| CN108028937A (zh) | 视频运动补偿装置和方法 | |
| WO2017004883A1 (zh) | 基于时域信息的自适应视频预处理方法 | |
| CN101841723B (zh) | 基于jnd和ar模型的感知视频压缩方法 | |
| CN1901620A (zh) | 一种基于运动检测和自适应滤波的视频图像降噪方法 | |
| JP2008541672A5 (zh) | ||
| JP2025508549A5 (zh) | ||
| CN117750020B (zh) | 可学习视频编码方法、系统、设备及存储介质 | |
| Li et al. | Depth video inter coding based on deep frame generation | |
| CN120088133B (zh) | 一种基于扩散模型的任意尺度视频超分辨率方法 | |
| CN101895762A (zh) | 基于零检测和矢量滤波的帧频提升算法 | |
| CN101313582A (zh) | 使用各种运动模型的编码器辅助式帧速率提升转换 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 15897539 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 12205A DATED 04/05/2018) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 15897539 Country of ref document: EP Kind code of ref document: A1 |










