WO2012031498A1 - 视频编码的滤波方法及装置 - Google Patents

视频编码的滤波方法及装置 Download PDF

Info

Publication number
WO2012031498A1
WO2012031498A1 PCT/CN2011/076286 CN2011076286W WO2012031498A1 WO 2012031498 A1 WO2012031498 A1 WO 2012031498A1 CN 2011076286 W CN2011076286 W CN 2011076286W WO 2012031498 A1 WO2012031498 A1 WO 2012031498A1
Authority
WO
WIPO (PCT)
Prior art keywords
macroblock
filtering
current macroblock
interest
threshold
Prior art date
Application number
PCT/CN2011/076286
Other languages
English (en)
French (fr)
Inventor
曾幸
鲁晓牧
吴钊
王宁
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2012031498A1 publication Critical patent/WO2012031498A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/85Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/17Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
    • H04N19/176Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/80Details of filtering operations specially adapted for video compression, e.g. for pixel interpolation

Definitions

  • the present invention relates to the field of communications, and in particular to a filtering method and apparatus for video encoding.
  • BACKGROUND OF THE INVENTION The tremendous advances in network technology and video coding technology have made it convenient for people to utilize various terminals for video conversations.
  • the image quality of video calls is currently unsatisfactory. Specifically, it is blurred in image, square in image, color distortion, and so on.
  • it is necessary to improve the quality of video images at low bit rates.
  • a video coding method based on the region of interest has been proposed.
  • the video coding method based on the region of interest distinguishes a portion of the video image that the user is interested in (for example, a face in a video call) from other image portions, and uses different coding methods for both, thereby improving the region of interest.
  • Image quality improve user experience. For example, in a mobile videophone, since the image size is small, the user's face is usually the most interesting part, and if the sharpness of the face in the encoded image can be improved, the subjective feeling of the user can be greatly improved.
  • encoding based on the region of interest improves the image quality of the region of interest by the quantization parameter of the region of interest in the Lowering.
  • a primary object of the present invention is to provide a filtering scheme for video encoding to at least solve the problem of low image quality based on encoding of a region of interest in the related art described above.
  • a filtering method for video encoding comprising: performing a region of interest detection on a video image before encoding each video image to obtain a sense of a macroblock unit.
  • the method further comprises: determining that the current macroblock belongs to the non-interest region and the current macroblock uses interframe coding; The current macroblock is filtered according to the pixel variance of the current macroblock, the variance of the residual coefficient, and the motion intensity.
  • the current macroblock is filtered according to the pixel variance, the residual coefficient variance and the motion intensity of the current macroblock: calculating an integrated filtering factor according to the pixel variance, the residual coefficient variance and the motion intensity of the current macroblock; Whether the filtering factor is greater than the first threshold, and if so, the residual filtering is not performed on the current macroblock; otherwise, the current macroblock is residual filtered using a predetermined filtering template.
  • performing residual filtering on the current macroblock using a predetermined filtering template comprises:
  • the pre-macroblock performs residual filtering; when the integrated filtering factor is greater than or equal to the second threshold, the filtering template is used 1 1 1
  • T F4" 1 1 1 Performs residual filtering on the current macroblock.
  • the method further comprises: filtering discrete cosine transform coefficients of each macroblock of the non-interest region.
  • filtering the discrete cosine transform coefficients of each macroblock of the non-interest region includes:
  • the discrete cosine transform coefficients are filtered, wherein A is a discrete cosine transform coefficient before filtering, and a discrete cosine transform coefficient after ⁇ filtering.
  • spatially filtering each macroblock according to the size of the distance comprises: determining whether the distance from the current macroblock to the central macroblock is greater than a third threshold, and if so, spatially filtering the current macroblock by using a predetermined filtering template, Otherwise, the current macroblock is not spatially filtered.
  • spatial filtering the current macroblock by using a predetermined filtering template comprises: using a filtering template 1 5 1 when the distance from the current macroblock to the central macroblock is greater than a third threshold and less than or equal to a fourth threshold.
  • T F1" 5 25 5 spatially filters the current macroblock; when the current macroblock to the central macroblock is significantly larger than 1 5 1
  • a video encoding filtering apparatus comprising: a detecting module configured to perform a region of interest detection on a video image before encoding each frame of video image, Obtaining a region of interest in units of macroblocks; a calculation module configured to calculate a distance of each macroblock in the video image to a central macroblock of the region of interest; a spatial filtering module, configured to follow each macroblock to the central macroblock The distance is spatially filtered separately for each macroblock.
  • the apparatus further comprises: a determining module, configured to determine that the current macroblock belongs to the non-interest area and the current macroblock uses inter-frame coding; and the filtering module is configured to adjust the pixel variance, the residual coefficient variance and the current macroblock according to the current macroblock The intensity of the motion filters the current macroblock.
  • a determining module configured to determine that the current macroblock belongs to the non-interest area and the current macroblock uses inter-frame coding
  • the filtering module is configured to adjust the pixel variance, the residual coefficient variance and the current macroblock according to the current macroblock The intensity of the motion filters the current macroblock.
  • FIG. 1 is a flowchart of a filtering method of video encoding according to an embodiment of the present invention
  • FIG. 2 is a schematic diagram of a filtering method according to an embodiment of the present invention
  • FIG. 3 is a spatial pixel domain filtering according to an embodiment of the present invention
  • FIG. 4 is a flowchart of a method for filtering motion compensation residuals according to an embodiment of the present invention
  • FIG. 5 is a structural block diagram of a filtering apparatus for video encoding according to an embodiment of the present invention
  • FIG. 7 is a block diagram showing a preferred structure of a filtering apparatus for a video encoding according to an embodiment of the present invention
  • FIG. 7 is a block diagram showing a preferred configuration of a filtering module 64 according to an embodiment of the present invention
  • FIG. 9 is a schematic diagram of a spatial pixel domain filtering apparatus according to an embodiment of the present invention
  • FIG. 10 is a schematic diagram of a motion compensation residual filtering apparatus according to an embodiment of the present invention
  • Embodiment 1 An embodiment of the present invention provides a filtering method for video coding. 1 is a flowchart of a method for filtering video coding according to an embodiment of the present invention. As shown in FIG.
  • the method includes: Step S102: performing a region of interest detection on a video image before encoding each video image,
  • the macroblock is the unit of interest;
  • step S106 spatially filter each macroblock according to a distance from each macroblock to the central macroblock.
  • This embodiment spatially filters each macroblock according to the distance from each macroblock to the center of the region of interest, which can reduce the high frequency information of the image in the non-interest region, thereby enabling more coding bits to be used for the region of interest, thereby improving The image quality of the region of interest.
  • the macroblock may be subjected to residual filtering. For example, if the current macroblock belongs to a non-interest region and the current macroblock uses inter-frame coding, the current macroblock may be filtered according to the pixel variance, the residual coefficient variance, and the motion intensity of the current macroblock.
  • This embodiment uses motion compensated residual filtering for non-interest regions, which reduces the number of bits required for non-region of interest coding, thereby enabling more coded bits to be used for the region of interest and improving the region of interest. quality.
  • the integrated filter factor may be calculated according to the pixel variance, the residual coefficient variance and the motion intensity of the current macroblock, and then it is determined whether the integrated filter factor is greater than the first threshold, and if so, the current macro is not
  • the block performs residual filtering, otherwise, the current macroblock is residual filtered using a predetermined filtering template.
  • the value of the first threshold can be set according to actual conditions. This embodiment controls whether or not filtering is required by an integrated filtering factor, and is highly practical. For example, when the integrated filter factor is greater than or equal to the first threshold and less than the second threshold, it can be used.
  • Filter template ⁇ F3" 2 4 2 Perform residual filtering on the current macroblock, when the integrated filter factor is greater than or equal
  • the value of the first threshold can be set according to the actual situation. This embodiment gives a preferred filtering template with high operability.
  • discrete cosine transform coefficients for each macroblock of the non-interest region may also be filtered. For example, you can use a formula 16 15 14 12 10 8 6 4
  • the string transform coefficients are filtered, wherein A is a discrete cosine transform coefficient before filtering, and 2 is a filtered discrete cosine transform coefficient.
  • This embodiment can ensure that more coded bits are used for the region of interest by multiple filtering, which improves the image quality of the region of interest.
  • spatial filtering of each macroblock according to the size of the distance may be implemented by: first, determining whether the distance from the current macroblock to the central macroblock is greater than a third threshold, and if so, using a predetermined filtering template pair The current macroblock is spatially filtered, otherwise the current macroblock may not be spatially filtered.
  • This embodiment determines whether filtering is needed by the distance from the current macroblock to the central macroblock, and the filtering effect is better. Wherein, when the macroblock of the current macroblock to the central macroblock is greater than the third threshold and less than or equal to the fourth threshold, filtering template filtering may be used; when the current macroblock
  • the filter template 1 3 1 can be used.
  • T F2— 3 8 3 spatially filter the current macroblock; when the distance from the current macroblock to the central macroblock is greater than 1 3 1
  • Inter-filtering use filter template when the distance from the current macroblock to the central macroblock is greater than the sixth threshold
  • T F4 1 1 1 Spatially filters the current macroblock. Among them, each threshold can be based on the actual situation 1 1 1
  • Embodiment 2 This embodiment provides a specific implementation manner of a filtering method for video coding according to an embodiment of the present invention.
  • the method includes: Step 1: Performing a region of interest on an input image to obtain a macroblock The area of interest of the unit.
  • Step 2 Perform spatial filtering on the entire frame of the image.
  • 2 is a schematic diagram of a filtering method according to an embodiment of the present invention. As shown in FIG. 2, spatial pixel domain filtering (ie, spatial domain filtering) occurs before encoding, and belongs to video encoding pre-processing.
  • FIG. 3 is a flowchart of a spatial pixel domain filtering method according to an embodiment of the present invention.
  • the spatial domain filtering includes the following steps: Step S301: Calculating each macroblock in the image to the central macroblock of the region of interest distance. The distance between the current macroblock and the central macroblock of the region of interest can be calculated using equation (1). The macro distance between the current macroblock and the central macroblock of the region of interest is:
  • Step S302 performing spatial filtering of different intensities for each macroblock according to the size of the distance. It is judged whether the distance from each macroblock to the central macroblock is greater than the threshold D1, and if so, step S303 is performed, otherwise, the process proceeds to step S305.
  • Step S303 For a macroblock whose distance to the central macroblock is greater than the threshold D1, an appropriate filtering template is selected according to the distance. Here you can choose different 3 ⁇ 3 filter templates. It should be noted that other filtering templates can also be selected, for example, selecting a 5 X 5 filtering template.
  • the corresponding filter template For a macroblock whose distance is greater than D 1 and less than or equal to D2, the corresponding filter template may be T F1 ; for a macroblock whose distance is greater than D2 and less than or equal to D3, the corresponding filter template may be T F2 ; D3 is less than or equal to D4 macroblock, and the corresponding filter template may be T F3 ; for a macroblock larger than D4, the corresponding filter template may be T F4 . among them,
  • Step S304 filtering the pixels of the macro block by using the selected filtering template.
  • Step S305 determining whether the current macroblock is the last macroblock, and if yes, the process ends; otherwise, returning to step 4 to gather S301.
  • D1 takes the value 2
  • D2 takes the value 3
  • D3 takes the value 4
  • D4 takes the value 5.
  • the third step performing motion compensation residual filtering on the inter-coded non-region of interest macroblock. As shown in Fig. 2, the process occurs after motion compensation and before Discrete Cosine Transform (DCT) conversion, which is used to reduce image noise in non-region of interest and high-frequency information that is insensitive to human eyes. , thereby reducing the number of bits required for encoding of macroblocks in non-region of interest.
  • DCT Discrete Cosine Transform
  • Step S401 calculating a pixel variance, a residual coefficient variance, and an exercise intensity of a current macroblock
  • Step S402 calculating an integrated filter factor according to formula (2): Where Er denotes the residual coefficient variance of the macroblock; Ec denotes the macroblock pixel variance; MVx, and MVy denote the motion vector horizontal and vertical components, respectively; MVmax denotes the motion search maximum range; a denotes the weighting factor.
  • the weighting factor a in equation (2) can take a value of 2; both the motion vector and the maximum motion search range in the equation can be obtained from the motion estimation and compensator.
  • step S403 it is determined whether the integrated filter factor is less than the threshold T1. If yes, the flow ends. Otherwise, the process proceeds to step S404.
  • Step S404 selecting a corresponding filtering template according to the integrated filtering factor. For example, different 3 x3 filter templates are selected based on the integrated filter factor.
  • Step S405 Perform residual coefficient filtering on the current macroblock by using the selected filtering template. The larger the integrated filter factor, the greater the filter strength. After the ideal motion estimation, the main components of the motion compensation residual are noise and high frequency information that is insensitive to the human eye.
  • Step 4 Filter the DCT coefficients of the macroblocks in the non-interest area. As shown in Figure 2, this process occurs after the DCT transform, before quantization.
  • the transform domain filtering is to filter the DCT coefficients of the 8x8 blocks of the inter-frame and intra-coded macroblocks of the non-interest region, that is, to perform matrix quantization on the DCT coefficients.
  • the luminance DCT coefficient and the chrominance DCT coefficient of the macroblock the same processing may be performed, or only the luminance DCT coefficients may be processed without processing the chrominance DCT coefficients.
  • the DCT coefficients can be filtered using equation (3): Where A is the discrete cosine transform coefficient before filtering, A is the filtered discrete cosine transform coefficient, and A and A are both 8 ⁇ 8 DCT coefficient matrices.
  • Step 5 At the end of the current image encoding, go to the first step to continue the next frame encoding.
  • This embodiment utilizes human visual characteristics, and by combining three filtering methods, appropriately reducing the image details of the non-region of interest, and using the saved bits for encoding the region of interest, thereby improving the image quality of the region of interest, Low-rate video calls can effectively improve the user experience.
  • the second step, the third step, and the fourth step may be used at the same time, or only one or several steps may be used, that is, only one or several filtering methods are used.
  • just use the second step then go to the fifth step, or use the second and third steps instead of the fourth step.
  • the effect of improving the image quality of the region of interest can be achieved by using only one or a few of them, but the effect of using the second step, the third step and the fourth step simultaneously is good.
  • the device embodiment corresponds to the foregoing method embodiment, and the embodiment of the present invention further provides a video encoding filter device.
  • FIG. 5 is a structural block diagram of a filtering apparatus for video encoding according to an embodiment of the present invention. As shown in FIG.
  • the apparatus includes: a detecting module 52 configured to perform a region of interest detection on a video image before encoding each video image. Obtaining a region of interest in units of macroblocks; computing module 54, coupled to detection module 52, configured to calculate a distance of each macroblock in the video image to a central macroblock of the region of interest; spatial filtering module 56, coupled to the computation The module 54, is configured to spatially filter each macroblock according to the distance from each macroblock to the central macroblock.
  • 6 is a block diagram of a preferred structure of a video encoding filter apparatus according to an embodiment of the present invention.
  • the apparatus further includes: a determining module 62 configured to determine that the current macroblock belongs to a non-interest area and And the current macroblock uses inter-frame coding; the filtering module 64 is coupled to the determining module 62, and is configured to filter the current macroblock according to the pixel variance, the residual coefficient variance, and the motion intensity of the current macroblock.
  • 7 is a block diagram of a preferred structure of the filtering module 64 according to the embodiment of the present invention.
  • the filtering module 64 includes: a calculating sub-module 72 configured to calculate the variance of the pixel, the variance of the residual coefficient, and the motion intensity according to the current macroblock.
  • the filtering sub-module 74 is coupled to the calculation sub-module 72, and is configured to determine whether the integrated filtering factor is greater than the first threshold. If yes, the residual filtering is not performed on the current macroblock. Otherwise, the predetermined filtering template is used to The macroblock performs residual filtering.
  • the filtering sub-module 74 is configured to have an integrated filtering factor greater than or equal to the first threshold and less than
  • FIG. 8 is another block diagram of another preferred structure of a video encoding filter apparatus according to an embodiment of the present invention. As shown in FIG. 8, the apparatus further includes: a coefficient filtering module 82 configured to each macroblock of a non-interest area. The discrete cosine transform coefficients are filtered.
  • the current macroblock performs spatial filtering; when the distance from the current macroblock to the central macroblock is greater than the fourth threshold and less than or
  • the spatial filtering module 56 uses the filtering mode.
  • the embodiment of the present invention further provides three filtering devices, which respectively perform the above three filtering functions, and correspond to corresponding modules of the filtering device.
  • 9 is a schematic diagram of a spatial pixel domain filtering apparatus according to an embodiment of the present invention.
  • the apparatus includes: a distance determiner 92 corresponding to the calculation module 52, configured to be based on coordinates and interest of each macroblock. The coordinates of the zone i or the central macroblock determine 3 macroranges from each macroblock to the central macroblock; a filter template determiner 94, coupled to the distance determiner 92, is arranged to select a corresponding filter template based on the distance determined by the distance determiner 92.
  • FIG. 10 is a schematic diagram of a motion compensation residual filtering apparatus according to an embodiment of the present invention. As shown in FIG.
  • the apparatus includes: a residual coefficient variance determiner 1002, configured to determine a residual coefficient variance of a current macroblock;
  • the pixel variance determiner 1004 is configured to determine a pixel variance of the current macroblock;
  • the motion intensity determiner 1006 is configured to determine the motion strength of the current macroblock according to the motion vector;
  • the integrated filter factor determiner 1008 is coupled to the variance of the residual coefficient
  • the determiner 1002, the pixel variance determiner 1004, and the exercise intensity determiner 1006, corresponding to the calculation sub-module 72, are set to determine the result determined by the residual coefficient variance determiner 1002, the pixel variance determiner 1004, and the exercise intensity determiner 1006.
  • Determining an integrated filter factor a filter template determiner 1010 coupled to the integrated filter factor determiner 1008, configured to perform overall filtering The factor determines a corresponding filter template; a convolver 1012, coupled to the filter template determiner 1010, in conjunction with the filter template determiner 1010, implements the function of the filter sub-module 74, and sets the residual coefficient of the current macroblock according to the determined filter template. Filtering is performed.
  • 11 is a schematic diagram of a transform domain filtering apparatus according to an embodiment of the present invention. As shown in FIG.
  • the apparatus includes: a filter coefficient matrix generator 1102 configured to generate a filter coefficient matrix; a multiplier 1104 coupled to the filter coefficient A matrix generator 1102, corresponding to the coefficient filtering module 82, is arranged to multiply the matrix of transform coefficients by the matrix of generated filter coefficients; a shifter 1106, coupled to the multiplier 1104, is arranged to generate the filtered transform coefficients.
  • the present invention reduces the noise in the video by using the above filtering technique, appropriately reduces the details of the non-interest region, and reduces the compression artifact, thereby using more bits for encoding the interest.
  • the area improves the image quality of the area of interest.
  • modules or steps of the present invention can be implemented by a general-purpose computing device, which can be concentrated on a single computing device or distributed over a network composed of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device, such that they may be stored in the storage device by the computing device and, in some cases, may be different from the order herein.
  • the steps shown or described are performed, or they are separately fabricated into individual integrated circuit modules, or a plurality of modules or steps are fabricated as a single integrated circuit module.
  • the invention is not limited to any specific combination of hardware and software.
  • the above is only the preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes can be made to the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the scope of the present invention are intended to be included within the scope of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

本发明公开了一种视频编码的滤波方法及装置,该方法包括:在编码每帧视频图像前对视频图像进行感兴趣区域检测,得到以宏块为单位的感兴趣区域;计算视频图像中每个宏块到感兴趣区域的中心宏块的距离;按照各个宏块到中心宏块的距离分别对各个宏块进行空间滤波。本发明达到了提高感兴趣区域图像质量的效果。

Description

视频编码的滤波方法^置 技术领域 本发明涉及通信领域, 具体而言, 涉及一种视频编码的滤波方法及装置。 背景技术 网络技术和视频编码技术的巨大进步使得人们可以方便地利用各种终端 进行视频对话。 但是, 由于信道带宽的限制, 目前视频通话的图像质量普遍不 能令人满意。 具体表现在图像模糊、 图像出现方块、 色彩失真等。 要想使用户 得到更好的视频通话体验, 必须改善低码率下的视频图像质量。 在目前通信带宽相对不足的情况下, 有人提出了基于感兴趣区域的视频编 码方法。 基于感兴趣区域的视频编码方法将视频图像中用户比较关注的部分 (例如, 视频通话中的人脸) 和其他图像部分区分开, 对两者釆用不同的编码 方法, 从而改善感兴趣区域的图像质量, 提高用户的感受。 例如, 在手机可视 电话中, 由于图像尺寸小, 用户的脸部通常是最受关注的部分, 如果能够提高 编码图像中脸部的清晰度, 就可以很大地提高使用者的主观感受。 通常, 基于感兴趣区域的编码会通过降氏感兴趣区域的量化参数来提高感 兴趣区域的图像质量。 由于各视频编码标准中对相邻宏块的量化参数改变范围 限定的很小, 仅仅通过降低量化参数的方法对感兴趣区域的图像质量改善有 限。 因此, 需要釆用其他的编码技术来提高感兴趣区域的图像质量。 发明内容 本发明的主要目的在于提供视频编码的滤波方案, 以至少解决上述的相关 技术中基于感兴趣区域的编码的图像质量不高的问题。 为了实现上述目的, 居本发明的一个方面, 提供了一种视频编码的滤波 方法, 该方法包括: 在编码每帧视频图像前对视频图像进行感兴趣区域检测, 得到以宏块为单位的感兴趣区域; 计算视频图像中每个宏块到感兴趣区域的中 心宏块的距离; 按照各个宏块到中心宏块的距离分别对各个宏块进行空间滤 波。 优选地, 在按照各个宏块到中心宏块的距离分别对各个宏块进行空间滤波 之后, 该方法还包括: 确定当前宏块属于非感兴趣区域并且当前宏块釆用帧间 编码; 才艮据当前宏块的像素方差、 残差系数方差和运动强度对当前宏块进行滤 波。 优选地, 才艮据当前宏块的像素方差、 残差系数方差和运动强度对当前宏块 进行滤波包括: 根据当前宏块的像素方差、 残差系数方差和运动强度计算综合 滤波因子; 判断综合滤波因子是否大于第一门限, 如果是, 则不对当前宏块进 行残差滤波, 否则, 使用预定的滤波模板对当前宏块进行残差滤波。 优选地, 使用预定的滤波模板对当前宏块进行残差滤波包括: 当综合滤波
1 2 1 因子大于或等于第一门限且小于第二门限时,使用滤波模板 TF3= 2 4 2 对当
1 2 1 前宏块进行残差滤波; 当综合滤波因子大于或等于第二门限时, 使用滤波模板 1 1 1
T F4" 1 1 1 对当前宏块进行残差滤波。
1 1 1 优选地, 在按照各个宏块到中心宏块的距离分别对各个宏块进行空间滤波 之后, 该方法还包括: 对非感兴趣区域的每个宏块的离散余弦变换系数进行滤 波。 优选地, 对非感兴趣区域的每个宏块的离散余弦变换系数进行滤波包括:
使用公式 , = a 对非感兴趣区域的每个宏块
Figure imgf000003_0001
的离散余弦变换系数进行滤波, 其中, A是滤波前的离散余弦变换系数, Δ疋 滤波后的离散余弦变换系数。 优选地, 按照距离的大小对每个宏块进行空间滤波包括: 判断当前宏块到 中心宏块的距离是否大于第三门限, 如果是, 则使用预定的滤波模板对当前宏 块进行空间滤波, 否则, 不对当前宏块进行空间滤波。 优选地, 使用预定的滤波模板对当前宏块进行空间滤波包括: 当当前宏块 到中心宏块的距离大于第三门限且小于或等于第四门限时, 使用滤波模板 1 5 1
T F1" 5 25 5 对当前宏块进行空间滤波; 当当前宏块到中心宏块的 巨离大于 1 5 1
1 3 1
第四门限且小于或等于第五门限时,使用滤波模板 TF2= 3 8 3 对当前宏块进
1 3 1
行空间滤波; 当当前宏块到中心宏块的距离大于第五门限且小于或等于第六门
1 2 1
限时, 使用滤波模板 T F3" 2 4 2 对当前宏块进行空间滤波; 当当前宏块到中
1 2 1
1 1 1
心宏块的距离大于第六门限时, 使用滤波模板 TF4= 1 1 1 对当前宏块进行:
1 1 1
间滤波。 为了实现上述目的, 才艮据本发明的另一方面, 提供了一种视频编码的滤波 装置, 该装置包括: 检测模块, 设置为在编码每帧视频图像前对视频图像进行 感兴趣区域检测, 得到以宏块为单位的感兴趣区域; 计算模块, 设置为计算视 频图像中每个宏块到感兴趣区域的中心宏块的距离; 空间滤波模块, 设置为按 照各个宏块到中心宏块的距离分别对各个宏块进行空间滤波。 优选地, 该装置还包括: 确定模块, 设置为确定当前宏块属于非感兴趣区 域并且当前宏块釆用帧间编码; 滤波模块, 设置为根据当前宏块的像素方差、 残差系数方差和运动强度对当前宏块进行滤波。 通过本发明, 釆用按照各宏块到感兴趣区域中心的距离的不同对各个宏块 进行空间滤波的方式, 解决了相关技术中基于感兴趣区域的编码的图像质量不 高的问题, 进而达到了提高感兴趣区域图像质量的效果。 附图说明 此处所说明的附图用来提供对本发明的进一步理解, 构成本申请的一部 分, 本发明的示意性实施例及其说明用于解释本发明, 并不构成对本发明的不 当限定。 在附图中: 图 1是根据本发明实施例的视频编码的滤波方法的流程图; 图 2是 居本发明实施例的滤波方法的示意图; 图 3是根据本发明实施例的空间像素域滤波方法的流程图; 图 4是 居本发明实施例的运动补偿残差滤波的方法的流程图; 图 5是才艮据本发明实施例的视频编码的滤波装置的结构框图; 图 6是才艮据本发明实施例的视频编码的滤波装置的一个优选结构框图; 图 7是 居本发明实施例的滤波模块 64的优选的结构框图; 图 8是 居本发明实施例的视频编码的滤波装置的另一个优选结构框图; 图 9是根据本发明实施例的空间像素域滤波装置的示意图; 图 10是才艮据本发明实施例的运动补偿残差滤波装置的示意图; 以及 图 11是才艮据本发明实施例的变换域滤波装置的示意图。 具体实施方式 下文中将参考附图并结合实施例来详细说明本发明。 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组合。 实施例一 本发明实施例提供了一种视频编码的滤波方法。 图 1是根据本发明实施例 的视频编码的滤波方法的流程图, 如图 1所示, 该方法包括: 步骤 S 102, 在编码每帧视频图像前对视频图像进行感兴趣区域检测, 得到 以宏块为单位的感兴趣区域; 步骤 S 104, 计算视频图像中每个宏块到该感兴趣区域的中心宏块的距离; 步骤 S 106,按照各个宏块到该中心宏块的距离分别对各个宏块进行空间滤 波。 该实施例根据各宏块到感兴趣区域中心的距离对各个宏块进行空间滤波, 这样可以减少非感兴趣区域的图像高频信息, 从而能够将更多的编码比特用于 感兴趣区域, 提高了感兴趣区域图像质量。 在本发明的一个实现方式中, 在按照各个宏块到中心宏块的距离分别对各 个宏块进行空间滤波之后, 可以对宏块进行残差滤波。 例如, 如果当前宏块属 于非感兴趣区域并且该当前宏块釆用了帧间编码, 则可以根据当前宏块的像素 方差、 残差系数方差和运动强度对该当前宏块进行滤波。 该实施例对非感兴趣 区域釆用了运动补偿残差滤波, 减少了非感兴趣区域编码所需的比特数, 从而 能够将更多的编码比特用于感兴趣区域, 提高了感兴趣区域图像质量。 在具体实现过程中, 可以才艮据当前宏块的像素方差、 残差系数方差和运动 强度计算综合滤波因子, 然后, 判断该综合滤波因子是否大于第一门限, 如果 是, 则不对该当前宏块进行残差滤波, 否则, 使用预定的滤波模板对该当前宏 块进行残差滤波。 其中, 第一门限的值可以根据实际情况进行设定。 该实施例 通过综合滤波因子控制是否需要滤波, 实用性较强。 例如, 当综合滤波因子大于或等于第一门限且小于第二门限时, 可以使用
1 2 1
滤波模板 τ F3" 2 4 2 对该当前宏块进行残差滤波,当综合滤波因子大于或等
1 2 1
1 1 1
于第二门限时, 可以使用滤波模板 TF4= 1 1 1 对该当前宏块进行残差滤波。
1 1 1
其中, 第一门限的值可以才艮据实际情况进行设定。 该实施例给出了优选的滤波 模板, 可操作性强。 在本发明的另一个实现方式中, 还可以对非感兴趣区域的每个宏块的离散 余 弦 变 换 系 数 进 行 滤 波 。 例 如 , 可 以 使 用 公 式 16 15 14 12 10 8 6 4
15 14 12 10 8 6 4 2
14 12 10 8 6 4 2 2
12 10 8 6 4 2 2 2
对非感兴趣区域的每个宏块的离散余
16 10 8 6 4 2 2 2 1
8 6 4 2 2 2 1 1
6 4 2 2 2 1 1 1
4 2 2 2 1 1 1 1
弦变换系数进行滤波, 其中, A是滤波前的离散余弦变换系数, 2是滤波后的 离散余弦变换系数。 该实施例通过多重滤波, 能够保证将更多的编码比特用于 感兴趣区域, 提高了感兴趣区域图像质量。 优选地, 按照距离的大小对每个宏块进行空间滤波可以通过以下方式实 现: 首先, 判断当前宏块到中心宏块的距离是否大于第三门限, 如果是, 则可 以使用预定的滤波模板对该当前宏块进行空间滤波, 否则, 可以不对该当前宏 块进行空间滤波。 该实施例通过当前宏块到中心宏块的距离判断是否需要滤 波, 滤波效果更好。 其中, 当当前宏块到中心宏块的 ϋ巨离大于第三门限且小于或等于第四门限 时, 可以使用滤波模板 滤波; 当当前宏块
Figure imgf000007_0001
到中心宏块的距离大于第四门限且小于或等于第五门限时, 可以使用滤波模板 1 3 1
T F2— 3 8 3 对当前宏块进行空间滤波; 当当前宏块到中心宏块的距离大于第 1 3 1
1 2 1
五门限且小于或等于第六门限时,使用滤波模板 TF3= 2 4 2 对当前宏块进行
1 2 1
间滤波; 当当前宏块到中心宏块的距离大于第六门限时, 使用滤波模板
1 1 1
T F4 1 1 1 对当前宏块进行空间滤波。 其中, 各个门限值都可以根据实际情 1 1 1
况设定。 该实施例给出了优选的滤波模板, 可操作性强 , 实施例二 本实施例给出了一种才艮据本发明实施例的视频编码的滤波方法具体实现 方式, 该方法包括: 第一步: 对输入图像进行感兴趣区域检测, 得到以宏块为单位的感兴趣区 域。 第二步: 对整帧图像进行空域滤波。 图 2是根据本发明实施例的滤波方法 的示意图, 如图 2所示, 空间像素域滤波 (即, 空域滤波) 发生在编码之前, 属于视频编码前处理。 空域滤波的作用在于减少非感兴趣区域的图像高频信 息, 以便将更多的编码比特用于感兴趣区域, 同时, 保证了感兴趣区域和非感 兴趣区域图像质量的平滑过渡。 对于宏块的亮度分量和色度分量, 既可以进行 相同处理, 也可以只对亮度分量进行处理, 而不对色度分量进行处理。 图 3是根据本发明实施例的空间像素域滤波方法的流程图, 如图 3所示, 空域滤波包括以下步 4聚: 步骤 S301 , 计算图像中每个宏块到感兴趣区域中心宏块的距离。 当前宏块 与感兴趣区域中心宏块的距离可以使用公式 ( 1 ) 计算。 当前宏块与感兴趣区 域中心宏块的 巨离为:
D(x, _y) = |_y - _yo| / 2 + |x - xo| (1) 其中, x、 y分别表示宏块的水平与垂直坐标; x。、 y。分别表示感兴趣区域 中心宏块的水平与垂直坐标, 符号" I |"表示对其中的数取绝对值。 步骤 S302 ,根据距离的大小对每个宏块进行不同强度的空间滤波。 判断各 个宏块到中心宏块的距离是否大于阈值 D 1 , 如果是, 则进行步骤 S303 , 否则, 进入步骤 S305。 步骤 S303 , 对到中心宏块的距离大于阈值 D 1的宏块, 才艮据距离选取合适 的滤波模板。 这里可以选择不同的 3 χ3滤波模板。 需要注意的是, 也可以选择 其他滤波模板, 例如选择 5 X 5滤波模板。 对于距离大于 D 1小于等于 D2的宏块, 相应的滤波模板可以为 TF1; 对于 距离大于 D2小于等于 D3的宏块, 相应的滤波模板可以为 TF2; 对于距离大于 D3小于等于 D4的宏块, 相应的滤波模板可以为 TF3; 对于 巨离大于 D4的宏 块, 相应的滤波模板可以为 TF4。 其中,
Figure imgf000009_0001
步骤 S304, 使用选取的滤波模板对宏块的像素进行滤波。 步骤 S305 , 判断当前的宏块是否为最后一个宏块, 如果是, 则流程结束, 否则, 返回步 4聚 S301。 其中, 对于距离感兴趣区域中心宏块小于等于阈值 D1 的宏块, 不需要进 行空 i或滤波。 上述阈值的选择主要与视频图像的尺寸有关。 例如, 对于 3G视频通话典 型的四分之一通用中间格式 ( Quarter common intermediate format, 简称为 QCIF ) 尺寸, Dl取值为 2, D2取值为 3 , D3取值为 4, D4取值为 5。 这样, 离感兴趣区域中心越近的区域细节保留越多, 编码失真越小。 由于人眼对远离 感兴趣区域的图像细节关注度低, 这些区域的细节被滤波基本不会造成主观感 受的下降, 反而节省了比特用于编码感兴趣区域。 第三步: 对帧间编码的非感兴趣区域宏块进行运动补偿残差滤波。 如图 2 所示, 该过程发生在运动补偿之后, 以及离散余弦变化 ( Discrete Cosine Transform, 简称为 DCT ) 变换之前, 其作用在于减少非感兴趣区域的图像噪 声和人眼不敏感的高频信息, 从而减少非感兴趣区域宏块的编码所需比特。 对 于宏块的亮度残差和色度残差, 可以进行相同处理, 也可以只对亮度残差进行 处理, 而不对色度残差进行处理。 图 4是 居本发明实施例的运动补偿残差滤波的方法的流程图, 如图 4所 示, 该流程包括: 步骤 S401 , 计算出当前宏块的像素方差、 残差系数方差和运动强度; 步骤 S402 , 根据公式 (2)计算出综合滤波因子:
Figure imgf000010_0001
其中, Er表示宏块的残差系数方差; Ec表示宏块像素方差; MVx,和 MVy 分别表示运动矢量水平和垂直分量; MVmax表示运动搜索最大范围; a表示权 重因子。 以 3G视频通话典型的 48 Kbps而言, 公式 (2)中的权重因子 a取值可 以为 2; 该公式中的运动矢量和最大运动搜索范围都可以从运动估计与补偿器 中得到。 步骤 S403 , 判断综合滤波因子是否小于阈值 T1 , 如果是, 则流程结束, 否则, 进入步 4聚 S404。 步骤 S404 , 根据综合滤波因子选取相应的滤波模板。 例如, 根据综合滤波 因子选择不同的 3 x3滤波模板。 具体的, 对于综合滤波因子大于等于 T1且小 于 T2的宏块, 相应的滤波模板可以为 TF3; 对于综合滤波因子大于等于 T2的 宏块, 相应的滤波模板为可以为 TF4。 其中, 综合滤波因子的阈值 T1取值可以 为 0. 1 , T2取值可以为 0.5。 步骤 S405 , 使用选取的滤波模板对该当前宏块进行残差系数滤波。 综合滤 波因子越大, 则滤波强度越大。 在理想的运动估计后, 运动补偿残差的主要成 分是噪声和人眼不敏感的高频信息。 残差宏块方差越大, 表示可以减少的高频 信息越多; 而人眼对运动剧烈的图像不敏感, 可以对大运动宏块进行更强的滤 波。 第四步: 对非感兴趣区域宏块的 DCT 系数进行滤波。 如图 2所示, 该过 程发生在 DCT 变换之后, 量化之前。 变换域滤波是对非感兴趣区域的帧间和 帧内编码宏块的 8x8块的 DCT系数滤波, 即对 DCT系数进行矩阵量化。 对于 宏块的亮度 DCT系数和色度 DCT系数, 可以进行相同处理, 也可以只对亮度 DCT 系数进行处理, 而不对色度 DCT 系数进行处理。 可以釆用公式 (3 ) 对 DCT系数进行滤波:
Figure imgf000011_0001
其中, 其中, A是滤波前的离散余弦变换系数, A是滤波后的离散余弦变 换系数, 并且, A和 A都是 8 X 8的 DCT系数矩阵。 第五步: 当前图像编码结束, 转至第一步继续下一帧编码。 该实施例利用人眼视觉特性, 通过将三种滤波方式组合, 适当减少非感兴 趣区域的图像细节, 将节省下的比特用于编码感兴趣区域, 从而提高了感兴趣 区域的图像质量, 在低码率视频通话中可以有效地提高用户的体验。 需要注意的是, 在本实施例中, 第二步、 第三步和第四步可以同时使用, 也可以仅使用其中的一步或几步, 即, 仅使用一种或几种滤波方式。 例如, 仅 使用第二步, 然后就进入第五步, 或者, 使用第第二步和第三步, 而不使用第 四步。 仅使用其中的一步或几步也可以达到提高感兴趣区域的图像质量的效 果, 只是没有第二步、 第三步和第四步同时使用产生的效果好。 装置实施例 对应于上述方法实施例, 本发明实施例还提供了一种视频编码的滤波装 置。 图 5是根据本发明实施例的视频编码的滤波装置的结构框图,如图 5所示, 该装置包括: 检测模块 52 , 设置为在编码每帧视频图像前对视频图像进行感兴 趣区域检测, 得到以宏块为单位的感兴趣区域; 计算模块 54, 耦合至检测模块 52 , 设置为计算视频图像中每个宏块到感兴趣区域的中心宏块的距离; 空间滤 波模块 56, 耦合至计算模块 54, 设置为按照各个宏块到中心宏块的距离分别 对各个宏块进行空间滤波。 图 6是才艮据本发明实施例的视频编码的滤波装置的一个优选结构框图, 优 选地, 该装置还包括: 确定模块 62 , 设置为确定当前宏块属于非感兴趣区域并 且当前宏块釆用帧间编码; 滤波模块 64 , 耦合至确定模块 62 , 设置为才艮据当 前宏块的像素方差、 残差系数方差和运动强度对当前宏块进行滤波。 其中, 图 7是才艮据本发明实施例的滤波模块 64的优选的结构框图, 滤波 模块 64包括: 计算子模块 72 , 设置为根据当前宏块的像素方差、 残差系数方 差和运动强度计算综合滤波因子; 滤波子模块 74 , 耦合至计算子模块 72 , 设 置为判断综合滤波因子是否大于第一门限, 如果是, 则不对当前宏块进行残差 滤波, 否则, 使用预定的滤波模板对当前宏块进行残差滤波。 其中, 滤波子模块 74设置为在综合滤波因子大于或等于第一门限且小于
1 2 1
第二门限时, 使用滤波模板 TF3= 2 4 2 对当前宏块进行残差滤波; 当综合滤
1 2 1
1 1 1
波因子大于或等于第二门限时, 使用滤波模板 TF4= 1 1 1 对当前宏块进行残
1 1 1
差滤波。 图 8是 居本发明实施例的视频编码的滤波装置的另一个优选结构框图, 如图 8所示, 该装置还包括: 系数滤波模块 82 , 设置为对非感兴趣区域的每个 宏块的离散余弦变换系数进行滤波。
16 15 14 12 10 8 6 4
15 14 12 10 8 6 4 2
14 12 10 8 6 4 2 2
12 10 8 6 4 2 2 2 其中,系数滤波模块 82使用公式 5. 对
16 10 8 6 4 2 2 2 1
8 6 4 2 2 2 1 1
6 4 2 2 2 1 1 1
4 2 2 2 1 1 1 1 非感兴趣区域的每个宏块的离散余弦变换系数进行滤波, 其中, A是滤波前的 离散余弦变换系数, 2是滤波后的离散余弦变换系数。 优选地, 空间滤波模块 56设置为判断当前宏块到中心宏块的距离是否大 于第三门限, 如果是, 则使用预定的滤波模板对当前宏块进行空间滤波, 否则, 不对当前宏块进行空间滤波。 其中, 当当前宏块到中心宏块的距离大于第三门 1 5 1 限且小于或等于第四门限时, 空间滤波模块 56使用滤波模板 TF1= 5 25 5 对
1 5 1 当前宏块进行空间滤波; 当当前宏块到中心宏块的距离大于第四门限且小于或
1 3 1
等于第五门限时, 空间滤波模块 56使用滤波模板 TF2= 3 8 3 对当前宏块进
1 3 1
行空间滤波; 当当前宏块到中心宏块的距离大于第五门限且小于或等于第六门
1 2 1
限时,空间滤波模块 56使用滤波模板 TF3= 2 4 2 对当前宏块进行空间滤波;
1 2 1
当当前宏块到中心宏块的距离大于第六门限时, 空间滤波模块 56 使用滤波模
1 1 1
板 T F4" 1 1 1 对当前宏块进行空间滤波。
1 1 1 在具体应用过程中, 本发明实施例还提出了三种滤波装置, 分别完成上述 三种滤波功能, 并对应上述滤波装置的相应模块。 图 9是根据本发明实施例的空间像素域滤波装置的示意图, 如图 9所示, 该装置包括: 距离确定器 92 , 对应于计算模块 52 , 设置为根据每个宏块的坐 标和感兴趣区 i或中心宏块的坐标确定每个宏块到中心宏块的 3巨离; 滤波模板确 定器 94 , 耦合至距离确定器 92 , 设置为根据距离确定器 92确定的距离选择相 应的滤波模板; 卷积器 96 , 耦合至滤波模板确定器 94 , 与滤波模板确定器 94 联合实现空间滤波模块 56 的功能, 设置为将宏块的像素和确定的滤波模板进 行卷积得到滤波后的宏块像素。 图 10是才艮据本发明实施例的运动补偿残差滤波装置的示意图, 如图 10所 示, 该装置包括: 残差系数方差确定器 1002 , 设置为确定当前宏块的残差系数 方差; 像素方差确定器 1004, 设置为确定当前宏块的像素方差; 运动强度确定 器 1006 , 设置为才艮据运动矢量确定当前宏块的运动强度; 综合滤波因子确定器 1008 , 耦合至残差系数方差确定器 1002、 像素方差确定器 1004和运动强度确 定器 1006 , 对应于计算子模块 72 , 设置为才艮据残差系数方差确定器 1002、 像 素方差确定器 1004和运动强度确定器 1006确定的结果确定综合滤波因子; 滤 波模板确定器 1010, 耦合至综合滤波因子确定器 1008 , 设置为根据综合滤波 因子确定相应的滤波模板; 卷积器 1012 , 耦合至滤波模板确定器 1010 , 与滤 波模板确定器 1010联合实现滤波子模块 74的功能, 设置为根据确定的滤波模 板对当前宏块的残差系数进行滤波。 图 11是才艮据本发明实施例的变换域滤波装置的示意图, 如图 11所示, 该 装置包括: 滤波系数矩阵生成器 1102 ,设置为生成滤波系数矩阵;乘法器 1104 , 耦合至滤波系数矩阵生成器 1102 , 对应于系数滤波模块 82 , 设置为将变换系 数矩阵与生成的滤波系数矩阵相乘; 移位器 1106 , 耦合至乘法器 1104 , 设置 为产生滤波后的变换系数。 综上所述, 本发明通过釆用上述滤波技术, 降低了视频中的噪声, 适当减 少了非感兴趣区域的细节信息, 同时减少了压缩伪效应, 从而将更多的比特用 于编码感兴趣区域, 提高了感兴趣区域图像质量的效果。 显然, 本领域的技术人员应该明白, 上述的本发明的各模块或各步骤可以 用通用的计算装置来实现, 它们可以集中在单个的计算装置上, 或者分布在多 个计算装置所组成的网络上, 可选地, 它们可以用计算装置可执行的程序代码 来实现, 从而, 可以将它们存储在存储装置中由计算装置来执行, 并且在某些 情况下, 可以以不同于此处的顺序执行所示出或描述的步骤, 或者将它们分别 制作成各个集成电路模块, 或者将它们中的多个模块或步骤制作成单个集成电 路模块来实现。 这样, 本发明不限制于任何特定的硬件和软件结合。 以上所述仅为本发明的优选实施例而已, 并不用于限制本发明, 对于本领 域的技术人员来说, 本发明可以有各种更改和变化。 凡在本发明的 ^"神和原则 之内, 所作的任何修改、 等同替换、 改进等, 均应包含在本发明的保护范围之 内。

Claims

权 利 要 求 书 一种视频编码的滤波方法, 包括:
在编码每帧视频图像前对所述视频图像进行感兴趣区域检测, 得到 以宏块为单位的感兴趣区域;
计算所述视频图像中每个宏块到所述感兴趣区域的中心宏块的距 离;
按照各个所述宏块到所述中心宏块的距离分别对各个所述宏块进行 空间滤波。 根据权利要求 1所述的方法, 其中, 在按照各个所述宏块到所述中心宏 块的距离分别对各个所述宏块进行空间滤波之后, 所述方法还包括: 确定当前宏块属于非感兴趣区域并且所述当前宏块釆用帧间编码; 才艮据所述当前宏块的像素方差、 残差系数方差和运动强度对所述当 前宏块进行滤波。 根据权利要求 2所述的方法, 其中, 根据所述当前宏块的像素方差、 残 差系数方差和运动强度对所述当前宏块进行滤波包括:
根据所述当前宏块的像素方差、 残差系数方差和运动强度计算综合 滤波因子;
判断所述综合滤波因子是否大于第一门限, 如果是, 则不对所述当 前宏块进行残差滤波, 否则, 使用预定的滤波模板对所述当前宏块进行 残差滤波。 根据权利要求 3所述的方法, 其中, 使用预定的滤波模板对所述当前宏 块进行残差滤波包括:
当所述综合滤波因子大于或等于第一门限且小于第二门限时, 使用
1 2 1
滤波模板 TF3= 2 4 2 对所述当前宏块进行残差滤波;
1 2 1 当所述综合滤波因子大于或等于第二门限时, 使用滤波模板 前宏块进行残差滤波。
Figure imgf000016_0001
根据权利要求 1至 4中任一项所述的方法, 其中, 在按照各个所述宏块 到所述中心宏块的距离分别对各个所述宏块进行空间滤波之后, 所述方 法还包括:
对非感兴趣区域的每个宏块的离散余弦变换系数进行滤波。
6. 根据权利要求 5所述的方法, 其中, 对非感兴趣区域的每个宏块的离散 余弦变换系数进行滤波包括:
16 15 14 12 10 8 6 4
15 14 12 10 8 6 4 2
14 12 10 8 6 4 2 2
12 10 8 6 4 2 2 2
使用公式 2 = ·_ 对非感兴趣区域的
16 10 8 6 4 2 2 2 1
8 6 4 2 2 2 1 1
6 4 2 2 2 1 1 1
4 2 2 2 1 1 1 1
每个宏块的离散余弦变换系数进行滤波, 其中, Α是滤波前的离散余弦 变换系数, 是滤波后的离散余弦变换系数。 根据权利要求 1所述的方法, 其中, 按照距离的大小对所述每个宏块进 行空间滤波包括:
判断当前宏块到所述中心宏块的距离是否大于第三门限, 如果是, 则使用预定的滤波模板对所述当前宏块进行空间滤波, 否则, 不对所述 当前宏块进行空间滤波。 根据权利要求 7所述的方法, 其中, 使用预定的滤波模板对所述当前宏 块进行空间滤波包括: 当所述当前宏块到所述中心宏块的距离大于所述第三门限且小于或
1 5 1
等于第四门限时, 使用滤波模板 TF1= 5 25 5 对所述当前宏块进行:
1 5 1
间滤波;
当所述当前宏块到所述中心宏块的距离大于第四门限且小于或等于
1 3 1
第五门限时,使用滤波模板 T F2— 3 8 3 对所述当前宏块进行空间滤波;
1 3 1 当所述当前宏块到所述中心宏块的距离大于第五门限且小于或等于 第六门限时, 使用滤波模板 T: 前宏块进行空间滤
Figure imgf000017_0001
当所述当前宏块到所述中心宏块的距离大于第六门限时, 使用滤波
1 1 1
模板 TF 1 1 1 对所述当前宏块进行空间滤波。
1 1 1
9. 一种视频编码的滤波装置, 包括:
检测模块, 设置为在编码每帧视频图像前对所述视频图像进行感兴 趣区域检测, 得到以宏块为单位的感兴趣区域;
计算模块, 设置为计算所述视频图像中每个宏块到所述感兴趣区域 的中心宏块的 ^巨离; 空间滤波模块, 设置为按照各个所述宏块到所述中心宏块的距离分 别对各个所述宏块进行空间滤波。
10. 根据权利要求 9所述的装置, 其中, 所述装置还包括:
确定模块, 设置为确定当前宏块属于非感兴趣区域并且所述当前宏 块釆用帧间编码;
滤波模块, 设置为根据所述当前宏块的像素方差、 残差系数方差和 运动强度对所述当前宏块进行滤波。
PCT/CN2011/076286 2010-09-10 2011-06-24 视频编码的滤波方法及装置 WO2012031498A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201010286637.3A CN101945281B (zh) 2010-09-10 2010-09-10 视频编码的滤波方法及装置
CN201010286637.3 2010-09-10

Publications (1)

Publication Number Publication Date
WO2012031498A1 true WO2012031498A1 (zh) 2012-03-15

Family

ID=43437005

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2011/076286 WO2012031498A1 (zh) 2010-09-10 2011-06-24 视频编码的滤波方法及装置

Country Status (2)

Country Link
CN (1) CN101945281B (zh)
WO (1) WO2012031498A1 (zh)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101945281B (zh) * 2010-09-10 2014-09-10 中兴通讯股份有限公司 视频编码的滤波方法及装置
US8964852B2 (en) * 2011-02-23 2015-02-24 Qualcomm Incorporated Multi-metric filtering
CN105516721A (zh) * 2014-10-20 2016-04-20 广东中星电子有限公司 一种视频编码器及其码率控制方法
CN106572351B (zh) * 2015-10-13 2019-11-15 腾讯科技(北京)有限公司 视频编码中的码率控制方法和装置、电子设备以及计算机可读存储介质
CN105930678A (zh) * 2016-05-19 2016-09-07 信阳农林学院 一种生猪口蹄疫实时预警的方法及系统
CN108063946B (zh) * 2017-11-16 2021-09-24 腾讯科技(成都)有限公司 图像编码方法和装置、存储介质及电子装置
CN111656785A (zh) * 2019-06-28 2020-09-11 深圳市大疆创新科技有限公司 可移动平台的图像处理方法、装置、可移动平台及介质
CN114531599B (zh) * 2022-04-25 2022-06-21 中国医学科学院阜外医院深圳医院(深圳市孙逸仙心血管医院) 一种用于医疗图像存储的图像压缩方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101102492A (zh) * 2007-07-26 2008-01-09 上海交通大学 实现基于感兴趣区域的压缩域mpeg-2到h.264视频转换的方法
CN101420618A (zh) * 2008-12-02 2009-04-29 西安交通大学 基于感兴趣区域的自适应可伸缩视频编解码结构设计方法
CN101882316A (zh) * 2010-06-07 2010-11-10 深圳市融创天下科技发展有限公司 一种图像区域划分/编码方法、装置及系统
CN101945281A (zh) * 2010-09-10 2011-01-12 中兴通讯股份有限公司 视频编码的滤波方法及装置

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8761538B2 (en) * 2008-12-10 2014-06-24 Nvidia Corporation Measurement-based and scalable deblock filtering of image data

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101102492A (zh) * 2007-07-26 2008-01-09 上海交通大学 实现基于感兴趣区域的压缩域mpeg-2到h.264视频转换的方法
CN101420618A (zh) * 2008-12-02 2009-04-29 西安交通大学 基于感兴趣区域的自适应可伸缩视频编解码结构设计方法
CN101882316A (zh) * 2010-06-07 2010-11-10 深圳市融创天下科技发展有限公司 一种图像区域划分/编码方法、装置及系统
CN101945281A (zh) * 2010-09-10 2011-01-12 中兴通讯股份有限公司 视频编码的滤波方法及装置

Also Published As

Publication number Publication date
CN101945281B (zh) 2014-09-10
CN101945281A (zh) 2011-01-12

Similar Documents

Publication Publication Date Title
WO2012031498A1 (zh) 视频编码的滤波方法及装置
Zhao et al. NSST: Non-separable secondary transforms for next generation video coding
TWI699996B (zh) 圖像處理裝置及方法
JP6615346B2 (ja) 符号化処理におけるリアルタイムビデオノイズ低減のための方法、端末、および、不揮発性コンピュータ可読記憶媒体
CN103124347B (zh) 利用视觉感知特性指导多视点视频编码量化过程的方法
US20170127062A1 (en) Adaptive Dequantization Method and Apparatus in Video Coding
US20110194615A1 (en) Video sequence encoding system and algorithms
CN109451310B (zh) 一种基于显著性加权的率失真优化方法及装置
CN1695381A (zh) 在数字视频信号的后处理中使用编码信息和局部空间特征的清晰度增强
WO2021185257A1 (zh) 图像编码方法、图像解码方法及相关装置
JP4794147B2 (ja) フレームシーケンスを符号化する方法、フレームシーケンスを復号する方法、その方法を実施する装置、その方法を実行するコンピュータプログラム、およびそのコンピュータプログラムを格納する記憶媒体
CN117640942A (zh) 一种用于视频图像的编码方法和装置
WO2006133613A1 (en) Method for reducing image block effects
CN108810552A (zh) 图像预测方法和相关产品
JP3046379B2 (ja) 画像符号化装置
WO2022252222A1 (zh) 编码方法和编码装置
CN113489976A (zh) 预测编码方法、装置和电子设备
KR20090017724A (ko) 동영상 부호화에 있어서 비트 발생 가능성 예측을 이용한블록 모드 결정 방법 및 장치
CN107409211B (zh) 一种视频编解码方法及装置
JP3150452B2 (ja) 動画像符号化装置
CN103313064B (zh) 基于帧间模式和运动修补的时域错误隐藏方法
WO2020147226A1 (zh) 信息处理方法及装置、设备、存储介质
Chon et al. Enabling access through real-time sign language communication over cell phones
JP4222274B2 (ja) 符号化モード選択装置及び符号化モード選択プログラム
CN105681784B (zh) 一种基于h264/avc视频的psnr盲估计方法

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: 11823021

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11823021

Country of ref document: EP

Kind code of ref document: A1