WO2009046601A1 - Procédé de compensation de mouvement - Google Patents

Procédé de compensation de mouvement Download PDF

Info

Publication number
WO2009046601A1
WO2009046601A1 PCT/CN2007/070858 CN2007070858W WO2009046601A1 WO 2009046601 A1 WO2009046601 A1 WO 2009046601A1 CN 2007070858 W CN2007070858 W CN 2007070858W WO 2009046601 A1 WO2009046601 A1 WO 2009046601A1
Authority
WO
WIPO (PCT)
Prior art keywords
filter
coding
subpixel
entropy
residual
Prior art date
Application number
PCT/CN2007/070858
Other languages
English (en)
Inventor
Hoi Ming Wong
Yan Huo
Original Assignee
Hong Kong Applied Science And Technology Research Institute Co. Ltd.
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 Hong Kong Applied Science And Technology Research Institute Co. Ltd. filed Critical Hong Kong Applied Science And Technology Research Institute Co. Ltd.
Priority to PCT/CN2007/070858 priority Critical patent/WO2009046601A1/fr
Publication of WO2009046601A1 publication Critical patent/WO2009046601A1/fr

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
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51—Motion estimation or motion compensation
    • H04N19/523—Motion estimation or motion compensation with sub-pixel accuracy
    • 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
    • H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/103—Selection of coding mode or of prediction mode
    • H04N19/105—Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction
    • 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
    • H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/117—Filters, e.g. for pre-processing or post-processing
    • 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
    • H04N19/136—Incoming video signal characteristics or properties
    • H04N19/137—Motion inside a coding unit, e.g. average field, frame or block difference
    • H04N19/139—Analysis of motion vectors, e.g. their magnitude, direction, variance or reliability
    • 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
    • H04N19/146—Data rate or code amount at the encoder output
    • H04N19/147—Data rate or code amount at the encoder output according to rate distortion criteria
    • 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
    • H04N19/169—Methods 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
    • H04N19/17—Methods 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
    • H04N19/172—Methods 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 picture, frame or field
    • 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
    • H04N19/169—Methods 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
    • H04N19/17—Methods 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
    • H04N19/174—Methods 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 slice, e.g. a line of blocks or a group of blocks
    • 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
    • H04N19/169—Methods 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
    • H04N19/17—Methods 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
    • H04N19/176—Methods 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
    • 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
    • H04N19/46—Embedding additional information in the video signal during the compression process
    • 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
    • H04N19/80—Details of filtering operations specially adapted for video compression, e.g. for pixel interpolation
    • H04N19/82—Details of filtering operations specially adapted for video compression, e.g. for pixel interpolation involving filtering within a prediction loop

Definitions

  • the claimed invention relates to a method for video compression.
  • the claimed invention relates to a method for removing temporal redundancy between video frames by motion compensation.
  • Video compression addresses the problem of reducing the amount of data required to represent video and a number of standards have been available for implementation.
  • the video coding standards in use today can be grouped into two broad categories according to their intended applications: (i) video teleconferencing standards such as H.261 and H.264 which have been defined by the International Telecommunications Union (ITU) and (ii) multimedia standards such as MPEG-2, MPEG-4 which are developed under the auspices of the Motion Picture Experts Group of the CCITT and ISO.
  • Video compression techniques include the still image compression techniques due to the fact that video is made up of a series of still images known as frames (or fields).
  • predictive coding techniques such as DPCM (Differential Pulse Code Modulation) which is a spatial domain method
  • DCT Discrete Cosine Transform
  • transform coding making use of a reversible, linear transform to map the image into a set of transform coefficients. The coefficients are then quantized and coded.
  • these video compression techniques also make use of motion estimation and motion compensation in order to reduce temporal redundancies.
  • Moving objects in video often appear from frame (or field, both terms can be used interchangeably for simplicity hereinafter) to frame with all or part of it relocated in those subsequent frames. Despite those relocations, correlation among the sequence of frames is high and gives rise to redundancy. This so-called temporal redundancy can be reduced by comparing and relating the pixel in the present frame to the location of the same object in the preceding reference frame.
  • an encoder is used to estimate the motion of object in the image in order to find the corresponding area in a previous frame. If moving objects in a video have non-integer pixel displacement between frames, motion vectors telling the decoder what portion of the previous frame for predicting the new frame will have fractional values instead of integer values. In order to cope with this sort of fractional-pixel motion, it is required to interpolate fractional pixels (hereinafter referred as subpixels) before combining with the residue to reconstitute the current frame.
  • subpixels fractional pixels
  • the interpolation is carried out either by solely a fixed filter or by solely an adaptive filter.
  • Fixed filters are symmetric and cannot capture any small motion offset between video frames.
  • the bandwidth of a fixed filter is fixed, inhibiting it from being adaptive to various video signals such as textured, noisy or smooth ones.
  • the complexity will be very high as a lot of costly multiplications are required, for example, each subpixel requires 36 multiplications if 36-tap filter is to interpolate each subpixel from 36 integer pixels.
  • the claimed invention is related to interpolating fractional pixels by using a combination of fixed filter and adaptive filter to generate fractional-pixel images.
  • the filters in use are not restricted to be symmetrical and the sum of the filter coefficient values of the interpolation filter are not fixed which is good for fading video. Fewer multiplications are required in the claimed invention, lowering the complexity of the interpolation with comparable and even better performance in term of video signal quality.
  • the reference pixels used by the disclosed adaptive filter are chosen from a set of subpixels, which are interpolated from a set of fixed filters.
  • the claimed invention provides a coding method adopted by the video encoder for the adaptive filter coefficient in which the adaptive filter coefficients are predicted from standard, upper or left filter such that to reduce the data required to be sent to the video decoder.
  • the disclosed coding of filter coefficients are no longer limited by any predefined filter.
  • a set of disclosed adaptive filter pattern and coefficients can be sent per frame, per field, or per number of macroblocks.
  • FIG. 1 shows a schematic diagram of a 2-step interpolation for an embodiment of the claimed invention
  • FIG. 2 shows a process of using intermediate subpixel value (interpolated by fixed filter) to interpolate final subpixel value for an embodiment of the claimed invention
  • FIG. 3 shows the reference image portion, intermediate subpixel image portion and reference subpixel image portion in a video frame
  • FIG. 4 shows the intermediate subpixel image portion with pixels required in the construction of interpolation filters for an embodiment of the claimed invention
  • FIG. 5 shows the interpolation of a video frame by a scalable filter for an embodiment of the claimed invention
  • FIG. 6 shows a flow chart for filter coefficient coding
  • FIG. 7 shows prediction modes available for coding filter coefficients at each subpixel location.
  • FIG. 8 shows a general video compression system model where the disclosed invention can be applied.
  • One embodiment of the claimed invention is to carry out the interpolation of a video frame for motion compensation in two steps as shown in FIG. 1.
  • each image in a video is made up of 2 fields. But for progressive scanning, each image is referred to be a frame. Therefore, these 3 terms: image, field and frame represent a still image in a video - a screenshot at a particular instance and can be used interchangeably throughout this disclosure.
  • image, field and frame represent a still image in a video - a screenshot at a particular instance and can be used interchangeably throughout this disclosure.
  • FIG. 1 after passing through filter 101, the resolution of frame 115 will be increased to that of frame 126 up to subpixel level as an intermediate subpixel image. Then frame 126 is further processed by filter 108 to obtain the desired output frame 133 as a reference subpixel image but this time the resolution remains the same as frame 126.
  • FIG. 2 further illustrates that in the 2-step filtering process, the first filter in use is a fixed filter 240 and the second filter is an adaptive filter 250.
  • the fixed filter 240 is applied on a reference image 210 which is an image previously received by the decoder (not shown).
  • an intermediate subpixel image 220 is generated which is then fed into the adaptive filter 250 to produce the reference subpixel image 230 from which the current image is reconstituted by arranging the pixels of the reference subpixel image 230 according to the estimated motion and adding the residue image (not shown) which is the prediction error for the current image.
  • the image is divided into a number of blocks which can be of any size such as 2x2, 4x4, 8x8, 16x16 pixels for processing.
  • image portion a portion of the image (6x6 pixels in size for an embodiment of the claimed invention) is used for exemplary illustration (hereafter referred as image portion) and the image portions are shown in FIG. 3. Therefore, when reconstituting the current image from the reference image, the reference image is manipulated in a portion by portion manner. Each image portion 310 from the reference image will then be fed into the fixed filter to build the intermediate subpixel image which is made up of image portions 320. The interpolation is continued by applying the adaptive filter on portions 320 to output the reference subpixel image which is made up of image portions 330.
  • Interpolation is to pad subpixels in between each existing pixel.
  • the value of each subpixel can be determined by those of existing pixels. The more subpixels are padded in between those existing pixels, the higher the precision of the fractional-pixel movement can be described. For example, if a moving object moves only half a pixel in a video, it is required to use at least half-pixel interpolation in which each existing pixel will become one pixel apart from its adjacent horizontal neighbors and from its adjacent vertical neighbors. Therefore, the half-pixel movement can be more accurately modeled by moving a "half step" to the interpolated subpixel in-between the existing pixels instead of a "full step” to the next existing pixel.
  • the precision of the fractional-pixel movement which can be catered for by the claimed invention can be up to any degree such as half pixel, quarter pixel, one-tenth pixel, etc. depending on how many subpixels are interpolated.
  • Reference image portion 310 contains 6x6 pixels known as reference pixels 302. Each of these reference pixels are named from A to F according to which rows they are in and numbered sequentially from 1 to 6 according to which columns they are in.
  • the intermediate subpixel image portion 320 is formed. Intermediate subpixels such as intermediate subpixel 308 are padded around each reference pixel 302. Now that all reference pixels such as reference pixel 302 are 3-pixel apart from each other in term of vertical and horizontally distance.
  • each reference pixel 302 has been interpolated into a size of 4x4 pixels in which the reference pixel 302 is located at the top left-hand corner in this 4x4 pixel area and the rest of the area will be filled by 15 intermediate subpixels 308. Then based on these intermediate subpixels 308 and the reference pixels 302, the adaptive filter is used to generate reference subpixels 316 as in the reference subpixel image portion 330 where there is no further padding of subpixels and the size of the reference subpixel image portion 330 remains the same as that of the intermediate subpixel image portion 320.
  • the interpolation is carried out in two steps but previously it is done in one operation.
  • the following will illustrate how the two-step interpolation is performed to achieve better prediction by combining fixed filter and adaptive filter with less taps and thus reduce the complexity of computation and implementation of the interpolation significantly.
  • the more taps in the filter applied such as adaptive filter the higher accuracy will be for the prediction.
  • adaptive filter coefficients must be computed by multiplications. For example, one may use an adaptive filter with 36 taps to interpolate each subpixel from a 6x6 image portion, i.e. 36 reference pixels:
  • SP ⁇ h ⁇ ) - AU h(l) - A2- ... H ⁇ (34) - F5 H h(35) - F ⁇ (1)
  • SP is the value for a reference subpixel 316
  • h(i) are the adaptive filter coefficients and Al
  • A2 ... Bl... F6 are the value of those reference pixels in the reference image portion 310.
  • This 36-tap filter needs 36 multiplication and the cost is very high as multiplication itself is costly.
  • SoC System-on-Chip
  • the area for a multiplier is approximately 10 times of that of the area for a shifter and the speed of a multiplier is much lower, i.e. comparable to a tenth of the speed of a shifter. Therefore, the claimed invention uses intermediate subpixel value (interpolated by fixed filter) to obtain final subpixel value (interpolated by adaptive filter with a few number of taps), providing much lower complexity. As such, a smaller size of adaptive filter is required for the same interpolation performance.
  • one embodiment for the fixed filter 240 can be the H.264 filter and which is used to compute the intermediate subpixel values spi'.
  • the H.264 filter takes the example of the values for the intermediate subpixels a', b', c', d', e', f , g', h', i', j', k', 1', m', n', o' in intermediate subpixel image portion 320 in FIG. 3, the H.264 half-pixel interpolation filter will be used to compute the intermediate subpixel values spi' for intermediate subpixels b', h' and j' with 6 fixed taps: [1, -5, 20, 20, -5, l]/32.
  • the intermediate subpixel values spi' for intermediate subpixels a', c', d', e', f , g', h', i', k', 1', m', n', o' will be computed by the H.264 quarter-pixel interpolation filter which is a bilinear one with filter coefficients of: [1, l]/2.
  • the computation is done by the following equations with respect to the subpixel values at subpixel locations as shown in image portion 400 in
  • filter coefficients for adaptive filter can be determined by a variety of methods during the coding stage such as the one used in VCEG Z- 17.
  • diamond pattern is adopted and the values for each reference subpixel 316 can be computed by the following equation where 5 taps are required:
  • SP ⁇ h(0) ⁇ spO'- h( ⁇ ) ⁇ spl ⁇ h(2) ⁇ spT ⁇ h(3) ⁇ sp3'- ⁇ h(4) ⁇ sp4' ⁇ (17)
  • SP is the reference subpixel value
  • h(i) are the filter coefficient
  • spi' are the intermediate subpixel values obtained from interpolating by fixed filter
  • Equation 17 can be written as:
  • the optimal filter coefficients for the adaptive filter 250 can be obtained by solving the following Sum of Squared Error (SSE) equation for the whole frame:
  • i, j are the filter indices and Z ⁇ j ⁇ ⁇ T ⁇ + ⁇ y+j
  • the reference subpixel values SP are determined for each reference subpixel in the order of a, b, c, d, e, f, g, h, i, j, k, 1, m, n in the reference image portion 330 as shown in FIG. 3.
  • a scalable adaptive filter is used in the disclosed embodiment.
  • FIG. 5 an intermediate subpixel image portion 500 is shown and the intermediate subpixels in different grayscale represent different filter scales.
  • the scale is selected from 1 to 6
  • the number of taps for the adaptive filter is 5
  • the filter pattern is diamond.
  • Each of these far-away intermediate subpixels 520 are either vertically or horizontally 3 pixels apart from the intermediate subpixel j'. Both the value of the 5 taps of coefficients and the value of the scale will be fed into the decoder when determining the value for the reference subpixel j.
  • coefficient taps For the adaptive filter used in the embodiment of the claimed invention, there are 5 coefficient taps as shown in equation 17. Since the interpolation enlarge each reference pixel into a size of 4x4 pixels where values for 15 subpixels are required to be determined through equation 17. The total number of coefficients will then be 75 and if they are coded in 16-bit integers, it will take 1200 bits per slice to store the coefficients. These coefficients may further be predicted and entropy coded in order to reduce the cost.
  • Filter coefficients can be predicted from the corresponding coefficients of neighboring subpixels. However, in the claimed invention, the correlation between filter coefficients is much smaller and the following coding method needs to be used instead of any symmetrical assumption.
  • FIG. 6 shows the steps for the prediction of filter coefficients. The filter coefficient for a portion is predicted and coded before being sent to the decoder. First of all, it is required to obtain the residual filter (Filter_residual) for each subpixel location as in residual filter determination step 610. Assuming the adaptive filter 250 in use has N taps, N can be any integer such as 5, the adaptive filter for each subpixel in a portion will also have N taps.
  • the rule for computing the residual filters is:
  • Each of the adaptive filter coefficients for all subpixel locations can be predicted from upper subpixel location (if available), left subpixel location (if available), or standard filter coefficients (std_filter), indicated by 1 to 2 bits.
  • the standard filter is a pre-defined (constant) filter for the whole video sequence and an example for the standard filter coefficients can be ⁇ 1,0,0,0,0 ⁇ and all operations between filters are tap-tap operations.
  • the N-tap residual filter at a particular subpixel location is the argument which can minimize the value of the Sum of Squared Error (SSE) between the filter at subpixel location and its corresponding resconstructed filters.
  • the residual filter can be determined by a subtraction of the standard filter from the filter at subpixel location.
  • the arrows such as arrow 730 represent how the filter at each subpixel location such as filter 720 depends on the reconstructed filter (not shown in FIG. 7).
  • subpixel filter at f (filter_f) is related to the reconstructed filters at subpixels b' (filter_br') and e' (filter_er').
  • the equations are given as follows:
  • Filter_residual(i') is the residual filter at subpixel location i
  • filter_i' is the filter at subpixel location i
  • filter_ir' is the reconstructed filter at subpixel location i
  • std_filter is the standard filter.
  • Filter_residual(a') filter_a' - std_filter (21)
  • Filter_residual(b') arg min ( SSE(filter_b'- filter_ar') , SSE(filter_b' - std_filter) )
  • Filter_residual(c') arg min ( SSE(filter_c'- filter_br') , SSE(filter_b' - std_filter) )
  • Filter_residual(d') filter_d' - std_filter (24)
  • Filter_residual(e') arg min ( SSE(filter_e'- filter_ar') , SSE(filter_e' - filter_dr'), SSE(filter_e' - std_filter) ) (25)
  • Filter_residual(f ) arg min ( SSE(filter_f - filter_er') , SSE(filter_f - filter_br'), SSE(filter_f - std_filter) ) (26)
  • Filter_residual(g') arg min ( SSE(filter_g'- filter_cr') , SSE(filter_g' - filter_fr'), SSE(filter_g' - std_filter) ) (27)
  • Filter_residual(h') arg min ( SSE(filter_ filter_h'- filter_dr') , SSE(filter_h' - filter_std_filter) ) (28)
  • Filter_residual(i') arg min ( SSE(filter_i'- filter, er') , SSE(filter_i' - filterjir'), SSE(filter_i' - std_filter) ) (29)
  • Filter_residual(j') arg min ( SSE(filterJ'- filter_fr') , SSE(filterJ' - filter_ir'), SSE(filterJ' - std_filter) ) (30)
  • Filter_residual(k') arg min ( SSE(filter_k'- filter_gr') , SSE(filter_k' - filter Jr'), SSE(filter_k' - std_filter) ) (31)
  • Filter_residual(n' ) arg min ( SSE(filter_n'- filterjr' ) , SSE(n' - filter_mr' ), SSE(filter_n' - std_filter) ) (33)
  • Filter_residual(o') arg min ( SSE(filter_o'- filter_kr') , SSE(filter_o' - filter_nr'), SSE(filter_o' - std_filter) ) (34)
  • All the filter coefficients are subsequently truncated to values with fixed number of bits, for example 10 bits, in the filter coefficient truncation step 620 as shown in FIG. 6.
  • All residual filters for all subpixels are entropy-coded by existing coding schemes such as Exp-Golumb code in the residual-filter coding step 630.
  • the entropy coding in use can be variable length coding or arithmetic coding such as CABAC (Context-based Adaptive Binary Arithmetic Coding).
  • the prediction mode for each subpixel location is also entropy-coded in the prediction-mode coding step 640. For example, if SSE(filter_e' - filter_ar') is the minimum, then the prediction mode will be "upper". This will be indicated by 1 - 2 bits.
  • the filter scale is also coded into binary code in the filter scale coding step 650. Finally, all coded residual filters and the coded prediction modes are sent to the decoder in the decoder communication step
  • 600 - 1200 bits are spent on average on filter coefficients per slice after using the aforesaid coding method.
  • FIG. 8 shows a general video compression model where the disclosed invention can be applied.
  • the input video 810 is encoded before sending to the receiver over the channel 840.
  • the encoding is carried out by the source encoder 820 and channel encoder 830.
  • the source encoder 820 and channel encoder 830 are together regarded as encoder 801.
  • the compressed video will be decoded by the channel decoder 805 and the source decoder 806 to generate the output video 807.
  • the channel decoder 805 and the source decoder 806 together can be regarded as decoder 802.
  • the interpolation method as described by the claimed invention is applied in encoding at the encoder 801 as well as decoding at the decoder 802.
  • the filter coefficients used in the interpolation can be coded at encoder 801 before sending to the decoder 802 in a way described in one embodiment of the claimed invention.
  • the disclosed method and related device have industrial applicability in the video industry.
  • the disclosed method and related device can find applications in the following: Cable TV (CATV) on optical networks, copper, etc.; Direct broadcast satellite (DBS) video services; Digital subscriber line (DSL) video services; Digital terrestrial television broadcasting (DTTB); Interactive storage media (ISM) such as optical disks, etc.; Multimedia mailing (MMM); Multimedia services over packet networks (MSPN); Real-time conversational (RTC) services such as videoconferencing, videophone, etc.; Remote video surveillance (RVS); Serial storage media (SSM) such as digital video tape recorder (DVTR), and others.

Landscapes

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

Abstract

La présente invention concerne un procédé destiné à être utilisé dans une compression vidéo. En particulier, l'invention concerne un procédé d'interpolation de pixels fractionnaires plus efficace en deux étapes par un filtre fixe (240) et un filtre adaptatif (250) pour une compensation de mouvement de pixels fractionnaires.
PCT/CN2007/070858 2007-10-09 2007-10-09 Procédé de compensation de mouvement WO2009046601A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2007/070858 WO2009046601A1 (fr) 2007-10-09 2007-10-09 Procédé de compensation de mouvement

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2007/070858 WO2009046601A1 (fr) 2007-10-09 2007-10-09 Procédé de compensation de mouvement

Publications (1)

Publication Number Publication Date
WO2009046601A1 true WO2009046601A1 (fr) 2009-04-16

Family

ID=40548941

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2007/070858 WO2009046601A1 (fr) 2007-10-09 2007-10-09 Procédé de compensation de mouvement

Country Status (1)

Country Link
WO (1) WO2009046601A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2709364A1 (fr) * 2011-06-27 2014-03-19 Nippon Telegraph and Telephone Corporation Procédé et dispositif de codage d'images vidéo, procédé et dispositif de décodage d'images vidéo, et programme associé

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1620821A (zh) * 2002-01-23 2005-05-25 索尼株式会社 图像信息编码装置和图像信息编码方法、图像信息解码装置和图像信息解码方法
CN1625902A (zh) * 2002-04-24 2005-06-08 日本电气株式会社 运动图片编码和解码方法以及使用该方法的设备和程序
US20060294171A1 (en) * 2005-06-24 2006-12-28 Frank Bossen Method and apparatus for video encoding and decoding using adaptive interpolation

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1620821A (zh) * 2002-01-23 2005-05-25 索尼株式会社 图像信息编码装置和图像信息编码方法、图像信息解码装置和图像信息解码方法
CN1625902A (zh) * 2002-04-24 2005-06-08 日本电气株式会社 运动图片编码和解码方法以及使用该方法的设备和程序
US20060294171A1 (en) * 2005-06-24 2006-12-28 Frank Bossen Method and apparatus for video encoding and decoding using adaptive interpolation

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2709364A1 (fr) * 2011-06-27 2014-03-19 Nippon Telegraph and Telephone Corporation Procédé et dispositif de codage d'images vidéo, procédé et dispositif de décodage d'images vidéo, et programme associé
EP2709364A4 (fr) * 2011-06-27 2014-11-26 Nippon Telegraph & Telephone Procédé et dispositif de codage d'images vidéo, procédé et dispositif de décodage d'images vidéo, et programme associé
US9667963B2 (en) 2011-06-27 2017-05-30 Nippon Telegraph And Telephone Corporation Method and apparatus for encoding video, method and apparatus for decoding video, and programs therefor

Similar Documents

Publication Publication Date Title
US7653132B2 (en) Method and system for fast implementation of subpixel interpolation
Ostermann et al. Video coding with H. 264/AVC: tools, performance, and complexity
EP1466477B1 (fr) Codage de filtres dynamiques
KR100919557B1 (ko) 향상된 코딩 모드 선택 방법 및 장치
CN102017615B (zh) 视频单元内的边界伪影校正
US20120033040A1 (en) Filter Selection for Video Pre-Processing in Video Applications
US20120076203A1 (en) Video encoding device, video decoding device, video encoding method, and video decoding method
US6628714B1 (en) Down converting MPEG encoded high definition sequences to lower resolution with reduced memory in decoder loop
EP1617672A1 (fr) Estimateur/compensateur de mouvement comportant un 16-bit 1/8 pel filtre d'interpolation
EP1690421B1 (fr) Procede de correction de valeurs de pixels interpoles
MXPA05000335A (es) Metodo y sistema para seleccionar tipo de filtro de interpolacion en codificacion de video.
US20140247890A1 (en) Encoding device, encoding method, decoding device, and decoding method
WO2004105399A1 (fr) Procede et dispositif de compression video
CN113728629A (zh) 视频译码中的运动向量推导
US20110228854A1 (en) Apparatus and method for encoding/decoding a video signal
JP4761390B2 (ja) 内挿される画素値の計算方法の改良
US8090031B2 (en) Method for motion compensation
EP1221261A1 (fr) Conversion de domaine tcd d'un signal video en un signal a definition reduite
US20050105611A1 (en) Video compression method
Qian et al. Transform domain transcoding from MPEG-2 to H. 264 with interpolation drift-error compensation
US20090180541A1 (en) Video motion compensation
US6510178B1 (en) Compensating for drift in the down conversion of high definition sequences to lower resolution sequences
US20060291743A1 (en) Configurable motion compensation unit
WO2009046601A1 (fr) Procédé de compensation de mouvement
US6061401A (en) Method and apparatus for selectively encoding/decoding a video signal

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: 07817049

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: 07817049

Country of ref document: EP

Kind code of ref document: A1