CN101888558B - A kind of chrominance interpolation method and device - Google Patents
A kind of chrominance interpolation method and device Download PDFInfo
- Publication number
- CN101888558B CN101888558B CN201010211400.9A CN201010211400A CN101888558B CN 101888558 B CN101888558 B CN 101888558B CN 201010211400 A CN201010211400 A CN 201010211400A CN 101888558 B CN101888558 B CN 101888558B
- Authority
- CN
- China
- Prior art keywords
- pixel
- interpolation
- integer pel
- apart
- nearest
- 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.)
- Active
Links
Landscapes
- Television Systems (AREA)
- Processing Of Color Television Signals (AREA)
Abstract
The invention provides a kind of chrominance interpolation method and device, chrominance interpolation method wherein specifically comprises: the Integer Pel point in the interpolation region of input reference frame; Utilize four the Integer Pel points nearest apart from 1/2 pixel, interpolation obtains described 1/2 pixel; Utilize apart from the nearest Integer Pel point of 1/4 pixel and 1/2 pixel, interpolation obtains described 1/4 pixel; Utilize Integer Pel point, 1/2 pixel and 1/4 pixel nearest apart from 1/8 pixel, interpolation obtains described 1/8 pixel.The present invention is mainly in order to improve the precision of chroma interpolation.
Description
Technical field
The present invention relates to video coding and decoding technology field, particularly relate to a kind of chrominance interpolation method and device.
Background technology
H.264 standard is a kind of video high compression techniques, and full name is MPEG (Motion Picture Experts Group, MovingPictureExpertGroup)-4AVC (advanced video coding, AdvancedVideoCoding) standard.
In the h .264 standard, a two field picture is divided into several macro blocks, and macro block is made up of 16 × 16 luminance block, Cb and the Cr chrominance block of 8 × 8; For improving the precision of prediction of estimation, H.264 standard adopts the estimation of 1/4 pixel precision, and also, prediction data to be added with residual error data by reference picture and to obtain, and motion vector is 1/4 precision to luminance block, is 1/8 precision to chrominance block.
Therefore, for reference chrominance block, H.264 the Interpolation Process of standard is exactly a process obtaining fraction pixel point (1/2 pixel, 1/4 pixel and 1/8 pixel) by calculating Integer Pel point.
With reference to Fig. 1, show the chroma interpolation schematic diagram in existing H.264 standard, it is input as:
1, two chroma reference image blocks (Cb, a Cr);
2, the chroma motion vector of two use 1/8 accuracy representings, i.e. mvCLX, wherein X represents 0 or 1, represent forward direction or backward, in addition, xFracc=mvCLX [0] & 7 represents the fraction pixel skew of horizontal direction, and yFracc=mvCLX [1] & 7 represents the mark item number skew of vertical direction.
And adopt following formulae discovery to divide pixel X:
X＝((8-xFracC)*(8-yFracC)*A+xFracC*(8-yFracC)*B+(8-xFracC)*yFracC*C+xFracC*yFracC*D+32)＞＞6
Can finding out, owing to only using A, B, C, D tetra-Integer Pel points, like this, the correlation information of some adjacent Integer Pel points can be lost, such as, when calculating 1/2 pixel on AB direction, being only the use of A, B two Integer Pel points; Therefore, interpolation precision is not high.
In a word, the technical problem needing those skilled in the art urgently to solve is exactly: how can improve chroma interpolation precision.
Summary of the invention
Technical problem to be solved by this invention is to provide a kind of chrominance interpolation method and device, in order to improve chroma interpolation precision.
In order to solve the problem, the invention discloses a kind of chrominance interpolation method, comprising:
The Integer Pel point in the interpolation region of input reference frame;
Utilize four the Integer Pel points nearest apart from 1/2 pixel, interpolation obtains described 1/2 pixel;
Utilize apart from the nearest Integer Pel point of 1/4 pixel and 1/2 pixel, interpolation obtains described 1/4 pixel;
Utilize Integer Pel point, 1/2 pixel and 1/4 pixel nearest apart from 1/8 pixel, interpolation obtains described 1/8 pixel.
Preferably, described interpolation obtains the step of described 1/2 pixel, comprising:
Respectively for 1/2 pixel in Integer Pel point horizontal direction and vertical direction, utilize and respective direction carries out interpolation apart from the adjacent Integer Pel point of its nearest four;
For 1/2 pixel in Integer Pel point diagonal, four the Integer Pel points nearest apart from it are utilized to carry out interpolation.
Preferably, described interpolation obtains the step of described 1/2 pixel, comprising:
Adopt 4 tap filters, interpolation is carried out to four the Integer Pel points nearest apart from 1/2 pixel, obtains described 1/2 pixel.
Preferably, described interpolation obtains the step of described 1/4 pixel, comprising:
For 1/4 pixel in 1/2 pixel diagonal, utilize and carry out interpolation apart from its nearest Integer Pel point and three 1/2 pixels;
For 1/4 pixel in non-1/2 pixel diagonal, utilize apart from its nearest Integer Pel point and 1/2 pixel, or two 1/2 pixels, carry out interpolation.
Preferably, the step that described interpolation obtains described 1/8 pixel comprises:
For 1/8 pixel in 1/4 pixel diagonal, utilize apart from its nearest Integer Pel point and three 1/4 pixels, or 1/2 pixel and three 1/4 pixels, carry out interpolation;
For 1/8 pixel in non-1/4 pixel diagonal, utilize apart from its nearest Integer Pel point and 1/4 pixel, or, 1/2 pixel and 1/4 pixel, or two 1/4 pixels, carry out interpolation.
The invention also discloses a kind of chroma interpolation device, comprising:
Input module, for inputting the Integer Pel point in the interpolation region of reference frame;
1/2 pixel interpolating module, for utilizing four the Integer Pel points nearest apart from 1/2 pixel, interpolation obtains described 1/2 pixel;
1/4 pixel interpolating module, for utilizing apart from the nearest Integer Pel point of 1/4 pixel and 1/2 pixel, interpolation obtains described 1/4 pixel;
1/8 pixel interpolating module, for utilizing Integer Pel point, 1/2 pixel and 1/4 pixel nearest apart from 1/8 pixel, interpolation obtains described 1/8 pixel.
Preferably, described 1/2 pixel interpolating module comprises:
One 1/2 pixel interpolation submodule, for respectively for 1/2 pixel in Integer Pel point horizontal direction and vertical direction, utilizes and respective direction carries out interpolation apart from the adjacent Integer Pel point of its nearest four;
2 1/2 pixel interpolation submodule, for for 1/2 pixel in Integer Pel point diagonal, utilizes four the Integer Pel points nearest apart from it to carry out interpolation.
Preferably, described 1/2 pixel interpolating module specifically for, adopt 4 tap filters, carry out interpolation to apart from nearest four the Integer Pel points of 1/2 pixel, obtain described 1/2 pixel.
Preferably, described 1/4 pixel interpolating module comprises:
One 1/4 pixel interpolation submodule, for for 1/4 pixel in 1/2 pixel diagonal, utilizes and carries out interpolation apart from its nearest Integer Pel point and three 1/2 pixels;
2 1/4 pixel interpolation submodule, for for 1/4 pixel in non-1/2 pixel diagonal, utilize apart from its nearest Integer Pel point and 1/2 pixel, or two 1/2 pixels, carries out interpolation.
Preferably, described 1/8 pixel interpolating module comprises:
One 1/8 pixel interpolation submodule, for for 1/8 pixel in 1/4 pixel diagonal, utilize apart from its nearest Integer Pel point and three 1/4 pixels, or 1/2 pixel and three 1/4 pixels, carries out interpolation;
2 1/8 pixel interpolation submodule, for for 1/8 pixel in non-1/4 pixel diagonal, utilizes apart from its nearest Integer Pel point and 1/4 pixel, or, 1/2 pixel and 1/4 pixel, or, two 1/4 pixels, carry out interpolation.
Compared with prior art, the present invention has the following advantages:
The present invention utilizes pixel nearest apart from interpolation fraction pixel point in current reference pixel source to carry out interpolation, because described current reference pixel source is constantly updated along with the change of interpolation fraction pixel point, like this, can ensure that the described pixel for interpolation is nearest apart from interpolation fraction pixel point, and the correlation of pixel and this interpolation fraction pixel point is maximum recently, therefore, it is possible to reach good interpolation, and then improve chroma interpolation precision.
First, the present invention utilizes four the Integer Pel points nearest apart from it, interpolation obtains described 1/2 pixel, relative to prior art, only adopt two adjacent Integer Pel points, or, although adopt four adjacent Integer Pel points, but these Integer Pel points are not nearest apart from described 1/2 pixel, the present invention can obtain accuracy and higher 1/2 pixel of precision, also just provides more high-precision reference pixel source for interpolation 1/4 pixel and 1/8 pixel.
Secondly, the present invention is from the current reference pixel source comprising Integer Pel point and 1/2 pixel, what selection distance 1/4 pixel was nearest carries out interpolation, therefore, relative to prior art, only select adjacent Integer Pel point, and these Integer Pel points not nearest apart from this 1/4 pixel, the present invention can obtain accuracy and higher 1/4 pixel of precision, also just provides more high-precision reference pixel source for interpolation 1/8 pixel.
Moreover, due to from the current reference pixel source comprising Integer Pel point, 1/2 pixel and 1/4 pixel, what selection distance 1/8 pixel was nearest carries out interpolation, therefore, adjacent Integer Pel point is only selected relative to prior art, and these Integer Pel points are not nearest apart from this 1/8 pixel, the present invention can obtain accuracy and higher 1/8 pixel of precision, thus can improve chroma interpolation precision.
Accompanying drawing explanation
Fig. 1 is the chroma interpolation schematic diagram in existing H.264 standard;
Fig. 2 is the flow chart of a kind of chrominance interpolation method embodiment of the present invention;
Fig. 3 is the schematic diagram of a kind of chroma interpolation of the present invention;
Fig. 4 is a kind of application example at coding side of the present invention;
Fig. 5 is a kind of application example in decoding end of the present invention;
Fig. 6 is the structure chart of a kind of chroma interpolation device of the present invention embodiment.
Embodiment
For enabling above-mentioned purpose of the present invention, feature and advantage become apparent more, and below in conjunction with the drawings and specific embodiments, the present invention is further detailed explanation.
One of core idea of the present invention is, pixel nearest apart from interpolation fraction pixel point in current reference pixel source is utilized to carry out interpolation, because described current reference pixel source is constantly updated along with the change of interpolation fraction pixel point, such as, the reference pixel source of interpolation 1/2 pixel comprises Integer Pel point, the reference pixel source of interpolation 1/4 pixel comprises Integer Pel point and 1/2 pixel, the reference pixel source of interpolation 1/8 pixel comprises Integer Pel point, 1/2 pixel and 1/4 pixel, like this, can ensure that the described pixel for interpolation is nearest apart from interpolation fraction pixel point, and the correlation of pixel and this interpolation fraction pixel point is maximum recently, therefore, good interpolation can be reached, and then improve chroma interpolation precision.
With reference to Fig. 2, show the flow chart of a kind of chrominance interpolation method embodiment of the present invention, specifically can comprise:
The Integer Pel point in the interpolation region of step 201, input reference frame;
The Integer Pel point of described input by as initial reference pixel source of the present invention, also, as the reference pixel source of interpolation 1/2 pixel; Here, described interpolation region can be the overall situation or the regional area of reference frame, and the present invention is not limited this.
Step 202, four the Integer Pel points utilizing distance 1/2 pixel nearest, interpolation obtains described 1/2 pixel;
For the reference frame of YCbCr4:2:0 sample format, no matter Cb or Cr chromatic component, all do 1/2 extraction to luminance component Y in horizontal and vertical direction to obtain, namely, Cb with Cr size is the same, horizontal and vertical direction is all the half (sum is exactly 1/4) of luminance component Y, and so adjacent 4 luminance pixels (2x2 block) share Cb and Cr.
Due to chromatic component be luminance component with 2: 1 sampling rate store, if select the adjacent Integer Pel point of large quantity to carry out interpolation, such as, when described quantity is 6 or 8, interpolation 1/2 pixel is very little with the correlation apart from its Integer Pel point farthest, even there is no correlation, can interpolation precision be affected; And if select the adjacent Integer Pel point of smallest number, such as, in Fig. 1, only select 2 adjacent Integer Pel points, then can lose the correlation information of some adjacent Integer Pel points, can interpolation precision be affected equally.
Further, inventor herein finds, concerning 1/2 pixel, four the Integer Pel points nearest apart from it are larger with its correlation, and wherein, described each Integer Pel point can be identical to the distance of this 1/2 pixel, also can be different.
In practice, the implementation of step 202 can comprise:
Sub-step A1, respectively for 1/2 pixel in Integer Pel point horizontal direction and vertical direction, utilize and respective direction carries out interpolation apart from the adjacent Integer Pel point of its nearest four;
With reference to Fig. 3, show the schematic diagram of a kind of chroma interpolation of the present invention, wherein, A ' is 1/2 pixel in Integer Pel point horizontal direction, comprises E, A, B, F apart from the adjacent Integer Pel point of its nearest four, and wherein, A and B is nearest apart from A ', E and F is recently secondary apart from A '; B ' is 1/2 pixel in Integer Pel point vertical direction, comprises G, A, C, H apart from the adjacent Integer Pel point of its nearest four, and wherein, A and C is nearest apart from B ', G and H is recently secondary apart from B '.
Sub-step A2, for 1/2 pixel in Integer Pel point diagonal, four the Integer Pel points nearest apart from it are utilized to carry out interpolation.
For Fig. 3,1/2 pixel in Integer Pel point diagonal, C ' is 1/2 pixel in Integer Pel point diagonal, and A, B, C, D are the Integer Pel point nearest apart from C '.
In specific implementation, 4 tap filters can be adopted, interpolation be carried out to four the Integer Pel points nearest apart from 1/2 pixel, obtains described 1/2 pixel; Filter factor wherein can be determined according to the position relationship of corresponding Integer Pel point and interpolation 1/2 pixel.
For the A ' in Integer Pel point horizontal direction, because A and B is nearest apart from A ', respective filter coefficient can be given with larger value; E and F is recently secondary apart from A ', can give respective filter coefficient with less value; Such as, adopt the filter factor of 4 tap filters can for (-1,5,5 ,-1), now, can adopt following 4 tap filter formula, interpolation obtains A ':
A′＝(-E+5*A+5*B-F+4)/8(1)
In like manner, for the B ' in Integer Pel point vertical direction, adopt 4 tap filters filter factor can be also (-1,5,5 ,-1), now, can adopt following 4 tap filter formula, interpolation obtains B ':
B′＝(-G+5*A+5*C-H+4)/8(2)
For the C ' in Integer Pel point diagonal, due to A, B, C, D to C ' distance equal, namely also weight is identical, therefore can adopt identical filter coefficient, be assumed to be (1,1,1,1), then can adopt following 4 tap filter formula, interpolation obtains C ':
C′＝(A+B+C+D+2)/4(3)
Be appreciated that above-mentioned (-1,5,5 ,-1) filter coefficient just exemplarily, those skilled in the art according to actual conditions, can adopt other filter coefficient, and such as, (1,7,7,1) etc., the present invention is not limited this.
Step 203, the Integer Pel point utilizing distance 1/4 pixel nearest and 1/2 pixel, interpolation obtains described 1/4 pixel;
Owing to utilizing four the Integer Pel points nearest apart from it, interpolation obtains described 1/2 pixel, two adjacent Integer Pel points are only adopted relative to prior art, or, although adopt four adjacent Integer Pel points, but these Integer Pel points are not nearest apart from described 1/2 pixel, the present invention can obtain accuracy and higher 1/2 pixel of precision, also just provides more high-precision reference pixel source for interpolation 1/4 pixel and 1/8 pixel.
Specific to interpolation 1/4 pixel of this step, its current reference pixel source comprises Integer Pel point and 1/2 pixel, in practice, and can according to 1/4 concrete pixel, determine to use which pixel in current reference pixel source, described step 203 specifically can comprise:
Sub-step B1, for 1/4 pixel in 1/2 pixel diagonal, utilize and carry out interpolation apart from its nearest Integer Pel point and three 1/2 pixels;
For Fig. 3, c ' be 1/4 pixel in 1/2 pixel diagonal, A is the Integer Pel point nearest apart from c ', and A ', B ', C ' be nearest 1/2 pixel of distance c '; Due to A, A ', B ', C ' be equal to the distance of C ', also namely weight is identical, therefore can adopt identical filter coefficient, and be assumed to be (1,1,1,1), in specific implementation, can adopt following 4 tap filter formula, interpolation obtains c ':
c′＝(A+A′+B′+C′+2)/4(4)
Sub-step B2, for 1/4 pixel in non-1/2 pixel diagonal, utilize apart from its nearest Integer Pel point and 1/2 pixel, or two 1/2 pixels, to carry out interpolation.
Such as, be apart from its nearest Integer Pel point for 1/4 pixel a ' A in non-1/2 pixel diagonal, A ' is 1/2 pixel nearest apart from it, and A, A ' equal to the distance of a ', also namely weight is identical, therefore identical filter coefficient can be adopted, be assumed to be (1,1); In specific implementation, adopt following 2 tap filter formula, interpolation obtains a ':
a′＝(A+A′+1)/2(5)
For 1/4 pixel d ' in non-1/2 pixel diagonal, B ', C ' are 1/2 pixel nearest apart from d ', and B ', C ' are equal to the distance of d ', and in like manner, can adopt following 2 tap filter formula, interpolation obtains d ':
d′＝(B′+C′+1)/2(6)
Step 204, the Integer Pel point, 1/2 pixel and 1/4 pixel that utilize distance 1/8 pixel nearest, interpolation obtains described 1/8 pixel.
Due to from the current reference pixel source comprising Integer Pel point and 1/2 pixel, what selection distance 1/4 pixel was nearest carries out interpolation, therefore, adjacent Integer Pel point is only selected relative to prior art, and these Integer Pel points are not nearest apart from this 1/4 pixel, the present invention can obtain accuracy and higher 1/4 pixel of precision, also just provides more high-precision reference pixel source for interpolation 1/8 pixel.
Specific to interpolation 1/8 pixel of this step, its current reference pixel source comprises Integer Pel point, 1/2 pixel and 1/4 pixel, in practice, can according to 1/8 concrete pixel, determine to use which pixel in current reference pixel source, described step 204 specifically can comprise following sub-step:
Sub-step C1, for 1/8 pixel in 1/4 pixel diagonal, utilize apart from its nearest Integer Pel point and three 1/4 pixels, or 1/2 pixel and three 1/4 pixels, to carry out interpolation;
In practice, for 1/8 pixel in 1/4 pixel diagonal, can adopt following 4 tap filter formula, interpolation obtains:
i＝(A+a′+b′+c′+2)/4(7)
Or, k=(B '+b '+c '+d '+2)/4 (8)
Wherein, i, k are 1/8 pixel in 1/4 pixel diagonal, and A is the Integer Pel point nearest apart from i, and a ', b ', c ' are 1/4 pixel nearest apart from i, and B ' is 1/2 pixel nearest apart from k, and b ', c ', d ' are 1/4 pixel nearest apart from k.
Sub-step C2, for 1/8 pixel in non-1/4 pixel diagonal, to utilize apart from its nearest Integer Pel point and 1/4 pixel, or, 1/2 pixel and 1/4 pixel, or two 1/4 pixels, carry out interpolation.
Such as, for the situation of an Integer Pel point and 1/4 pixel, the 2 tap filter formula of 1/8 pixel a can be:
a＝(A+a′+1)/2(9)
And for example, for the situation of 1/2 pixel and 1/4 pixel, the 2 tap filter formula of 1/8 pixel b can be:
b＝(A′+a′+1)/2(10)
For another example, for the situation of two 1/4 pixels, the 2 tap filter formula of 1/8 pixel e can be:
e＝(a′+c′+1)/2(11)
Wherein, a, i, b, e are 1/8 pixel, A, a ' be respectively apart from the nearest Integer Pel point of a and 1/4 pixel, A, c ' be respectively apart from the nearest Integer Pel point of i and 1/4 pixel, A ', a ' are respectively apart from 1/2 nearest pixel of b and 1/4 pixel, and a ', c ' are 1/4 pixel nearest apart from e.
Due to from the current reference pixel source comprising Integer Pel point, 1/2 pixel and 1/4 pixel, what selection distance 1/8 pixel was nearest carries out interpolation, therefore, adjacent Integer Pel point is only selected relative to prior art, and these Integer Pel points are not nearest apart from this 1/8 pixel, the present invention can obtain accuracy and higher 1/8 pixel of precision, thus can improve chroma interpolation precision.
It should be noted that, for the rectangular area of four adjacent Integer Pel point compositions, if its section of reaching the standard grade is positioned at the first row in interpolation region, such as, line segment AB in Fig. 3 is positioned at the first row in interpolation region, and interpolation B ' also needs to use Integer Pel point G, now, described Integer Pel point G can replace with Integer Pel point A;
Or its left line segment is positioned at the first row in interpolation region, such as, the line segment AC in Fig. 3 is positioned at the first row in interpolation region, and interpolation A ' also needs to use Integer Pel point E, and now, described Integer Pel point E can replace with Integer Pel point A.
Chrominance interpolation method of the present invention can be applied in video encoding-decoding process:
1, coding side;
When interframe encode, because the estimation of coding side is carried out for luminance block, therefore in the process of searching moving vector, do not use chroma interpolation; If when present frame needs reference frame as subsequent frame, so just need to utilize chroma motion vectors, interpolation obtains corresponding colorimetric prediction block.
With reference to Fig. 4, show a kind of application example at coding side of the present invention, wherein, reference frame Central Plains encoded chroma block is decoded, color heavy building block can be obtained; For described color heavy building block, according to the present invention, interpolation can be carried out successively to the Integer Pel point of input and obtains 1/2 pixel, 1/4 pixel and 1/8 pixel.
Further, those skilled in the art can estimate according to luma motion the motion vector that obtains, and in the reference frame after interpolation, the corresponding point block of pixels of search is used as the prediction block that current block is encoded.
2, decoding end.
After decoding end obtains motion vector, usually first carry out brightness interpolating, obtain brightness movement vector.
With reference to Fig. 5, show a kind of application example in decoding end of the present invention, this example operates for chrominance block, wherein, first can derive to brightness movement vector and obtain chrominance vector, then carry out corresponding interpolation to colourity and obtain predicting block; Owing to not relating to estimation in decode procedure, therefore, described brightness interpolating and chroma interpolation can carry out simultaneously, also, can obtain corresponding brightness and colorimetric prediction block simultaneously.
Corresponding to preceding method embodiment, the invention also discloses a kind of chroma interpolation device embodiment, with reference to Fig. 6, specifically can comprise:
Input module 601, for inputting the Integer Pel point in the interpolation region of reference frame;
1/2 pixel interpolating module 602, for utilizing four the Integer Pel points nearest apart from 1/2 pixel, interpolation obtains described 1/2 pixel;
1/4 pixel interpolating module 603, for utilizing apart from the nearest Integer Pel point of 1/4 pixel and 1/2 pixel, interpolation obtains described 1/4 pixel;
1/8 pixel interpolating module 604, for utilizing Integer Pel point, 1/2 pixel and 1/4 pixel nearest apart from 1/8 pixel, interpolation obtains described 1/8 pixel.
1/2 pixel interpolating module 602,
The reference pixel source of interpolation 1/2 pixel comprises Integer Pel point, and in practice, described 1/2 pixel interpolating module 602 may further include:
One 1/2 pixel interpolation submodule D1, for respectively for 1/2 pixel in Integer Pel point horizontal direction and vertical direction, utilizes and respective direction carries out interpolation apart from the adjacent Integer Pel point of its nearest four;
2 1/2 pixel interpolation submodule D2, for for 1/2 pixel in Integer Pel point diagonal, utilizes four the Integer Pel points nearest apart from it to carry out interpolation.
In specific implementation, described 1/2 pixel interpolating module 602 can be specifically for, adopts 4 tap filters, carries out interpolation, obtain described 1/2 pixel to four Integer Pel points nearest apart from 1/2 pixel.
Such as, described one 1/2 pixel interpolation submodule, can specifically for adopting following 4 tap filter formula, and interpolation obtains described 1/2 pixel:
A′＝(-E+5*A+5*B-F+4)/8
Wherein, A ' is 1/2 pixel in Integer Pel point horizontal direction or vertical direction, E, A, B, F and A ' and equidirectional, and A and B is the Integer Pel point nearest apart from A ', E and F is apart from the secondary nearest Integer Pel point of A '.
And for example, described 2 1/2 pixel interpolation submodule, can specifically for adopting following 4 tap filter formula, and interpolation obtains described 1/2 pixel:
C′＝(A+B+C+D+2)/4
Wherein, C ' is 1/2 pixel in Integer Pel point diagonal, and A, B, C, D are the Integer Pel point nearest apart from C '.
1/4 pixel interpolating module 603,
The reference pixel source of interpolation 1/4 pixel comprises Integer Pel point and 1/2 pixel, and the present invention can provide the following structon module of described 1/4 pixel interpolating module 603:
One 1/4 pixel interpolation submodule E1, for for 1/4 pixel in 1/2 pixel diagonal, utilizes and carries out interpolation apart from its nearest Integer Pel point and three 1/2 pixels;
Such as, described one 1/4 pixel interpolation submodule E1, can be used for adopting following 4 tap filter formula, interpolation obtains described 1/4 pixel:
c′＝(A+A′+B′+C′+2)/4
Wherein, c ' is 1/4 pixel in 1/2 pixel diagonal, and A is the Integer Pel point nearest apart from c ', and A ', B ', C ' are 1/2 pixel nearest apart from c '.
2 1/4 pixel interpolation submodule E2, for for 1/4 pixel in non-1/2 pixel diagonal, utilize apart from its nearest Integer Pel point and 1/2 pixel, or two 1/2 pixels, carries out interpolation.
And for example, described 2 1/4 pixel interpolation submodule E2, can be used for adopting following 2 tap filter formula, interpolation obtains described 1/4 pixel:
A '=(A+A '+1)/2, or, d '=(B '+C '+1)/2
Wherein, a ', d ' are 1/4 pixel in non-1/2 pixel diagonal, and A is the Integer Pel point nearest apart from a ', and A ' is 1/2 pixel nearest apart from a ', and B ', C ' are 1/2 pixel nearest apart from d '.
1/8 pixel interpolating module 604,
The reference pixel source of interpolation 1/8 pixel comprises Integer Pel point, 1/2 pixel and 1/4 pixel, and the present invention can provide the following project organization of described 1/8 pixel interpolating module 604:
According to 1/8 concrete pixel, determine to select which pixel from current reference pixel source, now, following submodule can be designed in described 1/8 pixel interpolating module:
One 1/8 pixel interpolation submodule F1, for for 1/8 pixel in 1/4 pixel diagonal, utilize apart from its nearest Integer Pel point and three 1/4 pixels, or 1/2 pixel and three 1/4 pixels, carries out interpolation;
In specific implementation, described one 1/8 pixel interpolation submodule F1, can be used for adopting following 4 tap filter formula, interpolation obtains described 1/8 pixel:
I=(A+a '+b '+c '+2)/4, or, k=(B '+b '+c '+d '+2)/4
Wherein, i, k are 1/8 pixel in 1/4 pixel diagonal, and A is the Integer Pel point nearest apart from i, and a ', b ', c ' are 1/4 pixel nearest apart from i, and B ' is 1/2 pixel nearest apart from k, and b ', c ', d ' are 1/4 pixel nearest apart from k.
2 1/8 pixel interpolation submodule F2, for for 1/8 pixel in non-1/4 pixel diagonal, utilizes apart from its nearest Integer Pel point and 1/4 pixel, or, 1/2 pixel and 1/4 pixel, or, two 1/4 pixels, carry out interpolation.
The course of work of described 2 1/8 pixel interpolation submodule F2 can be that adopt following 2 tap filter formula, interpolation obtains described 1/8 pixel:
A=(A+a '+1)/2, or, b=(A '+a '+1)/2, or, e=(a '+c '+1)/2
Wherein, a, b, e are 1/8 pixel in non-1/4 pixel diagonal, A, a ' be respectively apart from the nearest Integer Pel point of a and 1/4 pixel, A ', a ' are respectively apart from 1/2 nearest pixel of b and 1/4 pixel, and a ', c ' they are 1/4 pixel nearest apart from e.
For device embodiment, due to itself and embodiment of the method basic simlarity, so description is fairly simple, relevant part illustrates see the part of embodiment of the method.
Above to a kind of chrominance interpolation method provided by the present invention and device, be described in detail, apply specific case herein to set forth principle of the present invention and execution mode, the explanation of above embodiment just understands method of the present invention and core concept thereof for helping; Meanwhile, for one of ordinary skill in the art, according to thought of the present invention, all will change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.
Claims (4)
1. a chrominance interpolation method, is characterized in that, comprising:
The Integer Pel point in the interpolation region of input reference frame;
Utilize four the Integer Pel points nearest apart from 1/2 pixel, interpolation obtains described 1/2 pixel;
Described interpolation obtains the step of described 1/2 pixel, comprising:
Respectively for 1/2 pixel in Integer Pel point horizontal direction and vertical direction, utilize and respective direction carries out interpolation apart from the adjacent Integer Pel point of its nearest four;
For 1/2 pixel in Integer Pel point diagonal, four the Integer Pel points nearest apart from it are utilized to carry out interpolation;
Utilize apart from the nearest Integer Pel point of 1/4 pixel and 1/2 pixel, interpolation obtains described 1/4 pixel;
Described interpolation obtains the step of described 1/4 pixel, comprising:
For 1/4 pixel in 1/2 pixel diagonal, utilize and carry out interpolation apart from its nearest Integer Pel point and three 1/2 pixels;
For 1/4 pixel in non-1/2 pixel diagonal, utilize apart from its nearest Integer Pel point and 1/2 pixel, or two 1/2 pixels, carry out interpolation;
Utilize Integer Pel point, 1/2 pixel and 1/4 pixel nearest apart from 1/8 pixel, interpolation obtains described 1/8 pixel;
The step that described interpolation obtains described 1/8 pixel comprises:
For 1/8 pixel in 1/4 pixel diagonal, utilize apart from its nearest Integer Pel point and three 1/4 pixels, or 1/2 pixel and three 1/4 pixels, carry out interpolation;
For 1/8 pixel in non-1/4 pixel diagonal, utilize apart from its nearest Integer Pel point and 1/4 pixel, or, 1/2 pixel and 1/4 pixel, or two 1/4 pixels, carry out interpolation.
2. the method for claim 1, is characterized in that, described interpolation obtains the step of described 1/2 pixel, comprising:
Adopt 4 tap filters, interpolation is carried out to four the Integer Pel points nearest apart from 1/2 pixel, obtains described 1/2 pixel.
3. a chroma interpolation device, is characterized in that, comprising:
Input module, for inputting the Integer Pel point in the interpolation region of reference frame;
1/2 pixel interpolating module, for utilizing four the Integer Pel points nearest apart from 1/2 pixel, interpolation obtains described 1/2 pixel;
Described 1/2 pixel interpolating module comprises:
One 1/2 pixel interpolation submodule, for respectively for 1/2 pixel in Integer Pel point horizontal direction and vertical direction, utilizes and respective direction carries out interpolation apart from the adjacent Integer Pel point of its nearest four;
2 1/2 pixel interpolation submodule, for for 1/2 pixel in Integer Pel point diagonal, utilizes four the Integer Pel points nearest apart from it to carry out interpolation.
1/4 pixel interpolating module, for utilizing apart from the nearest Integer Pel point of 1/4 pixel and 1/2 pixel, interpolation obtains described 1/4 pixel;
Described 1/4 pixel interpolating module comprises:
One 1/4 pixel interpolation submodule, for for 1/4 pixel in 1/2 pixel diagonal, utilizes and carries out interpolation apart from its nearest Integer Pel point and three 1/2 pixels;
2 1/4 pixel interpolation submodule, for for 1/4 pixel in non-1/2 pixel diagonal, utilize apart from its nearest Integer Pel point and 1/2 pixel, or two 1/2 pixels, carries out interpolation;
1/8 pixel interpolating module, for utilizing Integer Pel point, 1/2 pixel and 1/4 pixel nearest apart from 1/8 pixel, interpolation obtains described 1/8 pixel;
Described 1/8 pixel interpolating module comprises:
One 1/8 pixel interpolation submodule, for for 1/8 pixel in 1/4 pixel diagonal, utilize apart from its nearest Integer Pel point and three 1/4 pixels, or 1/2 pixel and three 1/4 pixels, carries out interpolation;
2 1/8 pixel interpolation submodule, for for 1/8 pixel in non-1/4 pixel diagonal, utilizes apart from its nearest Integer Pel point and 1/4 pixel, or, 1/2 pixel and 1/4 pixel, or, two 1/4 pixels, carry out interpolation.
4. device as claimed in claim 3, is characterized in that, described 1/2 pixel interpolating module specifically for, adopt 4 tap filters, interpolation carried out to four the Integer Pel points nearest apart from 1/2 pixel, obtains described 1/2 pixel.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010211400.9A CN101888558B (en) | 2010-06-18 | 2010-06-18 | A kind of chrominance interpolation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010211400.9A CN101888558B (en) | 2010-06-18 | 2010-06-18 | A kind of chrominance interpolation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101888558A CN101888558A (en) | 2010-11-17 |
| CN101888558B true CN101888558B (en) | 2016-02-10 |
Family
ID=43074239
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010211400.9A Active CN101888558B (en) | 2010-06-18 | 2010-06-18 | A kind of chrominance interpolation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101888558B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103686029B (en) * | 2013-11-29 | 2016-09-28 | 浙江大学 | A kind of video pixel opposite-angle interpolation method for the many optical projection systems of giant-screen |
| CN104598929A (en) * | 2015-02-03 | 2015-05-06 | 南京邮电大学 | HOG (Histograms of Oriented Gradients) type quick feature extracting method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1543221A (en) * | 2003-11-04 | 2004-11-03 | 浙江大学 | Method and apparatus for 1/4 pixel precision interpolation |
| CN101141657A (en) * | 2007-07-24 | 2008-03-12 | 浙江大学 | Video image chroma interpolation method and device |
| EP1919212A1 (en) * | 2006-10-19 | 2008-05-07 | Samsung Electronics Co., Ltd. | Method and apparatus for interpolating chroma signal to minimize operation overhead |
-
2010
- 2010-06-18 CN CN201010211400.9A patent/CN101888558B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1543221A (en) * | 2003-11-04 | 2004-11-03 | 浙江大学 | Method and apparatus for 1/4 pixel precision interpolation |
| EP1919212A1 (en) * | 2006-10-19 | 2008-05-07 | Samsung Electronics Co., Ltd. | Method and apparatus for interpolating chroma signal to minimize operation overhead |
| CN101141657A (en) * | 2007-07-24 | 2008-03-12 | 浙江大学 | Video image chroma interpolation method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101888558A (en) | 2010-11-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI808365B (en) | Video data stream, video encoder, apparatus and methods for hrd timing fixes, and further additions for scalable and mergeable bitstreams | |
| CN105359531B (en) | Method and system for determining for the coder side of screen content coding | |
| CN104363451B (en) | Image prediction method and relevant apparatus | |
| CN101454990B (en) | Video compression method | |
| CN100551073C (en) | Decoding method and device, image element interpolation processing method and device | |
| CN102611884B (en) | Image encoding and decoding method and encoding and decoding device | |
| CN112997484A (en) | Multi-parameter based intra prediction | |
| TW202025752A (en) | History-based motion vector prediction for affine mode | |
| KR101868247B1 (en) | Image encoding and decoding method and device | |
| WO2011125256A1 (en) | Image encoding method and image decoding method | |
| CN109889850A (en) | Video coding-decoding method | |
| CN107105263B (en) | The video encoding/decoding method and device of motion compensation for being executed under inter-prediction | |
| CN101883286A (en) | Calibration method and device, and motion estimation method and device in motion estimation | |
| JP2022502912A (en) | Image coding / decoding method and equipment | |
| CN101765011A (en) | Method and device for scaling motion estimation | |
| CN101888558B (en) | A kind of chrominance interpolation method and device | |
| JP2022525943A (en) | Optical flow-based motion compensation | |
| CN104519363A (en) | Video encoding/decoding methods, corresponding computer programs and video encoding/decoding devices | |
| CN101931739A (en) | Absolute error sum estimation system and method | |
| JP2023528609A (en) | Encoding/decoding method, apparatus and device | |
| KR102606880B1 (en) | Encoders, decoders and corresponding methods for inter prediction | |
| TW202133619A (en) | History-based motion vector predictor constraint for merge estimation region | |
| CN112565768A (en) | Inter-frame prediction method, encoding and decoding system and computer readable storage medium | |
| CN103139561A (en) | Interpolation filter for half pixel and quarter sub-pixel | |
| CN104581186B (en) | Method for coding and decoding intra-frame motion vector in image compression |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| ASS | Succession or assignment of patent right |
Owner name: WUXI VIMICRO ELECTRONICS CO., LTD. Free format text: FORMER OWNER: BEIJING ZHONGXING MICROELECTRONICS CO., LTD. Effective date: 20110617 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| COR | Change of bibliographic data |
Free format text: CORRECT: ADDRESS; FROM: 100083 16/F, SHINING BUILDING, NO. 35, XUEYUAN ROAD, HAIDIAN DISTRICT, BEIJING TO: 214028 610, NATIONAL INTEGRATED CIRCUIT DESIGN PARK (CHUANGYUAN BUILDING), NO. 21-1, CHANGJIANG ROAD, WUXI NEW DISTRICT, JIANGSU PROVINCE |
|
| TA01 | Transfer of patent application right |
Effective date of registration: 20110617 Address after: 214028 national integrated circuit design (21-1), Changjiang Road, New District, Jiangsu, Wuxi, China, China (610) Applicant after: Wuxi Vimicro Co., Ltd. Address before: 100083 Haidian District, Xueyuan Road, No. 35, the world building, the second floor of the building on the ground floor, No. 16 Applicant before: Beijing Vimicro Corporation |
|
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: Wuxi District of Jiangsu city of Wuxi province Qingyuan Road 214000 No. 18 Taihu International Science Park sensor network university science and Technology Park 530 building A1001 Applicant after: WUXI ZHONGGAN MICROELECTRONIC CO., LTD. Address before: 214028 national integrated circuit design (21-1), Changjiang Road, New District, Jiangsu, Wuxi, China, China (610) Applicant before: Wuxi Vimicro Co., Ltd. |
|
| COR | Change of bibliographic data | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |