CN102547256B - Method and system for removing self-adaptive block effect - Google Patents
Method and system for removing self-adaptive block effect Download PDFInfo
- Publication number
- CN102547256B CN102547256B CN201010582772.2A CN201010582772A CN102547256B CN 102547256 B CN102547256 B CN 102547256B CN 201010582772 A CN201010582772 A CN 201010582772A CN 102547256 B CN102547256 B CN 102547256B
- Authority
- CN
- China
- Prior art keywords
- pixel
- new piece
- present frame
- frame
- new
- 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.)
- Expired - Fee Related
Links
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention provides a method for removing self-adaptive block effect, which includes partitioning a current frame of input video again to enable two central axes of each novel block which is divided again to right be borders of original blocks in the horizontal direction and the vertical direction of four adjacent original blocks; calculating difference value of pixels on two sides of the borders of the original blocks and difference value of pixels in the original blocks on two sides; and determining whether the calculated the difference value of the pixels on two sides of the borders of the original blocks is larger than the difference value of the pixels in the original blocks on two sides of the borders of the original blocks and whether the difference value of the pixels on two sides of the borders of the original blocks exceeds a predefined threshold. If the difference value of the pixels on two sides of the borders of the original blocks is larger than the e difference value of the pixels in the original blocks on two sides of the borders of the original blocks and the difference value of the pixels on two sides of the borders of the original blocks does not exceed the predefined threshold, an optimal matching novel block is sought in adjacent frames before and after the current frame, and the found optimal matching novel block in the adjacent frames is adopted to update the novel block in the current frame.
Description
Technical field
The present invention relates to a kind of blocking effect for video compression and video enhancing field and remove method and system, the method and system of blocking effect in a kind of adaptive removal vision signal based on time domain and spatial domain is especially provided.
Background technology
In video being divided to the video compression and video enhancing process of block operations, frame of video can produce a kind of blocking effect between piece and piece, at human eye, can see a kind of mosaic phenomenon.In order to eliminate this blocking effect, prior art has proposed the mode of various elimination blocking effects, wherein just comprises the technological means based on time domain and spatial domain elimination blocking effect.
US Patent No. 5512956A has just proposed a kind of post-processing approach to the image sequence of low bit rate based on time domain and spatial domain and has eliminated blocking effect.In this patent, the separable three-dimensional filter structure of the nonlinear filter of a kind of finite impulse based on spatial domain and intermediate value compound filter and the motion compensation based on time domain has been proposed, but this structure only limits to median filter in time-domain filtering, and do not consider the complexity of the content of frame of video completely, and need to all will take identical strategy one time to all pixels of frame of video.In addition, this patent has been introduced motion vector, and motion vector is associated with codec.
In Chinese patent application CN101494787A, a kind of method of the deblocking effect detecting based on blocking effect has been proposed.But this patent application has just used consecutive frame to carry out blocking effect detection, but do not utilize this information pixel value to recalculate.Especially, the detection of image true edge information is not considered in this patent application, thereby can cause image to seem excessively fuzzy after removing blocking effect.
In the prior art, the method that weighting that a kind of use is symmetrically distributed in the point on block edge removes blocking effect has also been proposed, but the method need to be taked identical strategy to all pixels in whole frame, and do not consider equally the difference of the content complexity of different masses, and the element that affects of considering also only relates to three pixels.Especially this method does not consider to eliminate the problem of blocking effect at all from the continuity of video time domain.Another kind of prior art has proposed a kind of fuzzy filter and has reduced the noise problem that affects in the multi-media signal having compressed.The filter of its use just utilize former frame, present frame and a rear frame fixing centered by calculating pixel to the pixel application filtering in the scope of external expansion certain width and height.Just say, although it has related to time domain problem, it does not expect the motion problems that in piece, object exists between different frame at all yet.
With the method for mentioning in prior art, all adopted the method based on spatial domain to remove blocking effect and do not consider the association of vision signal in time domain direction.To the processing on spatial domain, said method is all that the block boundary of horizontal direction and vertical direction is processed respectively.For different images content, the degree of blocking effect is actual in human eye is different, takes the strategy of unified level and smooth deblocking effect if therefore do not consider the difference of content for blocking effect, can cause the loss of the too fuzzy or detailed information of picture material.
Summary of the invention
In order to solve the above-mentioned technical problem existing in prior art, the present invention has gone out a kind of adaptive block effect removal method, comprise: to input video present frame again piecemeal, making each repartitioned two central shaft of new is just the horizontal direction of four original blocks and the original block border of vertical direction adjacent one another are; Calculate the difference of both sides pixel on described original block border and the difference of the pixel of original block inside, both sides; The difference of determining the described original block boundaries on either side pixel calculate whether be greater than described original block boundaries on either side original block inside pixel difference and and the difference of described original block boundaries on either side pixel whether surpass predefine threshold value; If being the difference of pixel and the difference of described original block boundaries on either side pixel that are greater than described original block boundaries on either side original block inside, the difference of described original block boundaries on either side pixel do not surpass predefine threshold value, in the consecutive frame of the front and back of present frame, search the new piece of optimum Match, adopt the new piece of optimum Match in the consecutive frame find to upgrade described new of present frame.
The adaptive block effect removal method according to the present invention, described step of searching the new piece of optimum Match in the consecutive frame of the front and back of present frame comprises: in the consecutive frame of the front and back of present frame, find the new piece with the consecutive frame of the new piece correspondence position of present frame, and to take the corresponding Xin Kuai center of consecutive frame be the center of circle, with the radius of being scheduled to, at front and back consecutive frame, the new piece of described correspondence is expanded, obtained and take the extended area that described corresponding Xin Kuai center is the center of circle; Before and after calculating, in consecutive frame, form each the new of described extended area absolute value sum with the difference of the gray value of the respective pixel of the new piece of present frame; And the new piece in the calculated minimum corresponding extended area of absolute value sum is defined as to the new piece of optimum Match in the front and back consecutive frame of new piece of present frame.
Above-mentioned technological means, when carrying out time domain processing, considers that the interior object of new piece exists the situation of moving displacement in time domain, therefore compares with the method for processing according to absolute coordinate merely, and the accuracy of the result that it obtains can be significantly improved.
According to the present invention, adaptive block effect removal method also comprises: before the new piece of present frame is upgraded, select the region of the required renewal in the new piece of described present frame according to the content complexity of the new piece of described present frame.
The adaptive block effect removal method according to the present invention, the step in the region of the required renewal in the new piece of the described present frame of described selection also comprises: according to the content complexity of the new piece of present frame, calculate the required step-length of expansion to the periphery of two central shafts by the new piece of described present frame; The cross region that the step length of calculating to external expansion along described two central shaft crosspoints is formed is defined as region to be updated;
According to the present invention, adaptive block effect removal method also comprises: before in adopting the consecutive frame finding, the new piece of optimum Match upgrades the step of described new of present frame, by following step, on spatial domain, recalculate the pixel value of each pixel to be updated in described region to be updated: for described each pixel to be updated, in described region to be updated, obtain the pixel set of the described pixel to be updated of impact; And the distance property of the pixel set based on obtained and described pixel to be updated and grey value difference attribute recalculate the weighted pixel value of described pixel to be updated.
The adaptive block effect removal method according to the present invention, the step of described new that in the consecutive frame that described employing is found, the new piece of optimum Match upgrades present frame is undertaken by the pixel value three of the respective pixel of the new piece of optimum Match of the new piece of present frame and determined front and back consecutive frame being weighted to summation.
According to another aspect of the present invention, provide a kind of adaptive block effect to remove system, comprise: piecemeal module again, for the present frame of input video is carried out to piecemeal again, making each repartitioned two central shaft of new is just the horizontal direction of four original blocks and the original block border of vertical direction adjacent one another are; Computing module, for calculating the difference of both sides pixel on described original block border and the difference of the pixel of original block inside, both sides; Determination module, for determine the difference of the described original block boundaries on either side pixel calculate whether be greater than described original block boundaries on either side original block inside pixel difference and and the difference of described original block boundaries on either side pixel whether surpass predefine threshold value; Search module, if the difference for described original block boundaries on either side pixel is that the difference of pixel and the difference of described original block boundaries on either side pixel that is greater than described original block boundaries on either side original block inside do not surpass predefine threshold value, in the consecutive frame of the front and back of present frame, search the new piece of optimum Match; And update module, adopt the new piece of optimum Match in the consecutive frame find to upgrade described new of present frame.
According to adaptive block effect of the present invention, remove system, the described module of searching comprises: extended area obtains module, in the consecutive frame of the front and back of present frame, find the new piece with the consecutive frame of the new piece correspondence position of present frame, and to take the corresponding Xin Kuai center of consecutive frame be the center of circle, with the radius of being scheduled to, at front and back consecutive frame, the new piece of described correspondence is expanded, obtained and take the extended area that described corresponding Xin Kuai center is the center of circle; Gray value differences absolute value and computing module, form each the new of described extended area absolute value sum with the difference of the gray value of the respective pixel of the new piece of present frame in consecutive frame before and after calculating; And the new piece determination module of optimum Match, the new piece in the calculated minimum corresponding extended area of absolute value sum is defined as to the new piece of optimum Match in the front and back consecutive frame of new piece of present frame.
According to adaptive block effect of the present invention, remove system, also comprise that more new region is selected module, according to the content complexity of the new piece of described present frame, select the region of the required renewal in the new piece of described present frame.
According to adaptive block effect of the present invention, remove system, it also comprises that pixel value recalculates module, its new piece of optimum Match in adopting the consecutive frame find upgrade described new of present frame before, on spatial domain, recalculate the pixel value of each pixel to be updated in described region to be updated.
The technical scheme of the present invention of above-mentioned concise and to the point narration according to the present invention, the present invention proposes a kind of method of the adaptive removal blocking effect based on time domain.In the present invention, frame of video is carried out to piecemeal again, the block boundary information of new piece and original block formed objects and the horizontal direction that comprises original block and vertical direction.To each new, at the front and back of this frame neighbours' frame, find the new piece of optimum Match, and different according to the content complexity of piece, determine different pixel set to be updated, rather than all pixels of new piece are all upgraded, this can greatly reduce the quantity of upgrading pixel the speed of mentioning Video processing.And, while frame of video being carried out to deblocking effect processing on spatial domain simultaneously, to each pixel to be updated, by consider that the original block border of horizontal direction and vertical direction obtains the set of pixels that affects this renewal pixel simultaneously, and according to obtain different pixels influence degree difference to pixel to be updated in set of pixels, especially according to obtain distance property and grey value difference of different pixels and pixel to be updated in set of pixels, carry out the renewal on spatial domain, therefore it is more accurate that the present invention can make deblocking effect acquisition results for video, and processing procedure is also more flexible.
Accompanying drawing explanation
Shown in Fig. 1 is according to the flow chart of the self adaptation deblocking effect method based on time domain and spatial domain of the present invention.
Shown in Fig. 2 is input video frame to be reclassified as to the schematic diagram of the process of new piece.
Shown in Fig. 3 is in the frame of the adjacent front and back of current video frame, to find the schematic diagram of the process of best matching blocks.
Shown in Fig. 4 is according to content complexity, to select the concrete more schematic diagram of the process of new region from new piece.
Shown in Fig. 5 is on spatial domain, for each pixel to be updated, to obtain the schematic diagram of the process that affects the set of the pixel of its pixel value.
Embodiment
For making the object, technical solutions and advantages of the present invention clearer, below in conjunction with embodiment and accompanying drawing, the present invention is described in more detail.
Shown in Fig. 1 is according to the flow chart of the self adaptation deblocking effect method based on time domain and spatial domain of the present invention.Self adaptation deblocking effect method based on time domain and spatial domain of the present invention is moved at a kind of image processing system, and this system is also referred to as deblocking effect system below.As shown in Figure 1, according to the self adaptation deblocking effect method based on time domain and spatial domain of the present invention, first at step S10 place, receive the video of inputting, and the current frame of video of inputting is carried out to piecemeal again.A common two field picture consists of a plurality of image blocks.A general image block consists of 8 * 8 pixels.Shown in Fig. 2 is input video frame to be reclassified as to the schematic diagram of the process of new piece.As shown in Figure 2, the original block that each minimum solid line square is frame of video, each minimum dotted line square is the new piece after repartitioning, each new has same size size with each original picture block.As shown in Figure 2, the original block border (solid line) of the inner horizontal direction of four original picture block adjacent one another are and vertical direction has formed two central shafts of a new piece of repartitioning, and in this new piece of illustrating separately at the right-hand component of Fig. 2, shows.Two central shafts that are new piece are just the block boundary of the horizontal direction of original picture block and vertical direction, the position that blocking effect occurs conventionally.By this mode of repartitioning, make to repartition rear obtained new piece comprise original picture block in the horizontal direction with the block boundary information of vertical direction.
Then, at step S11 place, whether the relevant calculating in deblocking effect system and determination module judge each new in described present frame, to determine whether described new need to carry out deblocking effect processing, namely need this new to carry out to upgrade to process.Before carrying out deblocking effect processing, first need whether to have blocking effect in definite described new.But in order to prevent that the real border in image is judged as to blocking effect, namely in order to retain the real border information of image, first need to judge whether the horizontal direction of original block and the block boundary of vertical direction are the real border information of image.While only having the real border information that this new block boundary comprising is not image, just need to implement to remove to new piece the processing of blocking effect.For this reason, at step S11 place, determine whether described new need to carry out deblocking effect and process and specifically by the pixel value difference of computation bound both sides block, to carry out.Detailed process is as follows.
Central shaft along new piece is the horizontal direction of original block and the block boundary of vertical direction, respectively the value differences of computation bound both sides pixel and the pixel value difference in two blocks.The pixel value difference of if block boundaries on either side is greater than the pixel value difference of both sides block and surpasses predetermined threshold T1, the real border information that this border is image.Provided by way of example the program of this algorithm below.
// to horizontal boundary
int?count[7]＝{0}；
For?i＝0?to?7
for?j＝0?to?7
{
if(|P
j，3-P
j，4|＞|P
j，3-P
j，2|)&&(|P
j，3-P
j，4|＞|P
j，5-P
j，4|)&&(|P
j，3-
P
j，4|＜threshold?1)
count[i]++；
else?break；
}
If?Max(count[])＞threshold2
return?true?edge；
//
Afterwards same processing carried out in vertical border.For saving space, therefore omit this description.
By above-mentioned processing, can retain and belong to real border in image, avoid unnecessary renewal.
If the border in the new piece of step S11 place judgement present frame is not the real border in image, at step S12 place, carry out following operation: in the consecutive frame of the front and back of present frame, search the new piece of optimum Match about the new piece of present frame, i.e. step S121 in Fig. 1 and S122.In step S121, in the last consecutive frame of present frame, search the new piece of optimum Match about the new piece of present frame, and in step S122, in a rear consecutive frame of present frame, search the new piece of optimum Match about the new piece of present frame.Why adopt this technological means to be because in time domain to existing the border of blocking effect to upgrade more accurately, thereby eliminate blocking effect.In view of the object of the boundary of the new piece of present frame likely moves, it is inadequate therefore according to same coordinate, looking for corresponding blocks in the consecutive frame of front and back merely.
Shown in Fig. 3 is in the frame of the adjacent front and back of current video frame, to find the schematic diagram of the process of best matching blocks.As shown in Figure 3, shown in the left side is the schematic diagram of present frame, and wherein (x, y) represent to form the pixel set of new piece, and shown in the right is the schematic diagram of consecutive frame, and wherein (x ', y ') represents a possible best matching blocks in consecutive frame.In Fig. 3, show, position in consecutive frame of the position of the new piece (x, y) in present frame in present frame and the piece in consecutive frame (x ', y ') is also not exclusively corresponding, and this position difference demonstrates movement between the two.In order more accurately the new piece that need to remove in the present frame of blocking effect to be upgraded in time domain, consider the possibility that existence is moved in consecutive frame in time domain of object in new piece, therefore, the present invention is in consecutive frame, take the consecutive frame Zhong Kuai center completely corresponding with the position of the new piece of present frame is the center of circle, (n is positive integer to take predefine step-length n, being generally 1-20) individual pixel distance is radius, in front and back, neighbours' frame is expanded this piece, thereby obtain the candidate matches piece of the new piece formed objects of a plurality of and present frame, there is each other lap in these candidate matches pieces.
After candidate matches piece in obtaining consecutive frame, calculate the absolute value and, i.e. sad value of difference of the gray value of the respective pixel between each candidate matches piece and the new piece of present frame.Compare subsequently sad value between each calculated candidate matches piece and the new piece of present frame, and the candidate matches piece with corresponding front and back two frames of minimum sad value is defined as to two new pieces of optimum Match.Concrete calculation procedure is as follows:
1.x′＝x+n；y′＝y+n(n∈(-step；0；step))
2. to each (x ', y ')
Calculate new piece in present frame and think the sad value of the piece B (x ', y ') at (x ', y ') center
If SAD < is T3, stop;
Otherwise, the B (x ', y ') with minimum SAD is required.
Wherein, (B (x ', y ') represents the candidate matches piece centered by (x ', y ').SAD be the respective pixel between any candidate matches piece and the new piece of present frame gray value difference absolute value and.Obviously, as mentioned above, thereby the amount of calculation of searching best matching blocks in order to reduce improves computational speed, the present invention is based on certain experience, be provided with a predefine threshold value T3, and when SAD is less than this threshold value between a candidate matches piece and the new piece of present frame arbitrarily, just directly the corresponding candidate matches piece of this SAD be defined as to the best matching blocks in consecutive frame, thereby stop the detection computations to follow-up candidate matches piece, as described in step 2 above.
In view of human eye for the susceptibility of blocking effect the difference of the complexity based on image and difference, particularly, for the level and smooth image of content, human eye is pair higher with the susceptibility of blocking effect, image for content complexity, human eye is lower to the susceptibility of blocking effect, therefore, all pixels of the new piece upgrading for fear of the needs in present frame are all upgraded the too fuzzy situation of image causing, when the present invention passes through the step S121 and S122 in execution step S12, also need to perform step S123, , according to the content complexity of the new piece of present frame (smoothness), select pixel set to be updated or the region of required renewal.From the new piece of present frame, select the process of pixel set to be updated as follows:
First, all pixel { P in the new piece of calculating present frame
ijthe average gray value μ of (i, j=0.....N-1),
Then,, according to following formula, calculate the new piece variance of present frame, i.e. all pixels of the new piece of present frame and the mean value σ of the difference side between average gray value μ
2, i.e. the complexity of the new piece of present frame.
Afterwards, the complexity σ of the new piece of the present frame based on calculated
2, calculate the abducent step-length of cross being formed by level and vertical block boundary and count step.
step＝max{1，log
10(σ
2+1)}
Finally, the cross region obtaining according to calculating gained step-length number obtains region to be calculated.
Shown in Fig. 4 is according to content complexity, to select the concrete more schematic diagram of the process of new region from new piece.As shown in Figure 4, different content complexities, has different expansion step-lengths, and wherein content is more complicated, and its expansion step-length is less, and the region to be updated forming is less; Content is more level and smooth, and the step-length of expansion is larger, and the region to be updated becoming is afterwards just larger.By the region selected to be updated corresponding to complex contents that Fig. 4 is shown, can know, the size in selected region to be updated be significantly less than new piece size, therefore, the quantity of the pixel to be updated significantly reducing, thereby the speed of the renewal improving, and, avoid the too fuzzy situation of complicated image piece after image, thereby retained as far as possible the detailed information of image.
In step S12, can also as required after step S123, at step S124 place, on spatial domain, recalculate the pixel value in selected region to be updated, first on spatial domain, carry out deblocking effect processing.Describedly on spatial domain, carry out deblocking effect processing and carry out as follows:
First, for each pixel to be updated in selected region to be updated, obtain a pixel set that affects this pixel to be updated on spatial domain.Concrete mode is exactly that the set of pixels of the described pixel to be updated of impact is combined into and take the rectangular area of described pixel to be updated centered by a summit and Yi Xinkuai center.Shown in Fig. 5 is on spatial domain, for each pixel to be updated, to obtain the schematic diagram of the process that affects the set of the pixel of its pixel value.As shown in Figure 5, wherein P represents any one pixel to be updated in selected region to be updated, and the pixel that the indicated solid-line rectangle frame of S (p) comprises is for affecting the pixel set of pixel P.
Then, for each pixel P to be updated, adopt following formula, according to affecting the distance property of each pixel and pixel P to be updated in pixel set and the determined weighted value of its gray value of grey value difference attribute, carry out weighted sum and calculate, thereby obtain the pixel value recalculating of pixel P to be updated.
Wherein, De is the distance of two pixels,
PD is the grey scale pixel value component of distance property direction,
PG is the pixel grey scale component in grey value difference direction.
Above-mentioned steps S124 realizes essential to the inventionly, but for the improvement that the block effect in time domain is processed to the present invention, thereby can obtain the effect of more perfect deblocking effect.
Then, at step S13 place, the best matching blocks of the front and back consecutive frame that utilization is found at step S12 place calculates the final pixel value of pixel to be updated in time domain.Concrete mode is as follows:
To selected each pixel to be updated in the new piece of present frame, its end value is obtained by the pixel value three weighting of the correspondence position pixel of the best matching blocks of himself pixel value two frames adjacent with front and back.
p
i，j＝α*p′
i，j+β*p
i，j+γ*p″
i，j
Wherein, p '
i, jand p "
i, jthe pixel value of the respective pixel in former frame and a rear frame the inside best matching blocks, and alpha+beta+γ=1, empirically, α=0.15, β=0.7, γ=0.15, this weighted value can rule of thumb be adjusted, as long as alpha+beta+γ=1, α=0.2 for example, β=0.6, γ=0.2; α=0.10, β=0.7, γ=0.2 etc.
By the technical scheme of the present invention of foregoing description, can utilize the time-domain information of vision signal, thus the more accurate result of the inconsistent acquisition of information that search optimum Match causes with the motion of avoiding due to object.And, owing to not adopting according to absolute location coordinates, obtain time domain respective pixel, the result obtaining by search best matching blocks is like this more accurate.Meanwhile, this process is last handling process, irrelevant with codec.
And, the block boundary of the present invention's while processing horizontal and vertical direction, thus the pixel of region of expansion as a reference rather than on difference processing horizontal and vertical two lines obtains more accurate result.
Especially, the present invention's selectable processing section pixel rather than travel through all pixels, the scope of pixel is adaptive to the content complexity of place piece.Therefore the degree of blocking effect is different in the zones of different of a frame of video, selects different pixel set according to content complexity, can effectively avoid so too fuzzy while of image to retain as much as possible the detailed information of image.And the present invention considers the Different Effects of the different attribute of surrounding pixel to pixel simultaneously, for each pixel to be updated, the different pixel of attribute is also different to the influence degree of this pixel.
The sequence of operations illustrating in specification can be carried out by the combination of hardware, software or hardware and software.When carrying out this sequence of operations by software, computer program wherein can be installed in the memory in the computer that is built in specialized hardware, make computer carry out this computer program.Or, computer program can be installed in the all-purpose computer that can carry out various types of processing, make computer carry out this computer program.
For example, can computer program is pre-stored in the hard disk or ROM (read-only memory) of recording medium.Or, can be temporarily or for good and all storage (record) computer program in removable recording medium, such as floppy disk, CD-ROM (compact disc read-only memory), MO (magneto-optic) dish, DVD (digital versatile disc), disk or semiconductor memory.So removable recording medium can be provided as canned software.
The present invention has been described in detail with reference to specific embodiment.Yet clearly, in the situation that not deviating from spirit of the present invention, those skilled in the art can carry out change and replace embodiment.In other words, the present invention is open by the form of explanation, rather than is limited to explain.Judge main idea of the present invention, should consider appended claim.
Claims (10)
1. an adaptive block effect removal method, comprising:
To input video present frame again piecemeal, making each repartitioned two central shaft of new is just the horizontal direction of four original blocks and the original block border of vertical direction adjacent one another are;
Calculate the difference of both sides pixel on described original block border and the difference of the pixel of original block inside, both sides;
The difference of determining the described original block boundaries on either side pixel calculate whether be greater than described original block boundaries on either side original block inside pixel difference and and the difference of described original block boundaries on either side pixel whether surpass predefine threshold value;
If being the difference of pixel and the difference of described original block boundaries on either side pixel that are greater than described original block boundaries on either side original block inside, the difference of described original block boundaries on either side pixel do not surpass predefine threshold value, in the consecutive frame of the front and back of present frame, search the new piece of optimum Match, adopt the new piece of optimum Match in the consecutive frame find to upgrade described new of present frame.
2. adaptive block effect removal method as described in claim 1, described step of searching the new piece of optimum Match in the consecutive frame of the front and back of present frame comprises:
In the consecutive frame of the front and back of present frame, find the new piece with the consecutive frame of the new piece correspondence position of present frame, and to take the corresponding Xin Kuai center of consecutive frame be the center of circle, with the radius of being scheduled to, at front and back consecutive frame, the new piece of described correspondence is expanded, obtained and take the extended area that described corresponding Xin Kuai center is the center of circle;
Before and after calculating, in consecutive frame, form each the new of described extended area absolute value sum with the difference of the gray value of the respective pixel of the new piece of present frame; And
New piece in the calculated minimum corresponding extended area of absolute value sum is defined as to the new piece of optimum Match in the front and back consecutive frame of new piece of present frame.
3. the adaptive block effect removal method described in claim 1 or 2, also comprises: before the new piece of present frame is upgraded, select the region of the required renewal in the new piece of described present frame according to the content complexity of the new piece of described present frame.
4. adaptive block effect removal method as described in claim 3, wherein, the step in the region of the required renewal in the new piece of the described present frame of described selection also comprises:
According to the content complexity of the new piece of present frame, calculate the required step-length of expansion to the periphery of two central shafts by the new piece of described present frame;
The cross region that the step length of calculating to external expansion along described two central shaft crosspoints is formed is defined as region to be updated.
5. adaptive block effect removal method as described in claim 4, wherein, described method also comprises: before in adopting the consecutive frame finding, the new piece of optimum Match upgrades the step of described new of present frame, the pixel value that recalculates each pixel to be updated in described region to be updated on spatial domain, comprising:
For described each pixel to be updated, in described region to be updated, obtain the pixel set of the described pixel to be updated of impact;
Pixel set based on obtained and the distance property of described pixel to be updated and grey value difference attribute recalculate the weighted pixel value of described pixel to be updated.
6. the adaptive block effect removal method described in claim 1 or 5, the step of described new that in the consecutive frame that described employing is found, the new piece of optimum Match upgrades present frame is undertaken by the pixel value three of the respective pixel of the new piece of optimum Match of the new piece of present frame and determined front and back consecutive frame being weighted to summation.
7. adaptive block effect is removed a system, comprising:
Again piecemeal module, for the present frame of input video is carried out to piecemeal again, making each repartitioned two central shaft of new is just the horizontal direction of four original blocks and the original block border of vertical direction adjacent one another are;
Computing module, for calculating the difference of both sides pixel on described original block border and the difference of the pixel of original block inside, both sides;
Determination module, for determine the difference of the described original block boundaries on either side pixel calculate whether be greater than described original block boundaries on either side original block inside pixel difference and and the difference of described original block boundaries on either side pixel whether surpass predefine threshold value;
Search module, if the difference for described original block boundaries on either side pixel is that the difference of pixel and the difference of described original block boundaries on either side pixel that is greater than described original block boundaries on either side original block inside do not surpass predefine threshold value, in the consecutive frame of the front and back of present frame, search the new piece of optimum Match; And
Update module, adopts the new piece of optimum Match in the consecutive frame find to upgrade described new of present frame.
8. adaptive block effect is removed system as described in claim 7, described in search module and comprise:
Extended area obtains module, in the consecutive frame of the front and back of present frame, find the new piece with the consecutive frame of the new piece correspondence position of present frame, and to take the corresponding Xin Kuai center of consecutive frame be the center of circle, with the radius of being scheduled to, at front and back consecutive frame, the new piece of described correspondence is expanded, obtained and take the extended area that described corresponding Xin Kuai center is the center of circle;
Gray value differences absolute value and computing module, form each the new of described extended area absolute value sum with the difference of the gray value of the respective pixel of the new piece of present frame in consecutive frame before and after calculating; And
The new piece determination module of optimum Match, is defined as the new piece in the calculated minimum corresponding extended area of absolute value sum the new piece of optimum Match in the front and back consecutive frame of new piece of present frame.
9. the adaptive block effect described in claim 7 or 8 is removed system, also comprises that more new region is selected module, selects the region of the required renewal in the new piece of described present frame according to the content complexity of the new piece of described present frame.
10. adaptive block effect is removed system as described in claim 9, it also comprises that pixel value recalculates module, its new piece of optimum Match in adopting the consecutive frame find upgrade described new of present frame before, on spatial domain, recalculate the pixel value of each pixel to be updated in region to be updated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010582772.2A CN102547256B (en) | 2010-12-10 | 2010-12-10 | Method and system for removing self-adaptive block effect |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010582772.2A CN102547256B (en) | 2010-12-10 | 2010-12-10 | Method and system for removing self-adaptive block effect |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102547256A CN102547256A (en) | 2012-07-04 |
| CN102547256B true CN102547256B (en) | 2014-12-03 |
Family
ID=46353063
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010582772.2A Expired - Fee Related CN102547256B (en) | 2010-12-10 | 2010-12-10 | Method and system for removing self-adaptive block effect |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102547256B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104505047B (en) * | 2014-12-31 | 2017-04-12 | 深圳市华星光电技术有限公司 | Display driving method, circuit and liquid crystal display |
| CN106228514B (en) * | 2016-07-11 | 2019-02-15 | 北京大学 | A method of eliminating compression video noise |
| CN106204493B (en) * | 2016-07-13 | 2019-01-15 | 北京大学 | A method of eliminating video blocking artifact |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0666695A2 (en) * | 1994-02-04 | 1995-08-09 | AT&T Corp. | Adaptive spatial-temporal postprocessing for low bit-rate coded image sequences |
| WO1998056184A1 (en) * | 1997-06-05 | 1998-12-10 | Wisconsin Alumni Research Foundation | Image compression system using block transforms and tree-type coefficient truncation |
| CN1972421A (en) * | 2005-11-25 | 2007-05-30 | 三星电子株式会社 | Frame interpolator, frame interpolation method and motion credibility evaluator |
| CN101494787A (en) * | 2009-02-10 | 2009-07-29 | 重庆大学 | De-block effect method based on block effect detection |
-
2010
- 2010-12-10 CN CN201010582772.2A patent/CN102547256B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0666695A2 (en) * | 1994-02-04 | 1995-08-09 | AT&T Corp. | Adaptive spatial-temporal postprocessing for low bit-rate coded image sequences |
| WO1998056184A1 (en) * | 1997-06-05 | 1998-12-10 | Wisconsin Alumni Research Foundation | Image compression system using block transforms and tree-type coefficient truncation |
| CN1972421A (en) * | 2005-11-25 | 2007-05-30 | 三星电子株式会社 | Frame interpolator, frame interpolation method and motion credibility evaluator |
| CN101494787A (en) * | 2009-02-10 | 2009-07-29 | 重庆大学 | De-block effect method based on block effect detection |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102547256A (en) | 2012-07-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8019124B2 (en) | Robust camera pan vector estimation using iterative center of mass | |
| KR100814424B1 (en) | Device for detecting occlusion area and method thereof | |
| US8588306B2 (en) | Temporal motion vector filtering | |
| US8237868B2 (en) | Systems and methods for adaptive spatio-temporal filtering for image and video upscaling, denoising and sharpening | |
| JP5520286B2 (en) | Filter processing method and apparatus for anti-aliasing | |
| US9277129B2 (en) | Robust image feature based video stabilization and smoothing | |
| WO2005098753A1 (en) | Image processing device and method, recording medium, and program | |
| WO2005098754A1 (en) | Image processing device and method, recording medium, and program | |
| CN101489121A (en) | Background model initializing and updating method based on video monitoring | |
| CN102547256B (en) | Method and system for removing self-adaptive block effect | |
| CN111046862A (en) | Character segmentation method and device and computer readable storage medium | |
| JP2000306108A (en) | Optical flow estimation method | |
| CN112215773A (en) | Local motion deblurring method and device based on visual saliency and storage medium | |
| KR101834512B1 (en) | Super-resolution image restoration apparatus and method based on consecutive-frame | |
| CN115546027B (en) | Image suture line determination method, device and storage medium | |
| KR100616755B1 (en) | Image processing device, image processing method, and storage medium | |
| CN116188521A (en) | Image processing method and device based on optical flow estimation and computer equipment | |
| KR102369802B1 (en) | Image processing apparatus and image processing method | |
| KR20170082331A (en) | Method and apparatus for predicting eye position | |
| Fisher | 3d plane labeling stereo matching with content aware adaptive windows | |
| KR100456102B1 (en) | Apparatus and Method of Motion Detection for Temporal Filtering | |
| JP4453202B2 (en) | Image processing apparatus, image processing method, and computer-readable recording medium | |
| EP1840823B1 (en) | Method for discriminating textures regions and homogeneous or flat regions in an image | |
| CN107154052A (en) | The method and device of Obj State estimation | |
| CN114666583B (en) | Video coding preprocessing method based on time-space domain filtering |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20141203 Termination date: 20171210 |