CN101699865A - Method for fast searching mass movement self-adapting sub pixel - Google Patents
Method for fast searching mass movement self-adapting sub pixel Download PDFInfo
- Publication number
- CN101699865A CN101699865A CN 200910229608 CN200910229608A CN101699865A CN 101699865 A CN101699865 A CN 101699865A CN 200910229608 CN200910229608 CN 200910229608 CN 200910229608 A CN200910229608 A CN 200910229608A CN 101699865 A CN101699865 A CN 101699865A
- Authority
- CN
- China
- Prior art keywords
- search
- pixel
- point
- optimal
- motion
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 229910003460 diamond Inorganic materials 0.000 claims abstract description 24
- 239000010432 diamond Substances 0.000 claims abstract description 24
- 239000013598 vector Substances 0.000 claims abstract description 23
- 230000003044 adaptive effect Effects 0.000 claims description 7
- PIDANAQULIKBQS-UHFFFAOYSA-N 17-hydroxy-17-(2-hydroxyacetyl)-10,13,16-trimethyl-6,7,8,9,12,14,15,16-octahydrocyclopenta[a]phenanthrene-3,11-dione Chemical compound C1CC2=CC(=O)C=CC2(C)C2C1C1CC(C)C(C(=O)CO)(O)C1(C)CC2=O PIDANAQULIKBQS-UHFFFAOYSA-N 0.000 claims description 4
- 230000000977 initiatory effect Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 description 6
- 230000006835 compression Effects 0.000 description 5
- 238000007906 compression Methods 0.000 description 5
- 230000007547 defect Effects 0.000 description 3
- 230000002123 temporal effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention provides a method for fast searching a mass movement self-adapting sub pixel, belonging to the technical field of motion estimation in video coding. The method comprises the steps of: 1) starting; 2) calculating motion vector; 3) searching integer pixel; 4) starting sub pixel search; 5) interpolating 1/2 and 1/4 pixel; 6) judging the motion violent degree of the current macro block; 7) carry out major diamond search; 8) carrying out first-step small diamond search; 9) judging whether an optimum point is at the center of the small diamond; 10) carrying out a second-step small diamond search; 11) stopping sub pixel search; and 12) returning the MV and the motion cost of the position of the optimum point of the sub pixel search, and stopping the motion estimation of the macro block. The method can lead AVS, H.264 to be used for more hardware platforms, obviously reduces the computation complexity of a coder, shortens coding time, and improves the performance.
Description
Technical Field
The invention belongs to the technical field of motion estimation in video coding, and particularly relates to a sub-pixel fast search method adaptive to block motion degree.
Background
The original motion video sequence contains a large amount of data information that is difficult to store and transmit without compression encoding. For video sequences, the temporal redundancy between adjacent frames is high, and motion estimation based on block matching is an effective method for reducing the redundancy between frames of the video sequences. Simple and effective block matching type motion estimation algorithms are widely applied in modern video compression coding standards. International standard h.261 employs integer-pixel motion estimation. However, the displacement of the actual object is rarely an integer pixel, and if the integer pixel is used as a unit for motion estimation, a large error will be generated; using 1/2 pixels for the description will reduce a significant part of the motion compensation error. So that international standards such as h.263, MPEG-1, MPEG-2, etc. successively adopt motion estimation with half-pixel accuracy. Meanwhile, because the accurate motion description can obtain smaller prediction error and better visual effect, 1/4-pixel-precision motion estimation technology is introduced into H.264 and AVS to improve the precision of motion compensation, further reduce the prediction error and improve the compression rate.
Motion estimation is one of the core techniques of video coding compression, which effectively reduces temporal redundancy between successive video frames. But at the same time it is also the most computationally intensive and time consuming part of video coding. Experiments prove that the motion estimation accounts for about 80% of the whole coding time, so that the improvement and optimization of the motion estimation method becomes the key for improving the coding speed.
For integer pixel motion estimation, a number of fast search algorithms have been proposed, such as three-step search method (TSS), two-dimensional logarithmic search method (TDL), four-step search method (FSS), diamond search method (DS), etc. Compared with full search, the algorithms greatly reduce the average search point number of the integer pixel and improve the search speed of the integer pixel. In addition, the asymmetric cross-shaped multi-level hexagonal lattice point Search (UMHexagonS, unmetric-cross multi-Hexagon-grid Search) adopts a plurality of motion vector predictions and Search modes with different shapes, can reduce the integer pixel motion Search complexity by 90%, and the reduction of the average peak signal-to-noise ratio (PSNR) is less than 0.05db, is the best fast integer pixel motion estimation algorithm at present, and is adopted by the h.264 and AVS standards.
With the great amount of proposed fast integer pixel motion estimation algorithms, the speed of integer pixel motion estimation is improved significantly, which also results in that the calculation amount of fractional pixel motion estimation accounts for more and more of the whole motion estimation calculation amount, and therefore how to reduce the calculation amount of fractional pixel motion estimation becomes another way to improve the video coding speed.
The currently widely used sub-pixel search methods include full search and diamond search. In the full search, firstly, the optimal point obtained by integer pixel search is used as the center starting point of the search, eight 1/2 pixel points around the optimal point are fully searched, difference and matching are carried out, and motion vectors with optimal 1/2 pixel points and 1/2 pixel accuracy are obtained; then, the optimal 1/2 pixel point is used as a center starting point of searching, eight 1/4 pixel points around the optimal 1/2 pixel point are searched, and the optimal 1/4 pixel point and the motion vector are obtained. The full search can obtain the most accurate 1/4 pixel points, but the calculation complexity is high, 16 pixel points need to be searched each time, and the searching speed is low. The diamond search method improves the defect of more search points in the full search, and is a quick search method. The method takes an optimal point obtained by integer pixel search as a center, takes a small prism with the step length of 1 as a template to search 4 1/4 pixel points around the optimal point, if the optimal point is still at the center, the search is ended, otherwise, the small prism search is repeated by taking a new optimal point as the center until the optimal point is at the center or the boundary of a sub-pixel is searched. Therefore, the number of pixel points searched each time is not fixed, at least 4 sub-pixel points are searched, at most 28 sub-pixel points need to be searched when the boundary of the sub-pixel is searched, statistics shows that the probability that the optimal sub-pixel point is close to the optimal integer pixel point is up to 90 percent; however, as the motion of the video sequence is more complex and violent, the optimal sub-pixel points are diffused to the periphery, and the probability of being close to the optimal integer pixel points is only about 30%. Therefore, for a video sequence with slow motion, diamond search can remarkably improve the search rate, and the optimal matching point can be found only by searching 4 or 8 sub-pixel points. However, for a video sequence with complex and violent motion, the average search point number of diamond search is obviously increased and is easy to fall into local optimum, so that the search speed and the image quality are deteriorated. Journal of the university of qinghua (natural science edition), 2007, Vo l.47, No. 1: 25-27, 31, the article entitled "fractional pixel exact motion search fast algorithm" written by manting and royal xi micro.
In conclusion, a fast sub-pixel search method is designed, so that better search speed and better search results can be obtained for video sequences with different motion degrees, and the method becomes a key for improving sub-pixel search and motion estimation.
Disclosure of Invention
In order to overcome the defects and shortcomings of the existing sub-pixel searching method, the invention provides a block motion adaptive sub-pixel fast searching method.
The technical scheme of the invention is realized by adopting the following mode.
A block motion adaptive sub-pixel fast search method comprises the following steps:
1) start of
2) Motion vector prediction
Before the motion search is carried out on the current block, firstly, the motion vector prediction is carried out to obtain a prediction motion vector pred _ MV and a motion compensation cost value pred _ mcost of a point pointed by the prediction motion vector;
3) integer pixel search
Taking the optimal prediction starting point obtained by motion vector prediction as the starting point of integer pixel search, and performing integer pixel search on the current block to obtain the position of the optimal point of the integer pixel and the motion compensation cost mincost;
4) initiating a sub-pixel search
Taking the optimal point of the integer pixel as a starting point of sub-pixel search;
5)1/2 and 1/4 pixel interpolation
Carrying out interpolation to obtain 1/2 pixel points and 1/4 pixel points;
6) judging the motion intensity of the current macro block (block)
Judging the intensity of the current block motion by adopting a method of solving difference values between motion compensation cost values, namely comparing the mincost and the pred _ mcost, if the mincost-pred _ mcost is less than beta pred _ mcost, indicating that the cost result of integer pixel search is close to a predicted value, and the macro block (block) moves slowly, and turning to step 8); otherwise, the difference between the integer pixel search result and the predicted value is large, the motion of the macro block (block) is violent, and the step 7) is carried out;
7) performing large diamond searches
The motion of the current block is violent, the distribution of the optimal sub-pixel points is dispersed, and the large prism with large search step length is adopted to search and position 1/2 pixel precision, so that the current block is prevented from falling into local optimal: in the step, an optimal integer pixel point is used as a search center, a large diamond with the step length of 2 is used as a template, 4 1/2 pixel points in the horizontal and vertical directions adjacent to the optimal integer pixel point are searched, cost values mcost of block matching of the 4 1/2 pixel points are calculated, the mcost and mincost of the optimal point of the central integer pixel are compared, if the mcost is smaller than the mincost, a 1/2 pixel point with the minimum mcost is used as a starting point of next search, and the mcost is used as a new mincost; otherwise, taking the optimal integer pixel point as the initial point of the next search, keeping mincost unchanged, and turning to the next step;
8) performing a first level of small diamond search
The motion of the current block is slow, or the current block is positioned by a large prismatic search, and the 1/4 pixel precision is refined by a small prismatic search with small step size: in the step, an integer pixel optimal point or a new optimal point obtained in the step 7) is used as a starting point of small prism search, a small prism with the step length of 1 is used as a template, 4 1/4 pixel points around the starting point are searched, and a cost value mcost of block matching is calculated to obtain a point with the minimum mcost;
9) judging whether the optimal point is at the center of the small diamond
After the first-stage small prismatic search, if the cost value of the central point is still the minimum, the point is the optimal point for sub-pixel search, and the step 11 is carried out); otherwise, taking the 1/4 pixel point with the minimum mcost as a new search center, and entering the step 10);
10) performing a second level of small diamond search
Taking the optimal point obtained by the first-stage small prismatic search as the center, taking the small prismatic with the step length of 1 as a template, performing the small prismatic search again, taking the point (including the center) with the minimum mcost in the search as the optimal point for the sub-pixel search, and turning to the next step;
11) end of sub-pixel search
Taking the central point of the first-stage small prismatic search (if the center is optimal in the first-stage small prismatic search) or the optimal point obtained by the second-stage small prismatic search as the optimal point of the sub-pixel search, and ending the sub-pixel search;
12) and returning the MV and the motion cost of the optimal point position of the sub-pixel search, and finishing the motion estimation of the macro block (block).
The motion vector prediction in the step 2) can adopt various prediction modes such as median prediction, origin prediction, uplayer prediction, prediction of a block corresponding to a previous frame, prediction of an adjacent reference frame and the like. Like motion vectors, the corresponding motion compensation cost mcost has a strong correlation. Therefore, the motion compensation cost values of the points pointed by the predicted motion vectors are respectively used as pred _ mcost as whether the mincost meets the accurate scale after the integer pixel search, and the satisfaction degree of the integer pixel search and the motion degree of the macro block (block) are judged by using the threshold judgment formula in the step 6).
The integer pixel search in step 3) may use the existing mature UMhexagonS algorithm.
And 6) judging the motion degree of the macro block (block) by comparing the mincost-pred _ mcost and the beta pred _ mcost to obtain a difference value. Wherein,
bsize is the size of the block, blocktype is the type of block, and α is a constant.
According to the judgment of the motion type of the current block in the step 6), if the current block moves slowly, only 1 or 2-level small prismatic search is carried out, and the optimal sub-pixel point can be quickly found by only searching 4 or 8 points; if the current block moves violently, the 1-level large prismatic search positioning is carried out firstly to prevent the current block from falling into local optimum, and then the 1-level or 2-level small prismatic search is carried out by taking the optimum point of the large prismatic search as the center, so that the search precision can be improved, and the number of search points can be effectively reduced.
A large number of experimental statistics show that the slower the macro block moves, the more the optimal sub-pixel point converges to the vicinity of the optimal integer pixel point, otherwise, the more divergent the macro block moves; meanwhile, the probability that the sub-pixel points in the horizontal and vertical directions are optimal is far higher than that of other sub-pixel points. The method of the invention makes full use of the relation between the intensity degree of macro block motion and the optimal sub-pixel point distribution characteristic, combines the motion vector prediction and the motion compensation cost prediction, adopts different search templates in a self-adaptive way according to the motion degree of the current block, makes up the defects of the original method, and can find the optimal point with quarter pixel precision for macro blocks with different motion degrees by the least number of search points. The method can be applied to video coding standards such as AVS, H.264 and the like to improve the performance of the encoder.
The method is suitable for the field of AVS, H.264 or other video coding adopting 1/4 pixel precision motion estimation, is a new sub-pixel searching method adaptive to the motion degree of macro blocks (blocks), can quickly find the optimal sub-pixel point with the least searching point number for the blocks with different motion degrees, and effectively reduces the time of video coding and the complexity of calculation under the condition of ensuring that the video compression quality is basically unchanged. The experimental result shows that the method has good search rate and search precision for video sequences with different motion degrees. The method is applied to an AVS reference model to code various video sequences, such as foreman _ qcif.yuv, akiyo _ qcif.yuv, contiiner _ qcif.yuv, mobile _ qcif.yuv and bus _ cif.yuv, and the coding time, peak signal-to-noise ratio (PSNR) and bit rate of the video sequences are respectively counted, and the result shows that the coding time is generally shortened by 20-30 percent compared with the original coding time, the peak signal-to-noise ratio (PSNR) is only reduced by 0.01-0.03db, and the bit rate is basically kept unchanged. The invention can apply AVS and H.264 to a plurality of hardware platforms, obviously reduce the calculation complexity of the encoder, shorten the encoding time and improve the performance.
Drawings
FIG. 1 is a block flow diagram of the method of the present invention. Wherein 1-12 are each steps in the method.
Fig. 2 shows the distribution of 1/2 pixel points and 1/4 pixel points obtained by interpolation around the optimal integer pixel point.
Fig. 3a shows that the current block is judged to be a block with violent motion, a large prismatic searching template is firstly adopted, and black in the graph is a pixel point to be searched. Fig. 3b and 3c show the case of 1-level and 2-level small prismatic searches performed after large prismatic searches, with the obtained optimal 1/2 pixel as the center (assuming that the top 1/2 pixel is optimal).
FIG. 4a shows a direct level 1 small prismatic search, in which the current block is determined to be a slow moving block. FIG. 4b shows a level 2 mini-prismatic search performed after a level 1 mini-prismatic search with the optimal point not at the center (assumed to be at the right).
Detailed description of the invention
The present invention is further described below with reference to the following drawings and examples, but is not limited thereto.
Example (b):
the embodiment of the invention is shown in FIG. 1, and discloses a block motion adaptive sub-pixel fast search method, which comprises the following steps:
1) start of
2) Motion vector prediction
Before the motion search is carried out on the current block, firstly, the motion vector prediction is carried out to obtain a prediction motion vector pred _ MV and a motion compensation cost value pred _ mcost of a point pointed by the prediction motion vector;
3) integer pixel search
Taking the optimal prediction starting point obtained by motion vector prediction as the starting point of integer pixel search, and performing integer pixel search on the current block to obtain the position of the optimal point of the integer pixel and the motion compensation cost mincost;
4) initiating a sub-pixel search
Taking the optimal point of the integer pixel as a starting point of sub-pixel search;
5)1/2 and 1/4 pixel interpolation
Carrying out interpolation to obtain 1/2 pixel points and 1/4 pixel points;
6) judging the motion intensity of the current macro block (block)
Judging the intensity of the current block motion by adopting a method of solving difference values between motion compensation cost values, namely comparing the mincost and the pred _ mcost, if the mincost-pred _ mcost is less than beta pred _ mcost, indicating that the cost result of integer pixel search is close to a predicted value, and the macro block (block) moves slowly, and turning to step 8); otherwise, the difference between the integer pixel search result and the predicted value is large, the motion of the macro block (block) is violent, and the step 7) is carried out;
7) performing large diamond searches
The motion of the current block is violent, the distribution of the optimal sub-pixel points is dispersed, and the large prism with large search step length is adopted to search and position 1/2 pixel precision, so that the current block is prevented from falling into local optimal: in the step, an optimal integer pixel point is used as a search center, a large diamond with the step length of 2 is used as a template, 4 1/2 pixel points in the horizontal and vertical directions adjacent to the optimal integer pixel point are searched, cost values mcost of block matching of the 4 1/2 pixel points are calculated, the mcost and mincost of the optimal point of the central integer pixel are compared, if the mcost is smaller than the mincost, a 1/2 pixel point with the minimum mcost is used as a starting point of next search, and the mcost is used as a new mincost; otherwise, taking the optimal integer pixel point as the initial point of the next search, keeping mincost unchanged, and turning to the next step;
8) performing a first level of small diamond search
The motion of the current block is slow, or the current block is positioned by a large prismatic search, and the 1/4 pixel precision is refined by a small prismatic search with small step size: in the step, an integer pixel optimal point or a new optimal point obtained in the step 7) is used as a starting point of small prism search, a small prism with the step length of 1 is used as a template, 4 1/4 pixel points around the starting point are searched, and a cost value mcost of block matching is calculated to obtain a point with the minimum mcost;
9) judging whether the optimal point is at the center of the small diamond
After the first-stage small prismatic search, if the cost value of the central point is still the minimum, the point is the optimal point for sub-pixel search, and the step 11 is carried out); otherwise, taking the 1/4 pixel point with the minimum mcost as a new search center, and entering the step 10);
10) performing a second level of small diamond search
Taking the optimal point obtained by the first-stage small prismatic search as the center, taking the small prismatic with the step length of 1 as a template, performing the small prismatic search again, taking the point (including the center) with the minimum mcost in the search as the optimal point for the sub-pixel search, and turning to the next step;
11) end of sub-pixel search
Taking the central point of the first-stage small prismatic search (if the center is optimal in the first-stage small prismatic search) or the optimal point obtained by the second-stage small prismatic search as the optimal point of the sub-pixel search, and ending the sub-pixel search;
12) and returning the MV and the motion cost of the optimal point position of the sub-pixel search, and finishing the motion estimation of the macro block (block).
Claims (1)
1. A block motion adaptive sub-pixel fast search method comprises the following steps:
1) start of
2) Motion vector prediction
Before the motion search is carried out on the current block, firstly, the motion vector prediction is carried out to obtain a prediction motion vector pred _ MV and a motion compensation cost value pred _ mcost of a point pointed by the prediction motion vector;
3) integer pixel search
Taking the optimal prediction starting point obtained by motion vector prediction as the starting point of integer pixel search, and performing integer pixel search on the current block to obtain the position of the optimal point of the integer pixel and the motion compensation cost mincost;
4) initiating a sub-pixel search
Taking the optimal point of the integer pixel as a starting point of sub-pixel search;
5)1/2 and 1/4 pixel interpolation
Carrying out interpolation to obtain 1/2 pixel points and 1/4 pixel points;
6) judging the intensity of the motion of the current macro block
Judging the intensity of the current block motion by adopting a method of solving difference values between motion compensation cost values, namely comparing the mincost and the pred _ mcost, if the mincost-pred _ mcost is less than beta pred _ mcost, indicating that the cost result of integer pixel search is close to a predicted value, the macro block moves slowly, and turning to step 8); otherwise, the difference between the integer pixel search result and the predicted value is large, the motion of the macro block is violent, and the step 7) is carried out;
7) performing large diamond searches
The motion of the current block is violent, the distribution of the optimal sub-pixel points is dispersed, and the large prism with large search step length is adopted to search and position 1/2 pixel precision, so that the current block is prevented from falling into local optimal: in the step, an optimal integer pixel point is used as a search center, a large diamond with the step length of 2 is used as a template, 4 1/2 pixel points in the horizontal and vertical directions adjacent to the optimal integer pixel point are searched, cost values mcost of block matching of the 4 1/2 pixel points are calculated, the mcost and mincost of the optimal point of the central integer pixel are compared, if the mcost is smaller than the mincost, a 1/2 pixel point with the minimum mcost is used as a starting point of next search, and the mcost is used as a new mincost; otherwise, taking the optimal integer pixel point as the initial point of the next search, keeping mincost unchanged, and turning to the next step;
8) performing a first level of small diamond search
The slow motion of the current block or the positioning which has undergone the large prismatic search adopts the small prismatic search with small step size to refine the accuracy of 1/4 pixels: in the step, an integer pixel optimal point or a new optimal point obtained in the step 7) is used as a starting point of small prism search, a small prism with the step length of 1 is used as a template, 4 1/4 pixel points around the starting point are searched, and a cost value mcost of block matching is calculated to obtain a point with the minimum mcost;
9) judging whether the optimal point is at the center of the small diamond
After the first-stage small prismatic search, if the cost value of the central point is still the minimum, the point is the optimal point for sub-pixel search, and the step 11 is carried out); otherwise, taking the 1/4 pixel point with the minimum mcost as a new search center, and entering the step 10);
10) performing a second level of small diamond search
Taking the optimal point obtained by the first-stage small prismatic search as the center, taking the small prismatic with the step length of 1 as a template, performing the small prismatic search again, taking the point with the minimum mcost in the search as the optimal point for the sub-pixel search, and turning to the next step;
11) end of sub-pixel search
Taking the central point of the first-stage small prismatic search or the optimal point obtained by the second-stage small prismatic search as the optimal point of the sub-pixel search, and finishing the sub-pixel search;
12) and returning the MV and the motion cost of the optimal point position of the sub-pixel search, and finishing the macro block motion estimation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910229608 CN101699865B (en) | 2009-10-28 | 2009-10-28 | Method for fast searching mass movement self-adapting sub pixel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910229608 CN101699865B (en) | 2009-10-28 | 2009-10-28 | Method for fast searching mass movement self-adapting sub pixel |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101699865A true CN101699865A (en) | 2010-04-28 |
| CN101699865B CN101699865B (en) | 2011-06-01 |
Family
ID=42148305
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200910229608 Expired - Fee Related CN101699865B (en) | 2009-10-28 | 2009-10-28 | Method for fast searching mass movement self-adapting sub pixel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101699865B (en) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101867815A (en) * | 2010-04-30 | 2010-10-20 | 西北工业大学 | Rapid fractional pixel hierarchical searching method |
| CN102123284A (en) * | 2011-03-18 | 2011-07-13 | 河北师范大学 | Quick search method in H.264 coding motion estimation |
| CN102164283A (en) * | 2011-05-30 | 2011-08-24 | 江苏大学 | Sub-pixel motion estimation method based on AVS (audio video standard) |
| CN102340660A (en) * | 2010-07-23 | 2012-02-01 | 联合信源数字音视频技术（北京）有限公司 | Motion vector prediction apparatus based on AVS encoder |
| CN102377995A (en) * | 2010-08-05 | 2012-03-14 | 富士通株式会社 | Fractional pixel motion estimation method and system thereof |
| CN103281533A (en) * | 2013-05-14 | 2013-09-04 | 芯原微电子(北京)有限公司 | Device and method for motion estimation in enhancement layer in scalable video coding |
| CN103384332A (en) * | 2013-06-29 | 2013-11-06 | 张新安 | Improved algorithm for HFPS algorithm of AVS video coding |
| CN103402084A (en) * | 2013-07-22 | 2013-11-20 | 张新安 | AVS (Audio Video coding Standard)-based motion estimation fast algorithm |
| CN103873875A (en) * | 2014-03-25 | 2014-06-18 | 山东大学 | Layering sub pixel motion estimation method for image super resolution |
| CN104215268A (en) * | 2014-07-07 | 2014-12-17 | 中国电子科技集团公司第四十一研究所 | Sextuple hardware interpolation method for laser rotary coded signals |
| CN104967858A (en) * | 2015-06-30 | 2015-10-07 | 北京奇艺世纪科技有限公司 | Sub-pixel motion vector estimation method and device |
| CN105611299A (en) * | 2015-12-25 | 2016-05-25 | 北京工业大学 | Motion estimation method based on HEVC |
| CN106331703A (en) * | 2015-07-03 | 2017-01-11 | 华为技术有限公司 | Video coding and decoding method, and video coding and decoding device |
| CN109348234A (en) * | 2018-11-12 | 2019-02-15 | 北京佳讯飞鸿电气股份有限公司 | A kind of efficient sub-picture element movement estimating method and system |
| CN109561315A (en) * | 2018-10-31 | 2019-04-02 | 北京金山云网络技术有限公司 | A kind of method for estimating, device, electronic equipment and storage medium |
| WO2019072248A1 (en) * | 2017-10-12 | 2019-04-18 | 北京金山云网络技术有限公司 | Motion estimation method and device, electronic apparatus and computer readable storage medium |
| CN109660799A (en) * | 2017-10-12 | 2019-04-19 | 北京金山云网络技术有限公司 | Method for estimating, device, electronic equipment and storage medium in Video coding |
| CN111837386A (en) * | 2018-09-21 | 2020-10-27 | Kddi 株式会社 | Image decoding device, image encoding device, image processing system, and program |
| CN112866699A (en) * | 2019-03-11 | 2021-05-28 | 杭州海康威视数字技术股份有限公司 | Encoding and decoding method, device and equipment |
| CN112866719A (en) * | 2019-11-27 | 2021-05-28 | 北京博雅慧视智能技术研究院有限公司 | Rapid pixel-division prediction method for AVS2 |
| CN113963433A (en) * | 2021-09-15 | 2022-01-21 | 北京百度网讯科技有限公司 | Motion search method and device, electronic equipment and storage medium |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1156168C (en) * | 2001-01-12 | 2004-06-30 | 北京航空航天大学 | Quick video motion estimating method |
| CN1245028C (en) * | 2003-03-03 | 2006-03-08 | 清华大学 | Non-uniform multilayer hexaploid lattice full pixel kinematic search method |
| CN1236624C (en) * | 2003-09-04 | 2006-01-11 | 上海大学 | Quick full pixel movement evaluation method of multiple kinds of modules |
| CN1237818C (en) * | 2003-11-18 | 2006-01-18 | 清华大学 | A coding method based on selfadaptive hexagon search |
-
2009
- 2009-10-28 CN CN 200910229608 patent/CN101699865B/en not_active Expired - Fee Related
Cited By (35)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101867815A (en) * | 2010-04-30 | 2010-10-20 | 西北工业大学 | Rapid fractional pixel hierarchical searching method |
| CN101867815B (en) * | 2010-04-30 | 2011-09-14 | 西北工业大学 | Rapid fractional pixel hierarchical searching method |
| CN102340660A (en) * | 2010-07-23 | 2012-02-01 | 联合信源数字音视频技术（北京）有限公司 | Motion vector prediction apparatus based on AVS encoder |
| CN102340660B (en) * | 2010-07-23 | 2013-10-30 | 联合信源数字音视频技术（北京）有限公司 | Motion vector prediction apparatus based on AVS encoder |
| CN102377995B (en) * | 2010-08-05 | 2015-03-18 | 富士通株式会社 | Fractional pixel motion estimation method and system thereof |
| CN102377995A (en) * | 2010-08-05 | 2012-03-14 | 富士通株式会社 | Fractional pixel motion estimation method and system thereof |
| CN102123284A (en) * | 2011-03-18 | 2011-07-13 | 河北师范大学 | Quick search method in H.264 coding motion estimation |
| CN102164283A (en) * | 2011-05-30 | 2011-08-24 | 江苏大学 | Sub-pixel motion estimation method based on AVS (audio video standard) |
| CN103281533A (en) * | 2013-05-14 | 2013-09-04 | 芯原微电子(北京)有限公司 | Device and method for motion estimation in enhancement layer in scalable video coding |
| CN103281533B (en) * | 2013-05-14 | 2016-02-24 | 芯原微电子(北京)有限公司 | For equipment and the method for enhancement layer estimation in scalable video |
| CN103384332A (en) * | 2013-06-29 | 2013-11-06 | 张新安 | Improved algorithm for HFPS algorithm of AVS video coding |
| CN103402084A (en) * | 2013-07-22 | 2013-11-20 | 张新安 | AVS (Audio Video coding Standard)-based motion estimation fast algorithm |
| CN103873875A (en) * | 2014-03-25 | 2014-06-18 | 山东大学 | Layering sub pixel motion estimation method for image super resolution |
| CN104215268A (en) * | 2014-07-07 | 2014-12-17 | 中国电子科技集团公司第四十一研究所 | Sextuple hardware interpolation method for laser rotary coded signals |
| CN104967858A (en) * | 2015-06-30 | 2015-10-07 | 北京奇艺世纪科技有限公司 | Sub-pixel motion vector estimation method and device |
| CN104967858B (en) * | 2015-06-30 | 2018-01-26 | 北京奇艺世纪科技有限公司 | A kind of sub-pixel motion is to amount estimation method and device |
| CN106331703A (en) * | 2015-07-03 | 2017-01-11 | 华为技术有限公司 | Video coding and decoding method, and video coding and decoding device |
| US10523965B2 (en) | 2015-07-03 | 2019-12-31 | Huawei Technologies Co., Ltd. | Video coding method, video decoding method, video coding apparatus, and video decoding apparatus |
| WO2017005146A1 (en) * | 2015-07-03 | 2017-01-12 | 华为技术有限公司 | Video encoding and decoding method and device |
| CN105611299A (en) * | 2015-12-25 | 2016-05-25 | 北京工业大学 | Motion estimation method based on HEVC |
| CN105611299B (en) * | 2015-12-25 | 2018-11-23 | 北京工业大学 | A kind of method for estimating based on HEVC |
| WO2019072248A1 (en) * | 2017-10-12 | 2019-04-18 | 北京金山云网络技术有限公司 | Motion estimation method and device, electronic apparatus and computer readable storage medium |
| CN109660799A (en) * | 2017-10-12 | 2019-04-19 | 北京金山云网络技术有限公司 | Method for estimating, device, electronic equipment and storage medium in Video coding |
| CN111837386A (en) * | 2018-09-21 | 2020-10-27 | Kddi 株式会社 | Image decoding device, image encoding device, image processing system, and program |
| CN109561315B (en) * | 2018-10-31 | 2020-10-02 | 北京金山云网络技术有限公司 | Motion estimation method and device, electronic equipment and storage medium |
| CN109561315A (en) * | 2018-10-31 | 2019-04-02 | 北京金山云网络技术有限公司 | A kind of method for estimating, device, electronic equipment and storage medium |
| CN109348234A (en) * | 2018-11-12 | 2019-02-15 | 北京佳讯飞鸿电气股份有限公司 | A kind of efficient sub-picture element movement estimating method and system |
| CN109348234B (en) * | 2018-11-12 | 2021-11-19 | 北京佳讯飞鸿电气股份有限公司 | Efficient sub-pixel motion estimation method and system |
| CN112866699A (en) * | 2019-03-11 | 2021-05-28 | 杭州海康威视数字技术股份有限公司 | Encoding and decoding method, device and equipment |
| CN112866699B (en) * | 2019-03-11 | 2022-11-01 | 杭州海康威视数字技术股份有限公司 | Encoding and decoding method, device and equipment |
| US11902563B2 (en) | 2019-03-11 | 2024-02-13 | Hangzhou Hikvision Digital Technology Co., Ltd. | Encoding and decoding method and device, encoder side apparatus and decoder side apparatus |
| CN112866719A (en) * | 2019-11-27 | 2021-05-28 | 北京博雅慧视智能技术研究院有限公司 | Rapid pixel-division prediction method for AVS2 |
| CN112866719B (en) * | 2019-11-27 | 2022-09-23 | 北京博雅慧视智能技术研究院有限公司 | Rapid pixel-division prediction method for AVS2 |
| CN113963433A (en) * | 2021-09-15 | 2022-01-21 | 北京百度网讯科技有限公司 | Motion search method and device, electronic equipment and storage medium |
| CN113963433B (en) * | 2021-09-15 | 2024-03-26 | 北京百度网讯科技有限公司 | Motion search method, motion search device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101699865B (en) | 2011-06-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101699865B (en) | Method for fast searching mass movement self-adapting sub pixel | |
| CN110087087B (en) | VVC inter-frame coding unit prediction mode early decision and block division early termination method | |
| CN101431675B (en) | Pixel motion estimating method and apparatus | |
| KR100275694B1 (en) | Hierarchical search block matching method by using multiple motion vector candidates | |
| CN102047665A (en) | Dynamic image encoding method and dynamic image decoding method | |
| CN101326550A (en) | Motion estimation using prediction guided decimated search | |
| CN108419082B (en) | Motion estimation method and device | |
| CN106604035B (en) | A method of the estimation for Video coding and compression | |
| CN102291581B (en) | Realizing method of self-adaptive motion estimation supporting frame field | |
| RU2573747C2 (en) | Video encoding method and apparatus, video decoding method and apparatus and programmes therefor | |
| CN103188496A (en) | Fast motion estimation video encoding method based on motion vector distribution forecast | |
| US20240031576A1 (en) | Method and apparatus for video predictive coding | |
| CN103327327B (en) | For the inter prediction encoding unit selection method of high-performance video coding HEVC | |
| CN100551071C (en) | Integer pixel quick mixing search method based on the center prediction | |
| CN102647598A (en) | H.264 inter-frame mode optimization method based on maximin MV (Music Video) difference value | |
| CN103957420B (en) | Comprehensive movement estimation modified algorithm of H.264 movement estimation code | |
| CN102164283A (en) | Sub-pixel motion estimation method based on AVS (audio video standard) | |
| CN105282557B (en) | A kind of H.264 rapid motion estimating method of predicted motion vector | |
| CN106878754B (en) | A kind of 3D video depth image method for choosing frame inner forecast mode | |
| CN101778281A (en) | Method for estimating H.264-based fast motion on basis of structural similarity | |
| Van et al. | Fast motion estimation for closed-loop HEVC transrating | |
| CN106658024A (en) | Fast video coding method | |
| CN113242427B (en) | Rapid method and device based on adaptive motion vector precision in VVC | |
| CN102143362B (en) | Video transcoding processing method and video transcoding processing device from MPEG2 format or H. 263 format to H. 264 format | |
| Luo et al. | A new algorithm on MPEG-2 target bit-number allocation at scene changes |
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 | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110601 Termination date: 20131028 |