CN107707916A - A kind of frame per second transfer algorithm based on scene change detecte - Google Patents
A kind of frame per second transfer algorithm based on scene change detecte Download PDFInfo
- Publication number
- CN107707916A CN107707916A CN201710944115.XA CN201710944115A CN107707916A CN 107707916 A CN107707916 A CN 107707916A CN 201710944115 A CN201710944115 A CN 201710944115A CN 107707916 A CN107707916 A CN 107707916A
- Authority
- CN
- China
- Prior art keywords
- frame
- carried out
- scene
- shear
- scene change
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/142—Detection of scene cut or scene change
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/587—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal sub-sampling or interpolation, e.g. decimation or subsequent interpolation of pictures in a video sequence
Abstract
The invention discloses a kind of frame per second transfer algorithm based on scene change detecte, step are as follows：(1) scene change detecte, is carried out to low frame-rate video sequence；(2) computing estimation, is carried out according to the result of scene shear；(3) adaptive motion compensated interpolation, is carried out after computing estimation, and judges whether moving scene occurs shear, forward interpolation processing is then carried out in the event of shear, two-way interpolation processing is carried out if shear not occurring；(4) result carried out after interpolation processing is high frame-rate video sequence.Processing speed of the present invention is fast, and image quality becomes apparent from.
Description
Technical field：
The invention belongs to technical field of video processing, more particularly to a kind of frame per second conversion based on scene change detecte is calculated
Method.
Background technology：
Video information is one of most important video source of the mankind, and frame per second conversion is group important in video format switch technology
Into part, while it is also a new study hotspot in current video form study on the transformation field.
Existing frame per second transfer algorithm mainly has two classes：(1) non-motion compensated interleave algorithm (2) motion compensation interleave is calculated
Method.Non-motion compensated algorithm is simple, and speed is fast, but non-motion compensated interpolation algorithm can produce motion when handling moving object
It is fuzzy, it is mainly used in the situation not high to video quality demands.With the development of science and technology, requirement of the people to video is increasingly
Height, movement compensating algorithm effectively improve the fuzzy of moving object so that video playback is more smooth.Therefore, motion compensation is inserted
Value-based algorithm also just becomes the new focus of research of people, and its application is also more extensive.
The content of the invention
Goal of the invention：The problem of existing for prior art, the present invention propose that a kind of processing speed is fast, and image quality becomes apparent from
The frame per second transfer algorithm based on scene change detecte.
Technical scheme：The present invention proposes a kind of frame per second transfer algorithm based on scene change detecte, comprises the following steps：
(1) scene change detecte, is carried out to low frame-rate video sequence；
(2) computing estimation, is carried out according to the result of scene shear；
(3) adaptive motion compensated interpolation, is carried out after computing estimation, and judges whether moving scene occurs shear, if
Generation shear then carries out forward interpolation processing, and two-way interpolation processing is carried out if shear not occurring；
(4) result carried out after interpolation processing is high frame-rate video sequence.
Further, comprising the following steps that for scene change detecte is carried out in the step (1)：First present frame is carried out non-
Even partition, make during segmentation central area proportion maximum, while the consecutive frame of present frame is also in compliance with identical partition principle
Same operation is carried out, seeks each piecemeal colour consistency vector CCV afterwards.
Further, estimation comprises the following steps that in the step (2)：By each piece in present frame, preceding
Match block is obtained according to matching criterior in one frame or a certain given region of a later frame, by match block and the relative position of current block
Calculate motion vector.
Further, motion compensation is carried out using adaptive motion compensated mode after estimation in the step (3)
Interpolation comprises the following steps that, wherein shown in interleave result equation below：
Wherein fn and fn+1 is two consecutive frames, and fn is to predict the frame come, and s represents a segmentation block, and MV is that the block is corresponding
Motion vector, λ is weight.
Further, judge whether moving scene occurs shear and judged by weight λ in the step (3), wherein weight λ's
Calculation formula is as follows：
Wherein CCVD is that the colour consistency vector of the macro block at two frame same positions make the difference and compares to obtain；If
CCVD between two segmentation blocks has exceeded the empirical value T initially set up, now adaptive weight λ=1 or 0, Ye Jiwei
Occurrence scene shear, then row interpolation is entered to present frame using forward frame or backward frame, otherwise adaptive weight λ=0.5, field
Shear occurs for scape, now needs to enter row interpolation using bidirectional frame.
Beneficial effect：It is of the invention compared with existing frame rate conversion, including advantages below and beneficial effect.
1. carrying out scene change detecte before estimation, more video sequence occurs for those scene shears,
The accuracy of estimation can greatly improve.
2. selecting different motion compensated interpolation modes according to the testing result of scene shear, it can so improve video
Image quality, make video more smooth.
Brief description of the drawings
Fig. 1 is the overall flow figure of the present invention.
Embodiment
With reference to specific embodiment, the present invention is furture elucidated, it should be understood that these embodiments are merely to illustrate the present invention
Rather than limitation the scope of the present invention, after the present invention has been read, various equivalences of the those skilled in the art to the present invention
The modification of form falls within the application appended claims limited range.
As shown in figure 1, it is the execution flow chart of the frame per second transfer algorithm of the invention based on scene change detecte.This method bag
Include following steps：
Scene change detecte：According to human visual system (Human Visual System, HVS) feature, often in order to prominent
Non-uniformly distributed load can be carried out by image by going out the importance of central area feature, then we can also be in order that the standard of change detecte
True property is higher, and each frame equally is carried out into non-uniformly distributed load, when carrying out piecemeal, make region shared by the macro block of central area occupy compared with
Large scale, each consecutive frame carry out same piecemeal also in compliance with identical segmentation criterion and operated, and then calculate the color of each macro block
Parity vector CCV (Color Coherence Vectors), and the colour consistency of the macro block at two frame same positions is sweared
Amount make the difference comparing, by obtained CCVD (Color Coherence Vectors Difference) and the experience threshold of setting
Value T relatively decides whether to produce shear.
Estimation：With the searching algorithm of Block- matching, on the basis of based on Block- matching, according to the result of scene shear
Simplify the selection of optimal vector in motion estimation process.It is a certain given in former frame or a later frame by each piece in present frame
Region in match block obtained according to matching criterior, motion vector is calculated by the relative position of match block and current block.
Motion compensated interpolation：After estimation, row interpolation is entered by the way of adaptive motion compensated, according to scene
The result of shear, using different interpolation methods.
The implementation process in described scene change detecte stage is：During judging whether to produce scene shear, because
To be to compare consecutive frame by calculating the method for colour consistency vector, so CCVD calculation is based on pixel,
The computation complexity in scene detection-phase can so be increased.Therefore in calculating process, it is not necessary to CCVD is solved to each piece,
But a part of macro block is chosen to calculate CCVD by the method for down-sampling, taken in experiment be according to the macro block divided by
Number of macroblocks is chosen according to ratio to calculate CCVD, then judges scene shear by comparing CCVD and the empirical value T set
Whether.
The implementation process of described motion estimation stage is：Block- matching searching algorithm, profit are improved according to scene shear result
It is a certain given in former frame or a later frame by each piece in present frame with the correlation of motion vector over time and space
Region in match block obtained according to matching criterior, motion vector is calculated by the relative position of match block and current block.
The implementation process in described motion compensated interpolation stage is：After estimation, use is adaptive motion compensated
Mode enter row interpolation, according to the result of scene shear, using different interpolation methods.
In specific embodiments, can be operated by following mode
The scene change detecte stage：The scene detection mode that uses herein is to compare consecutive frame based on pixel value.Therefore,
Before estimation, non-uniformly distributed load is carried out to present frame, makes during segmentation central area proportion maximum, while present frame
Consecutive frame carry out same operation also in compliance with identical partition principle, CCV is asked to each piecemeal afterwards.
Calculating to CCV is actually based on pixel, and this can make in the increase of scene shear stage computation complexity.It is based on
This consideration, during CCV is solved, utilizes the spatial coherence between macro block, it is not necessary to all calculates CCV to all blocks.
In an experiment, what we took is according to the ratio followed when most starting and splitting block, is selected according to this ratio from each segmentation block
A number of macro block is taken to be calculated.As a result show, fortune in this way and not in this way to scene shear whether
Judged result is consistent, but in execution efficiency, extraction block, which calculate, in proportion possesses bigger advantage.
Motion estimation stage：Frame per second transfer algorithm based on motion compensation depends on the motion vector that estimation obtains,
Therefore, if the accuracy of motion vector is high, the performance of follow-up interpolation algorithm can be improved to a certain extent, by present frame
Each piece, match block is obtained according to matching criterior in a certain given region of former frame or a later frame, by match block and currently
The relative position of block calculates motion vector.
Block matching algorithm is the classic algorithm in images match, because block matching algorithm realizes simple, the accurate spy of calculating
Put and be widely used in estimation, its main thought is to divide the image into the fritter for 16*16 or 8*8 pixels, is utilized
The correlation of motion vector over time and space, each block search out most in the hunting zone of reference frame according to matching criterior
Good match block position.But the shortcomings that this method, is that amount of calculation is very big.Herein, in order to reduce amount of calculation, we
Using the method that frame is carried out to non-uniformly distributed load, during segmentation, central area is occupied larger proportion, then taken out according to the ration of division
Take the block of each region varying number to be calculated, be thus not required to calculate each macro block so that amount of calculation greatly reduces.
The motion compensated interpolation stage：After estimation, motion compensation is carried out using adaptive motion compensated mode and inserted
Value.Shown in interleave result equation below (1)：
Wherein fn and fn+1 is two consecutive frames, and fn is to predict the frame come, and s represents a segmentation block, and MV is that the block is corresponding
Motion vector, λ is weight, herein we weight λ is adjusted according to the result of scene shear, equation below (2)
It is shown：
If the CCVD between two segmentation blocks has exceeded threshold value T, now adaptive weight λ=1 or 0, namely do not send out
Raw scene shear, then row interpolation is entered to present frame using forward frame or backward frame, otherwise adaptive weight λ=0.5, scene
Generation shear, now need to enter row interpolation using bidirectional frame.
Claims (5)
1. a kind of frame per second transfer algorithm based on scene change detecte, it is characterised in that comprise the following steps：
(1) scene change detecte is carried out to low frame-rate video sequence；
(2) computing estimation is carried out according to the result of scene shear；
(3) adaptive motion compensated interpolation is carried out after computing estimation, and judges whether moving scene occurs shear, if it happens cut
Become and then carry out forward interpolation processing, two-way interpolation processing is carried out if shear not occurring；
(4) result carried out after interpolation processing is high frame-rate video sequence.
A kind of 2. frame per second transfer algorithm based on scene change detecte according to claim 1, it is characterised in that the step
Suddenly comprising the following steps that for scene change detecte is carried out in (1)：Non-uniformly distributed load first is carried out to present frame, makes center during segmentation
Domain proportion is maximum, while the consecutive frame of present frame carries out same operation also in compliance with identical partition principle, afterwards to every
One piecemeal seeks colour consistency vector CCV.
A kind of 3. frame per second transfer algorithm based on scene change detecte according to claim 1, it is characterised in that the step
Suddenly estimation comprises the following steps that in (2)：By each piece in present frame, in a certain given area of former frame or a later frame
Match block is obtained according to matching criterior in domain, motion vector is calculated by the relative position of match block and current block.
A kind of 4. frame per second transfer algorithm based on scene change detecte according to claim 1, it is characterised in that the step
Suddenly in (3) after estimation, comprising the following steps that for motion compensated interpolation is carried out using adaptive motion compensated mode, wherein
Shown in interleave result equation below：
<mrow>
<msub>
<mover>
<mi>f</mi>
<mo>^</mo>
</mover>
<mi>n</mi>
</msub>
<mo>&lsqb;</mo>
<mi>s</mi>
<mo>,</mo>
<mi>v</mi>
<mo>&rsqb;</mo>
<mo>=</mo>
<msub>
<mi>&lambda;f</mi>
<mrow>
<mi>n</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>&lsqb;</mo>
<mi>s</mi>
<mo>-</mo>
<mi>M</mi>
<mi>V</mi>
<mo>&rsqb;</mo>
<mo>+</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mi>&lambda;</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>f</mi>
<mrow>
<mi>n</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>&lsqb;</mo>
<mi>s</mi>
<mo>+</mo>
<mi>M</mi>
<mi>V</mi>
<mo>&rsqb;</mo>
</mrow>
Wherein fn and fn+1 is two consecutive frames, and fn is to predict the frame come, and s represents a segmentation block, and MV is transported corresponding to the block
Dynamic vector, λ are weights.
A kind of 5. frame per second transfer algorithm based on scene change detecte according to claim 4, it is characterised in that the step
Suddenly judge whether moving scene occurs shear and judged by weight λ in (3), wherein weight λ calculation formula is as follows：
Wherein CCVD is that the colour consistency vector of the macro block at two frame same positions make the difference and compares to obtain；If two
CCVD between segmentation block has exceeded the empirical value T initially set up, now adaptive weight λ=1 or 0, namely do not occur
Scene shear, then row interpolation is entered to present frame using forward frame or backward frame, otherwise adaptive weight λ=0.5, scene hair
Raw shear, now needs to enter row interpolation using bidirectional frame.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710944115.XA CN107707916A (en) | 2017-09-30 | 2017-09-30 | A kind of frame per second transfer algorithm based on scene change detecte |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710944115.XA CN107707916A (en) | 2017-09-30 | 2017-09-30 | A kind of frame per second transfer algorithm based on scene change detecte |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107707916A true CN107707916A (en) | 2018-02-16 |
Family
ID=61184153
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710944115.XA Pending CN107707916A (en) | 2017-09-30 | 2017-09-30 | A kind of frame per second transfer algorithm based on scene change detecte |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107707916A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110049205A (en) * | 2019-04-26 | 2019-07-23 | 湖南科技大学 | The detection method that video motion compensation frame interpolation based on Chebyshev matrix is distorted |
| CN111277895A (en) * | 2018-12-05 | 2020-06-12 | 阿里巴巴集团控股有限公司 | Video frame interpolation method and device |
| WO2022143078A1 (en) * | 2020-12-28 | 2022-07-07 | 深圳创维－Rgb电子有限公司 | Video automatic motion compensation method, apparatus, and device, and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102665061A (en) * | 2012-04-27 | 2012-09-12 | 中山大学 | Motion vector processing-based frame rate up-conversion method and device |
| CN104219533A (en) * | 2014-09-24 | 2014-12-17 | 苏州科达科技股份有限公司 | Bidirectional motion estimating method and video frame rate up-converting method and system |
| CN104378589A (en) * | 2014-11-11 | 2015-02-25 | 深圳市视晶无线技术有限公司 | Low-bit-rate video code transmission reconstruction method and system |
-
2017
- 2017-09-30 CN CN201710944115.XA patent/CN107707916A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102665061A (en) * | 2012-04-27 | 2012-09-12 | 中山大学 | Motion vector processing-based frame rate up-conversion method and device |
| CN104219533A (en) * | 2014-09-24 | 2014-12-17 | 苏州科达科技股份有限公司 | Bidirectional motion estimating method and video frame rate up-converting method and system |
| CN104378589A (en) * | 2014-11-11 | 2015-02-25 | 深圳市视晶无线技术有限公司 | Low-bit-rate video code transmission reconstruction method and system |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111277895A (en) * | 2018-12-05 | 2020-06-12 | 阿里巴巴集团控股有限公司 | Video frame interpolation method and device |
| CN111277895B (en) * | 2018-12-05 | 2022-09-27 | 阿里巴巴集团控股有限公司 | Video frame interpolation method and device |
| CN110049205A (en) * | 2019-04-26 | 2019-07-23 | 湖南科技大学 | The detection method that video motion compensation frame interpolation based on Chebyshev matrix is distorted |
| WO2022143078A1 (en) * | 2020-12-28 | 2022-07-07 | 深圳创维－Rgb电子有限公司 | Video automatic motion compensation method, apparatus, and device, and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102291578B (en) | Apparatus and method for frame rate conversion | |
| CN104219533B (en) | A kind of bi-directional motion estimation method and up-conversion method of video frame rate and system | |
| CN101692692A (en) | Method and system for electronic image stabilization | |
| US20070009038A1 (en) | Motion estimator and motion estimating method thereof | |
| CN107707916A (en) | A kind of frame per second transfer algorithm based on scene change detecte | |
| CN102025960B (en) | Motion compensation de-interlacing method based on adaptive interpolation | |
| CN106303546B (en) | Conversion method and system in a kind of frame rate | |
| US8532409B2 (en) | Adaptive motion search range determining apparatus and method | |
| US20110050993A1 (en) | Motion estimating method and image processing apparatus | |
| CN101557516B (en) | Video quality evaluation method and device | |
| CN108615241B (en) | Rapid human body posture estimation method based on optical flow | |
| CN104123733B (en) | A kind of method of motion detection and reduction error rate based on Block- matching | |
| KR20060093787A (en) | Apparatus and method for estimating motion | |
| CN111178261A (en) | Face detection acceleration method based on video coding technology | |
| CN116170650A (en) | Video frame inserting method and device | |
| CN103152566B (en) | A kind of video frame rate method for improving | |
| EP3596698B1 (en) | Motion estimation method and apparatus for plurality of frames | |
| TW200941405A (en) | Motion estimation method and related apparatus | |
| CN106658024B (en) | A kind of quick method for video coding | |
| US20080187049A1 (en) | Moving image encoding device and moving image encoding method | |
| CN105306952B (en) | A method of it reducing side information and generates computation complexity | |
| CN104243991A (en) | Side information generation method and device | |
| JP4608136B2 (en) | Motion vector and parallax vector detection device | |
| CN201509247U (en) | Image processing apparatus for deciding pixel value of interpolation position | |
| JP3094893B2 (en) | Video encoding device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180216 |