CN110072103A - Video Fast Compression method, HD video system, 4K video system based on ROI - Google Patents
Video Fast Compression method, HD video system, 4K video system based on ROI Download PDFInfo
- Publication number
- CN110072103A CN110072103A CN201910198941.3A CN201910198941A CN110072103A CN 110072103 A CN110072103 A CN 110072103A CN 201910198941 A CN201910198941 A CN 201910198941A CN 110072103 A CN110072103 A CN 110072103A
- Authority
- CN
- China
- Prior art keywords
- block
- video
- coefficient
- roi
- ctu
- 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
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/13—Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
-
- 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
- H04N19/167—Position within a video image, e.g. region of interest [ROI]
-
- 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
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/96—Tree coding, e.g. quad-tree coding
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention belongs to HD video processing technology fields, disclose a kind of video Fast Compression method, HD video system, 4K video system based on ROI;The video sequence of input is pre-processed respectively, obtains different image blocks;Calculate the characteristic value of each image block static region;Based on the coordinate of block, the position feature value of calculation block;Using frame difference method, the motion feature coefficient of calculation block;Static nature and position feature to block carry out mathematical modeling, obtain the static concern coefficient of the overall situation of block;Static state concern coefficient and kinematic coefficient are normalized respectively, are mapped in fixed interval range;Weight is distributed with dynamic concern coefficient for the static concern coefficient after normalization, the human eye concern coefficient of block is calculated;Based on the human eye of block concern coefficient, the quad-tree partition model of video compression algorithm is corrected.The present invention makes video compression algorithm on the basis of not influencing people's perception, significantly reduces the processing time of algorithm.
Description
Technical field
The invention belongs to HD video processing technology field more particularly to a kind of video Fast Compression method based on ROI,
HD video system, 4K video system.
Background technique
Currently, the immediate prior art: with the appearance of the video of the higher resolutions such as HD video, 4K video, passing
The video compression algorithm of system can no longer meet the application of these scenes.Although this Lossy Compression Algorithm of such as H264 is theoretically
The size of video can be compressed as much as possible, but the content of video can also lose most, for the user of decoding end
For, it is a kind of loss that can not be born.So being woven in the video volume for having developed a new generation on the basis of H264 via JCTVC group
Decoding standard HEVC, also known as H265.HEVC is improved on the basis of H264, and the multiple technologies in H264 are optimized, including
Bigger CTU, more prediction modes etc..The algorithm, can be under identical compression performance, by the big of video compared with H264
Small to save nearly half or so, so HEVC, which can have, preferably to be showed under identical channel width, user is it can also be seen that thin
Save video content more abundant.In the cataloged procedure of HEVC, need to carry out CTU the division of quaternary tree form, to obtain
Optimal coding CU.In partition process, HEVC has generallyd use RDO to calculate the rate distortion costs of current CU block, by right
Each CU carries out recursive calculation, obtains the smallest cost value.A kind of coding mode, HEVC are realized in compression bit rate in this way
Best adaptation between video distortion.
In the encoder of HEVC, due to needing the CU to each layer to carry out recursive calculation cost value, algorithm is caused to need wave
Optimal coding parameter can just be obtained by taking a large amount of time.And by the compressed video of these coding parameters, in decoding end
User may can't feel apparent difference.This has resulted in some disadvantages of the HEVC in video compression: as encoded
Overlong time, can not after real time codec, video decoding user experience it is unobvious etc..Simultaneously as HEVC needs on coding
The a large amount of time is wasted, which results in HEVC can not apply in some real-time scenes, such as high definition conference, live streaming.
In conclusion problem of the existing technology is: HEVC needs to consume a large amount of calculating time, causes HEVC can not
It applies in some real-time scenes, in high definition conference, high-definition live broadcasting, has greatly delayed the popularization of HEVC.
The present invention carries out mathematical modeling by paying close attention to characteristic to human eye, and is drawn using the quaternary tree of Modifying model HEVC
Point, the compression time of video can be reduced significantly while not influencing user's subjective feeling, solving HEVC can not be real-time
The defect of application.
Solve the difficulty of above-mentioned technical problem:
Mathematical modeling is carried out by paying close attention to characteristic to human eye, obtains human eye to the concern coefficient of video to correct the four of HEVC
Fork tree partition mode, is Major Difficulties of the invention.
Solve the meaning of above-mentioned technical problem:
Mathematical modeling is carried out by paying close attention to characteristic to human eye, the quad-tree partition of HEVC is corrected using mathematical model, it can
To reduce compression time while not influencing user's subjective experience.
Summary of the invention
In view of the problems of the existing technology, the present invention provides a kind of video Fast Compression method, high definition based on ROI
Video system, 4K video system.
The invention is realized in this way a kind of video Fast Compression method based on ROI, the video based on ROI are fast
Fast compression method includes:
The first step pre-processes the video of input, obtains corresponding block;
Second step calculates each piece of comentropy, obtains the static nature F1 of block;
Third step calculates each piece of position feature F2 according to the coordinate of block;
4th step obtains the static concern coefficient w1 of block based on static nature and position feature;
5th step calculates each piece of motion feature using frame difference method, obtains the movement concern coefficient w2 of block;
6th step is respectively normalized the static concern coefficient of block and dynamic concern coefficient, is two coefficient distribution
Specific weight obtains final ROI coefficient, corrects quad-tree partition model using ROI coefficient.
Further, the video of input is pre-processed in the first step, obtains the process of relevant block are as follows: video sequence
Into after encoder, a space coordinates are established according to the size of the resolution ratio of video and CTU block, each CTU block is mapped
Onto corresponding spatial position coordinate, each CTU block is indexed by space coordinate.
Further, in the second step calculation block static nature process are as follows: in statistics coordinate system in each coordinate points
The pixel value of CTU block, obtains picture element matrix;The comentropy of each CTU block is calculated according to picture element matrix;The wherein calculating of comentropy
Mode is as follows:
Further, in the third step calculation block position feature process are as follows: the coordinate of each CTU is sent into encoder,
The space eigenvalues of the CTU are calculated as follows:
Enhance the weight of intermediate region, and weakens neighboring area.
Further, the calculating process of static concern coefficient is as follows in the 4th step: using comentropy as each CTU block
Weight factor considers human eye for the concern characteristic of CTU block position, and the static concern coefficient of each CTU block calculates as follows:
According to above formula, the static concern coefficient of CTU is related to two values: entropy and coordinate;Only current CTU block is close to center
Cut the information for including it is more when, cause the concern of people.
Further, the calculating process of motion feature is as follows in the 5th step: according to the playing sequence of video, using current
Frame subtracts former frame and obtains residual matrix, and residual matrix is divided into different sub- residual matrixes according to the size of CTU block, statistics
The sum of each residual matrix;Using the sum of residual error submatrix divided by the size for corresponding to CTU block, the residual error for obtaining each CTU block is equal
Value, using the residual error mean value of CTU block divided by global mean value, obtains the motion feature of the CTU.
Further, global concern coefficient is calculated in the 6th step and corrects the calculating process of quad-tree partition model are as follows:
Pay close attention to coefficient normalization:
W=α ws+(1-α)wd;
In formula, wsAnd wdRespectively indicate the static state and dynamic concern coefficient of CTU block;α is 0~0.5 according to statistics;
Correct quad-tree partition mode:
When the concern coefficient of CU is more than threshold value, it is determined as ROI block, using the coding mode that HEVC is set.And it is less than threshold
When value, it is determined as NROI block, since user will not generate apparent reaction to NROI block, so by reducing the volume to NROI block
Code complexity, it is possible to reduce the compression time of video.
Another object of the present invention is to provide the high definition of a kind of video Fast Compression method described in application based on ROI views
Display system.
Another object of the present invention is to provide the 4K videos of a kind of video Fast Compression method described in application based on ROI
System.
In conclusion advantages of the present invention and good effect are as follows: the present invention pays close attention to characteristic using human eye to correct HEVC's
Quad-tree partition, the HEVC compression algorithm of comparison with standard, the present invention can save while not influencing user's subjective feeling
30% or so video encoding time solves the problems, such as that current HEVC can not be applied to real time video processing.
Detailed description of the invention
Fig. 1 is the video Fast Compression method flow diagram provided in an embodiment of the present invention based on ROI.
Fig. 2 is eye tracker test data schematic diagram provided in an embodiment of the present invention.
Fig. 3 is the static ROI concern model schematic provided by the invention based on variance.
Fig. 4 is the static ROI concern model schematic provided by the invention based on entropy.
Fig. 5 is position ROI concern model schematic provided by the invention.
Fig. 6 is dynamic ROI concern model schematic provided by the invention.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to embodiments, to the present invention
It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to
Limit the present invention.
It needs to consume a large amount of calculating time for HEVC, causes HEVC that can not apply in some real-time scenes, such as high definition
In meeting, high-definition live broadcasting, greatly delayed HEVC promote the problem of.Present invention utilizes the ROI characteristic revision HEVC of human eye
Quad-tree partition structure, solve the problems, such as that current HEVC can not be applied to real time video processing.
Application principle of the invention is explained in detail with reference to the accompanying drawing.
As shown in Figure 1, the video Fast Compression method provided in an embodiment of the present invention based on ROI the following steps are included:
S101: pre-processing the video of input, obtains corresponding block;
S102: each piece of comentropy is calculated, the static nature F1 of block is obtained;
S103: according to the coordinate of block, each piece of position feature F2 is calculated;
S104: based on static nature and position feature, the static concern coefficient w1 of block is obtained;
S105: calculating each piece of motion feature using frame difference method, obtains the movement concern coefficient w2 of block;
S106: being respectively normalized the static concern coefficient of block and dynamic concern coefficient, is that two coefficient distribution are special
Fixed weight obtains final ROI coefficient, corrects quad-tree partition model using ROI coefficient.
In a preferred embodiment of the invention, S101 step pre-processes the video of input, obtains corresponding block
Mode are as follows: after video sequence enters encoder, a space coordinates are established according to the size of the resolution ratio of video and CTU block,
Each CTU block is mapped on corresponding spatial position coordinate, in this way by establishing mathematical coordinates system to input video sequence, just
Each CTU block can be indexed by space coordinate.After having pre-processed, the coordinate information of CTU is sent together with video
Enter encoder, subsequent algorithm can extract position feature according to the coordinate of block.
In a preferred embodiment of the invention, the static nature mode of S102 step calculation block are as follows: by the place of previous step
Reason, encoder can be indexed according to each piece of coordinate pair of block, therefore, carry out circular treatment to the coordinate of image first,
Obtain each piece of static nature.Fig. 3, Fig. 4, which are shown, carries out showing for mathematical modeling to the static nature of block using variance and entropy
It is intended to.In view of the variation range of variance is excessive, the present invention models the static region of block using entropy, and the calculating of entropy is public
Formula:
In cyclic process, it is contemplated that the pixel variation range of image block is 0~255, so passing through radix sorting
The number that mode counts each point in current block can be in the hope of the entropy of current block.
In a preferred embodiment of the invention, the position feature mode of S103 step calculation block are as follows: by each piece of coordinate
It is sent into encoder, the space eigenvalues of the block are calculated as follows:
Eye tracker test result according to Fig.2, it was demonstrated that human eye is more sensitive to the middle section of image, this quick
Perception is mostly derived from two aspects: the shooting behavior of photographer and user's watches behavior attentively.Always will just because of photographer
Interested region takes the middle section of video, so we can not prove that the impact factor of image location information is higher than
Texture information, so the present invention has carried out mathematical modeling to the location information of image.Fig. 5 show the present invention to position feature into
The schematic diagram of row mathematical modeling introduces the exponential increase factor, in this way when the position feature to each piece carries out mathematical modeling
When far from picture centre, ROI impact factor will become smaller block, and when close to picture centre, ROI impact factor becomes larger.
In a preferred embodiment of the invention, S104 step calculates the mode of global static concern coefficient are as follows: according to fig. 2
Eye tracker test data, discovery human eye is all more sensitive for location information and content information, so in order to balance in image
This characteristic is set in invention with entropy as concern coefficient, the coordinate of block is the calculation of weight coefficient, real process
In, calculation formula is as follows:
For each of image piece, it is calculated the entropy of current block first, the calculation of entropy such as step S102, then
By the feature of the position of the step S103 current block being calculated, two characteristic values are finally combined as image current block
Static state concern coefficient.Concern coefficient reflects the content and the two information of position of block, only when the content of block is more complicated, and
When positioned at central region, higher concern coefficient can be just obtained.
In a preferred embodiment of the invention, S105 step obtains the motion feature of image using frame difference method, the process
Processing mode are as follows: see that the movement of block in image strongly influences than stronger the concern characteristic of human eye according to Fig. 4, so according to
The playing sequence of video obtains the motion feature of current block using frame difference method.Firstly, using present frame subtract former frame obtain it is residual
Poor matrix:
D (x, y)=| fk(x,y)-fk-1(x,y)| (4)
Herein, it is poor make using the pixel of same coordinate in two field pictures, acquires residual matrix.The residual matrix
The mass motion information of image is represented, and HEVC is using single block as coding region, so needing the movement of image
Residual matrix is divided into residual error submatrix according to the size of block in each piece by information MAP, makes residual error submatrix
Size is equal to the size of block.For each of image piece, have:
Above formula indicates that the element of residual error submatrix and the pixel of current block correspond.Using residual error submatrix and make
For the motion information of current block:
F=Sum (i, j) (6)
Since the resolution ratio of input video sequence can not just be the integral multiple of current block, so motion information needs are examined
Consider the number in block comprising pixel:
By the processing of formula (7), the motion information of block is just unrelated with the size of block.
In a preferred embodiment of the invention, the process of the quad-tree partition mode of S106 step amendment HEVC are as follows: pass through
Static ROI characteristic and dynamic ROI characteristic to video carry out the concern coefficient that mathematical modeling obtains and are not in a magnitude, institute
First two coefficients to be normalized respectively:
Formula (8), formula (9) divide ROI characteristic for a magnitude, in order to by static ROI characteristic and dynamic ROI characteristic into
Row combination needs to match two characteristics using a weight coefficient, using following model foundation static state ROI characteristic and dynamic ROI
Relationship between characteristic:
W=α ws+(1-α)wd (10)
By largely testing, weight factor size that the present invention uses for 0.3, can take into account well static characteristic with
Dynamic characteristic.Quaternary tree in HEVC calculates the cost that mode calculates each layer using RDO, decides whether to continue drawing for CU
Point:
J=D+ λ R (11)
When carrying out the division of quaternary tree, HEVC calculates the cost of every layer of CU according to above formula, the calculating for R, needs time
Each coding mode is gone through, is encoded, and coding is obtained into staining effect to encoder, encoder loses according to finally obtained rate
True cost J decides whether to continue to divide.This traversal calculating process complexity of HEVC is very high, and causes the scramble time
Too long principal element.In the present invention, non-by quickly stopping using the quad-tree partition mode of concern coefficient amendment HEVC
The recurrence of ROI region reduces the scramble time:
The method of determination of threshold value in formula are as follows:
Above formula is analyzed it is found that the concern coefficient ratio of ROI region is larger, so algorithm will not modify the depth of ROI block
Degree, but for non-ROI block, since concern coefficients comparison is small, so non-ROI block can be quick when carrying out quaternary tree recursive calculation
Cut-off, to realize the function of video Fast Compression.
Application effect of the invention is explained in detail below with reference to experiment.
It is the present invention and comparison of the standard HEVC in performance as shown in table 1 below, the test software of HEVC uses HM16.0.
Table 1 and HM16.0 coding efficiency comparison (opening code control)
Table 1 lists the comparing result of the present invention with HM16.0, and cycle tests is the standard that resolution ratio is 3840*2160
Yuv format simulates channel from the situation under 4Mbps~16Mbps respectively.The two is given in table in scramble time, the area ROI
The Δ PSNR in the domain and Δ PSNR of non-ROI region.As can be seen that can be seen according to the analysis to ROI region and the region NROI
Out, inside ROI region, performance is slightly improved, and inside the region NROI, performance causes a degree of decline, and encodes
Time then saves 30% or so.Due to reducing the encoder complexity of the region NROI part, the section on the scramble time has been obtained
It saves.And the performance of ROI region part is slightly promoted, and is and the area NROI since code rate control needs to keep certain compression bit rate
Domain produces bigger distortion (distortion is bigger, and code rate is smaller) since coding depth reduces, so can vacant a part of code rate out
ROI region is distributed to, which results in the slightly promotions of ROI region PSNR.Table 2 illustrates the performance pair when closing code control
Than.
Table 2 and HM16.0 coding efficiency comparison (closing code control)
As the test result that table 2 obtains, in the region NROI, a part of performance can be lost still to promote the coding of HEVC
Speed (30% or so), but in ROI region, then performance is no longer slightly to rise, and occurs the fluctuation of certain amplitude instead, but
It is this fluctuation is by a small margin our acceptables.
Traditional HEVC video compression algorithm gives the weight of same degree to ROI region and the region NROI, carries out same
Etc. degree compression, the shortcomings that causing video compress overlong time, and the present invention considers human eye when watching video not
Same degree of concern gives the region NROI lesser CU depth, and to save the scramble time, and the details of ROI region is due to CU depth
Spend constant so will not influence too much.The present invention can be substantially reduced video in the case where not influencing user's perception
Handle the time.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.
Claims (9)
1. a kind of video Fast Compression method based on ROI, which is characterized in that the video Fast Compression method based on ROI
Include:
The first step pre-processes the video of input, obtains corresponding block;
Second step calculates each piece of comentropy, obtains the static nature F1 of block;
Third step calculates each piece of position feature F2 according to the coordinate of block;
4th step obtains the static concern coefficient w1 of block based on static nature and position feature;
5th step calculates each piece of motion feature using frame difference method, obtains the movement concern coefficient w2 of block;
6th step is respectively normalized the static concern coefficient of block and dynamic concern coefficient, is that two coefficient distribution are specific
Weight, obtain final ROI coefficient, use ROI coefficient correct quad-tree partition model.
2. the video Fast Compression method based on ROI as described in claim 1, which is characterized in that defeated in the first step
The video entered is pre-processed, and the process of relevant block is obtained are as follows: after video sequence enters encoder, according to the resolution ratio of video with
The size of CTU block establishes a space coordinates, and each CTU block is mapped on corresponding spatial position coordinate, space is passed through
Coordinate is indexed each CTU block.
3. the video Fast Compression method based on ROI as described in claim 1, which is characterized in that calculated in the second step
The static nature process of block are as follows: the pixel value of the CTU block in statistics coordinate system in each coordinate points obtains picture element matrix;According to
Picture element matrix calculates the comentropy of each CTU block;Wherein the calculation of comentropy is as follows:
4. the video Fast Compression method based on ROI as described in claim 1, which is characterized in that calculated in the third step
The position feature process of block are as follows: the coordinate of each CTU is sent into encoder, the space eigenvalues of the CTU are calculated as follows:
Enhance the weight of intermediate region, and weakens neighboring area.
5. the video Fast Compression method based on ROI as described in claim 1, which is characterized in that static in the 4th step
The calculating process for paying close attention to coefficient is as follows: using comentropy as the weight factor of each CTU block, considering human eye for CTU block position
Concern characteristic, the static concern coefficient of each CTU block calculates as follows:
According to above formula, the static concern coefficient of CTU is related to two values: entropy and coordinate;Only current CTU block anxious packet in
When the information contained is more, cause the concern of people.
6. the video Fast Compression method based on ROI as described in claim 1, which is characterized in that moved in the 5th step
The calculating process of feature is as follows: according to the playing sequence of video, subtracting former frame using present frame and obtains residual matrix, by residual error
Matrix is divided into different sub- residual matrixes according to the size of CTU block, counts the sum of each residual matrix;Use residual error submatrix
Sum divided by the size of corresponding CTU block, obtain the residual error mean value of each CTU block, using CTU block residual error mean value divided by it is global
Value, obtains the motion feature of the CTU.
7. the video Fast Compression method based on ROI as described in claim 1, which is characterized in that calculated in the 6th step
Overall situation concern coefficient and the calculating process for correcting quad-tree partition model are as follows:
Pay close attention to coefficient normalization:
W=α ws+(1-α)wd;
In formula, wsAnd wdRespectively indicate the static state and dynamic concern coefficient of CTU block;α is 0~0.5 according to statistics;
Correct quad-tree partition mode:
When the concern coefficient of CU is more than threshold value, it is determined as ROI block, using the coding mode that HEVC is set.And when being less than threshold value,
It is determined as NROI block, since user will not generate apparent reaction to NROI block, so multiple to the coding of NROI block by reducing
Miscellaneous degree, it is possible to reduce the compression time of video.
8. a kind of HD video system using the video Fast Compression method described in claim 1~7 any one based on ROI
System.
9. a kind of 4K video system using the video Fast Compression method described in claim 1~7 any one based on ROI.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910198941.3A CN110072103A (en) | 2019-03-15 | 2019-03-15 | Video Fast Compression method, HD video system, 4K video system based on ROI |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910198941.3A CN110072103A (en) | 2019-03-15 | 2019-03-15 | Video Fast Compression method, HD video system, 4K video system based on ROI |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110072103A true CN110072103A (en) | 2019-07-30 |
Family
ID=67366221
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910198941.3A Pending CN110072103A (en) | 2019-03-15 | 2019-03-15 | Video Fast Compression method, HD video system, 4K video system based on ROI |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110072103A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101853510A (en) * | 2010-04-20 | 2010-10-06 | 上海大学 | Movement perception model extraction method based on time-space domain |
CN104079934A (en) * | 2014-07-14 | 2014-10-01 | 武汉大学 | Method for extracting regions of interest in real-time video communication |
CN104392212A (en) * | 2014-11-14 | 2015-03-04 | 北京工业大学 | Method for detecting road information and identifying forward vehicles based on vision |
CN106937118A (en) * | 2017-03-13 | 2017-07-07 | 西安电子科技大学 | A kind of bit rate control method being combined based on subjective area-of-interest and time-space domain |
US20180139456A1 (en) * | 2008-11-17 | 2018-05-17 | Checkvideo Llc | Analytics-modulated coding of surveillance video |
CN109379594A (en) * | 2018-10-31 | 2019-02-22 | 北京佳讯飞鸿电气股份有限公司 | Video coding compression method, device, equipment and medium |
-
2019
- 2019-03-15 CN CN201910198941.3A patent/CN110072103A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180139456A1 (en) * | 2008-11-17 | 2018-05-17 | Checkvideo Llc | Analytics-modulated coding of surveillance video |
CN101853510A (en) * | 2010-04-20 | 2010-10-06 | 上海大学 | Movement perception model extraction method based on time-space domain |
CN104079934A (en) * | 2014-07-14 | 2014-10-01 | 武汉大学 | Method for extracting regions of interest in real-time video communication |
CN104392212A (en) * | 2014-11-14 | 2015-03-04 | 北京工业大学 | Method for detecting road information and identifying forward vehicles based on vision |
CN106937118A (en) * | 2017-03-13 | 2017-07-07 | 西安电子科技大学 | A kind of bit rate control method being combined based on subjective area-of-interest and time-space domain |
CN109379594A (en) * | 2018-10-31 | 2019-02-22 | 北京佳讯飞鸿电气股份有限公司 | Video coding compression method, device, equipment and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9143776B2 (en) | No-reference video/image quality measurement with compressed domain features | |
JP5149188B2 (en) | Content-driven transcoder that uses content information to coordinate multimedia transcoding | |
CN104853191B (en) | A kind of HEVC fast encoding method | |
Ray et al. | A low-complexity video encoder for equirectangular projected 360 video content | |
US12028519B2 (en) | System and method for video coding | |
CN110521206A (en) | Code device, decoding apparatus, coding method and coding/decoding method | |
CN110495168A (en) | Code device, decoding apparatus, coding method and coding/decoding method | |
CN100366091C (en) | Video frequency compression | |
Lei et al. | Region adaptive R-$\lambda $ model-based rate control for depth maps coding | |
Xu et al. | Consistent visual quality control in video coding | |
US11368686B2 (en) | Encoder and decoder | |
US20230344996A1 (en) | Encoder, decoder, encoding method, and decoding method | |
Lin et al. | Efficient quadtree search for HEVC coding units for V-PCC | |
Boujut et al. | Weighted-MSE based on Saliency map for assessing video quality of H. 264 video streams | |
CN110072103A (en) | Video Fast Compression method, HD video system, 4K video system based on ROI | |
Chen et al. | Video coding using dynamic texture synthesis | |
Wang et al. | Quality assessment for MPEG-2 video streams using a neural network model | |
CN110495179A (en) | Code device, decoding apparatus, coding method and coding/decoding method | |
Jung | Comparison of video quality assessment methods | |
Mallik et al. | HEVC based mixed-resolution stereo video codec | |
Zhang et al. | Scalable Point Cloud Attribute Compression | |
Uhrina et al. | Impact of compression on the video quality | |
Tancharoen et al. | Automatic face color segmentation based rate control for low bit-rate video coding | |
CN109889829A (en) | The point self-adapted compensation of the quick sample product of 360 degree of videos | |
Miličević et al. | An approach to interactive multimedia systems through subjective video quality assessment in H. 264/AVC standard |
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 |
Application publication date: 20190730 |
|
RJ01 | Rejection of invention patent application after publication |