CN108596855A - A kind of video image quality Enhancement Method, device and video picture quality enhancement method - Google Patents
A kind of video image quality Enhancement Method, device and video picture quality enhancement method Download PDFInfo
- Publication number
- CN108596855A CN108596855A CN201810403772.8A CN201810403772A CN108596855A CN 108596855 A CN108596855 A CN 108596855A CN 201810403772 A CN201810403772 A CN 201810403772A CN 108596855 A CN108596855 A CN 108596855A
- Authority
- CN
- China
- Prior art keywords
- image
- video
- frame number
- quality
- default frame
- 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
- 238000000034 method Methods 0.000 title claims abstract description 119
- 238000013507 mapping Methods 0.000 claims description 108
- 230000005484 gravity Effects 0.000 claims description 26
- 238000012545 processing Methods 0.000 claims description 22
- 238000001514 detection method Methods 0.000 claims description 12
- 238000013527 convolutional neural network Methods 0.000 claims description 11
- 239000000284 extract Substances 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 5
- 230000002708 enhancing effect Effects 0.000 description 17
- 230000000694 effects Effects 0.000 description 9
- 239000011159 matrix material Substances 0.000 description 9
- 238000004590 computer program Methods 0.000 description 6
- 238000000605 extraction Methods 0.000 description 5
- 230000007423 decrease Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- YBJHBAHKTGYVGT-ZKWXMUAHSA-N (+)-Biotin Chemical compound N1C(=O)N[C@@H]2[C@H](CCCCC(=O)O)SC[C@@H]21 YBJHBAHKTGYVGT-ZKWXMUAHSA-N 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 230000009191 jumping Effects 0.000 description 2
- FEPMHVLSLDOMQC-UHFFFAOYSA-N virginiamycin-S1 Natural products CC1OC(=O)C(C=2C=CC=CC=2)NC(=O)C2CC(=O)CCN2C(=O)C(CC=2C=CC=CC=2)N(C)C(=O)C2CCCN2C(=O)C(CC)NC(=O)C1NC(=O)C1=NC=CC=C1O FEPMHVLSLDOMQC-UHFFFAOYSA-N 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 238000001303 quality assessment method Methods 0.000 description 1
- 238000012372 quality testing Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Image Processing (AREA)
Abstract
This application provides a kind of video image quality Enhancement Method, device and video picture quality enhancement method, which includes：The consecutive image of the first default frame number is obtained from video current location；Whether detect in the consecutive image of the described first default frame number includes the benchmark image for meeting preset quality condition；If there are the benchmark image, using the image in the consecutive image of the described first default frame number in addition to benchmark image as reconstructed image is waited for, wait for that reconstructed image executes image reconstruction process to described based on the benchmark image.The application can use the image of high quality in video that low-quality image is reconstructed, and to enhance video image quality, improve video image quality, meet the use needs of user.
Description
Technical field
This application involves technical field of image processing, in particular to a kind of video image quality Enhancement Method, device
And video picture quality enhancement method.
Background technology
Video coding refers to through specific compress technique, and the video file of some video format, which is converted into another, to be regarded
The video file of frequency format.When by network transmission video data, in order to reduce the data traffic for needing to transmit,
Reduce transmission of video to the pressure of network, generally require and Internet video is encoded in advance, to realize the compression to video.Depending on
Although volume decreases after frequency encodes, there is distortion in the video council after coding, and video pictures quality is caused to decline.
In addition, video stream data during by network transmission, frame loss, mistake can occur due to the problems such as network delay, interim card
The problems such as position, also results in the decline of video pictures quality.Therefore, in order to improve video image quality it is necessary to video file into
The processing of row picture quality enhancement.
Current image quality Enhancement Method is carried out generally be directed to each frame image in video file at picture quality enhancement
Reason, namely the data based on each frame image carry out resolution ratio, tone, contrast, brightness, pixel etc. to image and are adjusted,
Finally obtain the video file of picture quality enhancement.But conventional picture quality enhancement method, only increased on the basis of original image
Strength is managed, or since original image is second-rate, the effect after enhancing is not obvious, and picture quality promotes limited, Huo Zheyou
It is excessive to the change of original image in enhancing processing, differing greatly between continuous different images is caused, video is not enough " smooth ".
Therefore, the effect is unsatisfactory for conventional video image picture quality enhancement, is unable to reach use demand.
Invention content
In view of this, the embodiment of the present application be designed to provide a kind of video image quality Enhancement Method, device and
Video picture quality enhancement method can be based on the higher image of quality in video and the lower image of other quality is reconstructed, weight
Image after structure can largely be influenced by the higher image of quality, and picture quality can be substantially improved, and met and used
The use demand at family.On the other hand, video image quality Enhancement Method, device and the video that the embodiment of the present application is provided are drawn
Matter Enhancement Method, can be while enhancing picture quality so that the difference between continuous different images is smaller, excessively more flat
It is sliding, improve user's impression.
In a first aspect, the embodiment of the present application provides a kind of video image quality Enhancement Method, including：
The consecutive image of the first default frame number is obtained from video current location；
Whether detect in the consecutive image of the described first default frame number includes the benchmark image for meeting preset quality condition；
If there are the benchmark image, by the image in the consecutive image of the described first default frame number in addition to benchmark image
As reconstructed image is waited for, wait for that reconstructed image executes image reconstruction process to described based on the benchmark image.
Second aspect, the embodiment of the present application provide a kind of video picture quality enhancement method, including：
From the initial position of pending video or video clip, cycle perform claim requires 1-8 is any described to regard
Frequency method for enhancing image quality, until pending video or video clip are disposed.
The third aspect, the embodiment of the present application provide a kind of video image quality intensifier, including：
Acquiring unit, the consecutive image for obtaining the first default frame number from video current location；
Whether detection unit includes meeting preset quality condition in the consecutive image for detecting the described first default frame number
Benchmark image；
Reconfiguration unit, for executing：If, will be in the consecutive image of the described first default frame number there are the benchmark image
Image in addition to benchmark image, which is used as, waits for reconstructed image, and waits for that reconstructed image executes image weight to described based on the benchmark image
Structure processing.
The embodiment of the present application first has to obtain from current location when carrying out quality enhancing to the image in video
The consecutive image of first default frame number detects whether to include meeting preset quality condition then from the consecutive image obtained
Benchmark image.If there are benchmark image, it is based on the benchmark image, figure is carried out to other images in acquired consecutive image
As reconstruction processing, the lower image of quality can be also reconstructed based on quality higher image, thus for matter itself
For measuring poor image, the image after reconstruct can largely be influenced by the higher image of quality, can be substantially
Picture quality is promoted, the use demand of user is met.On the other hand, the video image quality enhancing that the embodiment of the present application is provided
Method, apparatus and video picture quality enhancement method, can be while enhancing picture quality so that between continuous different images
Difference is smaller, excessively more smooth, improves user's impression.
To enable the above objects, features, and advantages of the application to be clearer and more comprehensible, preferred embodiment cited below particularly, and coordinate
Appended attached drawing, is described in detail below.
Description of the drawings
It, below will be to needed in the embodiment attached in order to illustrate more clearly of the technical solution of the embodiment of the present application
Figure is briefly described, it should be understood that the following drawings illustrates only some embodiments of the application, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 shows a kind of flow chart for video image quality Enhancement Method that the embodiment of the present application one is provided；
Fig. 2 shows the specific method flow charts for the image reconstruction process that the embodiment of the present application two is provided；
Fig. 3 shows the flow chart for another video image quality Enhancement Method that the embodiment of the present application three is provided；
Fig. 4 shows a kind of flow chart for video picture quality enhancement method that the embodiment of the present application four is provided；
Fig. 5 shows that the embodiment of the present application five provides a kind of concrete structure signal of video image quality intensifier
Figure；
Fig. 6 shows the structural schematic diagram for the computer equipment that the embodiment of the present application six is provided.
Specific implementation mode
To keep the purpose, technical scheme and advantage of the embodiment of the present application clearer, below in conjunction with the embodiment of the present application
Middle attached drawing, technical solutions in the embodiments of the present application are clearly and completely described, it is clear that described embodiment is only
It is some embodiments of the present application, instead of all the embodiments.The application being usually described and illustrated herein in the accompanying drawings is real
Applying the component of example can be arranged and designed with a variety of different configurations.Therefore, below to the application's for providing in the accompanying drawings
The detailed description of embodiment is not intended to limit claimed scope of the present application, but is merely representative of the selected reality of the application
Apply example.Based on embodiments herein, institute that those skilled in the art are obtained without making creative work
There is other embodiment, shall fall in the protection scope of this application.
Current video picture quality enhancement is enhanced on the basis of source images for each frame image in video file
Processing, for example, carry out resolution ratio, tone, brightness, contrast with or pixel adjustment, enhancing effect is poor, cannot be satisfied use
Demand.Based on this, a kind of video image quality Enhancement Method provided by the present application and device, can based on multiple sequential charts
The highest frame image of picture quality as in, enhances video image quality, can obtain better picture quality enhancement effect.
For ease of understanding the present embodiment, a kind of video image quality disclosed in the embodiment of the present application is increased first
Strong method describes in detail, and this method is mainly used for carrying out picture quality enhancing by the video of coding.
Shown in Figure 1, the video image quality Enhancement Method that the embodiment of the present application one provides includes：
S101：The consecutive image of the first default frame number is obtained from video current location.
When specific implementation, video is made of multiple images with sequential, is being compiled to video using compress technique
After code, the image quality of every image of composition video can be caused to have a degree of decline, and under the image quality of different images
Drop degree may be different.With under efficient video coding (High Efficiency Video Coding, HEVC) compressed encoding
For video, 6 frames or so are differed from the image in quality low ebb to the image in quality peak adjacent thereto
Image；And the video under other format compression codings, it is in quality peak from the image in quality low ebb to adjacent thereto
Image between differ more or less image.Therefore in order to realize based on the image in quality peak to other figures
As carrying out picture quality enhancing, the consecutive image that the first default frame number is obtained from the current location of video is needed.
First default frame number can specifically be set according to the actual needs, that is, the first default frame number can be with
In the video generated according to compress technique, in quality low ebb image to the image in quality peak adjacent thereto it
Between frame number difference specifically set.Such as if video carries out compressed encoding using HEVC, then can by this
One default frame number is set as any one in 3 frames to 6 frames.
S102：Whether detect in the consecutive image of the first default frame number includes the benchmark image for meeting preset quality condition.
It, be to acquired sequential chart after obtaining the consecutive image of the first default frame number when specific implementation
The picture quality of picture is detected or assesses, and whether detection wherein includes the benchmark image of preset quality condition.
Herein, preset quality condition can specifically be set according to the actual needs；For different videos,
Its picture quality is actual to be presented different trend, and the overall picture quality included by partial video is higher, can should
Preset quality condition setting it is more stringenter；For including the lower video of overall picture quality, can be by default matter
Measure the more loose of condition setting.
Specifically, may be used in following manner in the consecutive image of any one the first default frame number of detection whether include
Meet the benchmark image of preset quality condition：
One：The quality of consecutive image based on the first default frame number of non-reference picture method for evaluating quality pair is commented
Valence finds the benchmark image for meeting preset quality condition.
Non-reference picture method for evaluating quality (No Reference Image Quality Assessment, NR-IQA),
The image quality measure method of i.e. no original image information as reference.There are many realize for non-reference picture method for evaluating quality
Mode can be conventional appraisal procedure, such as calculate the corresponding index of reflection picture quality, to weigh the quality of picture,
Can be the method based on machine learning or the non-reference picture method for evaluating quality based on image information entropy etc..
Secondly：The consecutive image of first default frame number is input to advance trained two-value grader,
Judge whether to include meeting to preset according to the classification results of the consecutive image of the first default frame number of two-value grader pair
The benchmark image of quality requirements.
When specific implementation, two-value grader is to be instructed in advance using the training image for meeting preset quality condition
White silk obtains.When carrying out quality testing to the company's figure image obtained, obtained consecutive image can be sequentially input
To this, trained two-value grader, two-value grader can export the two-value classification results of each image, be in advance：Meet default
Quality requirements, or it is unsatisfactory for preset quality condition.
S103：If including to meet preset quality in the consecutive image of the first default frame number obtained from video current location
The benchmark image of condition, using the image in the consecutive image of the first default frame number in addition to benchmark image as waiting for reconstructed image, base
Reconstructed image, which is treated, in benchmark image executes image reconstruction process.
The embodiment of the present application first has to obtain the from current location when carrying out quality enhancing to the image in video
The consecutive image of one default frame number detects whether to include the base for meeting preset quality condition then from the consecutive image obtained
Quasi- image.If there are benchmark image, it is based on the benchmark image, image is carried out to other images in acquired consecutive image
Reconstruction processing can also be reconstructed the lower image of quality based on the higher image of quality, thus for quality itself
For poor image, can largely it be influenced by the higher image of quality after reconstruct, it can be largely
Promote the quality of this parts of images.Also, from the consecutive image obtained, the image content of benchmark image and other images is usual
Difference very little carries out enhancing processing using benchmark image to other images, and handling result more " smooth ", image quality is enabled to increase
Strong effect is more preferable, promotes user's impression.
Shown in Figure 2, the embodiment of the present application two provides a kind of specific method of image reconstruction process, and this method is by benchmark
Image and CNN networks are input to reconstructed image, do multilayer feature map superposition weighting processing, obtain reconstructed image.The party
Method specifically includes：
S201：By benchmark image and wait for that reconstructed image is input to convolutional neural networks, on the basis of image zooming-out first it is special
Sign mapping, and to wait for that reconstructed image extracts second feature mapping.
S202：Weighted overlap-add procedure is carried out to fisrt feature mapping and second feature mapping, obtains reconstruct Feature Mapping.
S203：Reconstructed image is obtained based on reconstruct Feature Mapping, reconstructed image is waited for using reconstructed image replacement.
When specific implementation, to benchmark image and before waiting for reconstructed image extraction Feature Mapping, first to use full
The training image set of sufficient preset quality condition is trained convolutional neural networks so that convolutional neural networks on the basis of scheme
The fisrt feature mapping extracted when picture is with waiting for that reconstructed image extracts Feature Mapping (feature map) and second feature
Mapping can be more likely to extract higher clear feature.On the basis of using convolutional neural networks image and wait for reconstructed image extract
It is to constituting benchmark image and waiting for that reconstructed image carries out convolution algorithm, extraction respectively characterizes benchmark image when Feature Mapping
Fisrt feature map and wait for reconstructed image second feature mapping, this feature mapping be characterization benchmark image eigenmatrix or
Eigenmatrix set.
On the basis of image zooming-out can characterize the benchmark image fisrt feature mapping, and for wait for reconstructed image extraction can
The second feature for waiting for reconstructed image mapping is characterized, and fisrt feature mapping is identical with the dimension that second feature maps.
There are one fisrt feature mappings, and the quantity of second feature mapping is equal with the quantity of reconstructed image is waited for.To first
It is to a fisrt feature mapping and all second when Feature Mapping and second feature mapping carry out weighted overlap-add procedure
Feature Mapping is weighted overlap-add procedure.Weighted superposition is carried out to fisrt feature mapping and second feature mapping, two can be divided into
A process, weighted sum superposition.Weighting is to be carried out to fisrt feature mapping and second feature mapping according to preset specific gravity factor
Weighting is handled.Herein, fisrt feature mapping and second feature mapping are corresponding with specific gravity factor, cut the proportion of fisrt feature mapping
Coefficient is more than the specific gravity factor of second feature mapping, and the proportion of the specific gravity factor of fisrt feature mapping and institute's second feature mapping
Coefficient and be equal to 1.
Herein, it should be noted that difference waits for that the specific gravity factor of the corresponding second feature mapping of reconstructed image can be identical
Can not also be identical, it can specifically be set according to actual reconstruct.Such as：When the specific gravity factor of second feature mapping
When identical, the needed reconstructed image of the default frame number of currently obtained first is replaced using same reconstructed image, in this way can
Simplify and calculates.And the specific gravity factor of second feature mapping can also be different, such as：Calculating a certain wait for corresponding to reconstructed image
When reconstructed image, other than the specific gravity factor corresponding to the Feature Mapping of benchmark image is big, this waits for that the feature of reconstructed image is reflected
Penetrating corresponding specific gravity factor can also be more than the specific gravity factor corresponding to other Feature Mappings with reconstructed image, with better
Retain " primitive character " that waits for reconstructed image.Since fisrt feature mapping and second feature mapping are actually multiple by feature
It is worth the matrix or set of matrices constituted, is in mapping fisrt feature when being weighted processing to fisrt feature mapping
Weighting coefficient corresponding with fisrt feature mapping is multiplied each characteristic value successively, obtains the weighting matrix of fisrt feature mapping or adds
Weight matrix set；Processing is weighted to second feature mapping, be during second feature is mapped each characteristic value successively with second
The corresponding weighting coefficient of Feature Mapping obtains the weighting matrix or weighting matrix set of second feature mapping.
It is to add fisrt feature mapping when carrying out overlap-add procedure to fisrt feature mapping and second feature mapping
Each characteristic value of weight matrix or weighting matrix set, the characteristic value with the weighting matrix corresponding position of all second feature mapping
Or the characteristic value of the corresponding position of homography is added in weighting matrix set.
Such as following examples one：
First default frame number is 4, and the consecutive image of the first default frame number obtained from the current location of video is respectively to scheme
As A, image B, image C and image D；Wherein image B is the benchmark image for meeting preset quality condition, image A, image C and figure
As D is to wait for reconstructed image.
It is the first of image B extractions after image A, image B, image C and image D are input to convolutional neural networks
Feature Mapping is (for simplification example, only by taking the calculating of eigenmatrix as an example)：
It is respectively for the second feature mapping that image A, image C and image D are extracted respectively：
The specific gravity factor of image A, image C and image D is respectively：S1、S2、S3；The specific gravity factor of image B is L, then to figure
As A, image B, image C and image D are weighted superposition, obtained reconstruct Feature Mapping N satisfactions：
Wherein, reconstruct Feature Mapping N is image A, image B and the common reconstruct Feature Mappings of image C.When calculating, figure
As the specific gravity factor S of A, image C and image D1、S2、S3Can be identical, at this point, the figure after image A, image C and image D reconstruct
As identical, calculating can be simplified.S1、S2、S3Can not also be identical, for example, when calculating the corresponding reconstructed image of image A, S1
More than S2、S3, to more preferably retain the primitive character of image A.
After obtaining reconstruct Feature Mapping, reconstruct Feature Mapping can be based on and obtain reconstructed image, and use reconstructed image
Replace needed reconstructed image.Such as：Coding (encoder) process of CNN networks can be used to obtain Feature Mapping, by solution
Code (decoder) process obtains image.
The use of the obtained reconstructed images of reconstruct Feature Mapping N is E, and specific gravity factor S such as in above-mentioned example one kind1、S2、
S3When identical, after replacing needed reconstructed image using reconstructed image, the figure after obtained reconstruct corresponding with original image sequence
As sequence is：Image E, image B, image E, image E, to complete to treat the reconstruct of reconstructed image.Specific gravity factor S1、S2、S3No
Identical result can with and so on, details are not described herein again.
In addition, other than image reconstruction process method as shown in Figure 2, other calculations can also be used.For example,
After obtaining the fisrt feature with benchmark image and mapping the second feature mapping for waiting for reconstructed image with multiple, based on benchmark image
The second feature mapping that fisrt feature maps and every is waited for reconstructed image, waits for that reconstructed image carries out at image reconstruction to every respectively
Reason.
When specific implementation, the fisrt feature mapping based on benchmark image and every second feature for waiting for reconstructed image
Mapping when waiting for that reconstructed image carries out image reconstruction process to every respectively, can also be used fisrt feature mapping and the
Two Feature Mappings are weighted the mode of superposition, and obtain second feature mapping characterization waits for that the reconstruct feature of reconstructed image is reflected
It penetrates.Wherein, the sum of every specific gravity factor for waiting for reconstructed image and benchmark image is 1, is waiting for that reconstructed image carries out figure to different
When as reconstruction processing, the specific gravity factor of benchmark image may be the same or different.
Example two：First default frame number is 4, the consecutive image point of the first default frame number obtained from the current location of video
It Wei not image A, image B, image C and image D；Wherein image B is the benchmark image for meeting preset quality condition, image A, image
C and image D is to wait for reconstructed image.
It is the first of image B extractions after image A, image B, image C and image D are input to convolutional neural networks
Feature Mapping is (for simplification example, only by taking the calculating of eigenmatrix as an example)：
It is respectively for the second feature mapping that image A, image C and image D are extracted respectively：
In above-mentioned example one, the specific gravity factor point of the image B as the image A for waiting for reconstructed image and as benchmark image
It Wei not α1And α2, the fisrt feature mapping to the mapping of the second feature of image A and image B be weighted overlap-add procedure when
It waits, obtains the band reconstruct Feature Mapping N of image AAMeet：
The specific gravity factor of image B as the image C for waiting for reconstructed image and as benchmark image is respectively β1And β2, right
The second feature mapping of image C and when being weighted overlap-add procedure of fisrt feature mapping of image B, obtain image C's
Band reconstruct Feature Mapping NCMeet：
The specific gravity factor of image B as the image D for waiting for reconstructed image and as benchmark image is respectively γ1And γ2,
When being weighted overlap-add procedure of fisrt feature mapping of second feature mapping and image B to image D, obtains image D
Band reconstruct Feature Mapping NDMeet：
Every is being obtained after the reconstruct Feature Mapping of reconstructed image, the reconstruct feature that can wait for reconstructed image based on every
Mapping obtains this and waits for the corresponding reconstructed image of reconstructed image, and is replaced using reconstructed image and corresponding wait for reconstructed image.
Such as in above-mentioned example two, the reconstruct Feature Mapping N of image A is usedAObtained reconstructed image is A ', uses figure
As the reconstruct Feature Mapping N of CCObtained reconstructed image is C ', uses the reconstruct Feature Mapping N of image DDObtained reconstructed image
For D ', corresponding after reconstructed image, the image after obtained reconstruct corresponding with original image sequence is replaced using reconstructed image
Sequence is：Image A ', image B, image C ', image D ', to complete to treat the reconstruct of reconstructed image.
Shown in Figure 3, the embodiment of the present application three also provides another video image quality Enhancement Method, including：
S301：The consecutive image of the first default frame number is obtained from video current location；
S302：Whether detect in the consecutive image of the first default frame number includes the benchmark image for meeting preset quality condition；
If so, jumping to S303；If it is not, then jumping to S305.
S303：Using the image in the consecutive image of the first default frame number in addition to benchmark image as reconstructed image is waited for, it is based on
Benchmark image treats reconstructed image and executes image reconstruction process.Jump to S304.
S301-S303 can be found in aforementioned S101-S103, and details are not described herein.
S304：Current location movement third is preset to the position after frame number, the new current location as video；Third
Default frame number is less than or equal to the first default frame number.Redirect S301.
When specific implementation, if including from the consecutive image for the first default frame number that video current location obtains
Benchmark image is then based on benchmark image in consecutive image, waiting for that reconstructed image carries out at image reconstruction in addition to benchmark image
Reason.After having executed image reconstruction process, current location is moved to the position after third presets frame number, newly as video
Current location re-executes above-mentioned S301, until video or video clip have been processed.
S305：The consecutive image for continuing the second default frame number of acquisition backward, until finding the base for meeting preset quality condition
Quasi- image.Jump to S306.
S306：Using the image from all consecutive images acquired current location in addition to benchmark image as waiting reconstructing
Image treats reconstructed image based on benchmark image and executes image reconstruction process.Jump to S307.
When specific implementation, continues the consecutive image for obtaining the second default frame number backward, refer to and the first default frame
The consecutive image of the continuous second default frame number of several consecutive images.
Herein, the consecutive image of the second default frame of acquisition backward can be continued by executing following step：
With the next frame image of the consecutive image obtained be it is interim obtain position, from obtaining position temporarily, obtain the
The consecutive image of two default frame numbers；Second default frame number is less than or equal to the first default frame number.
For example, include in video image 1, image 2, image 3, image 4 ..., totally 100 images of image 100.First
Default frame number is 3, and the second default frame number is 2.If 3 consecutive images obtained from current location are image 4, image 5 and figure
Do not include the benchmark image for meeting preset quality condition as 6, and in image 4, image 5 and image 6, with the next frame of image 6
Image namely the interim position that obtains of image 7 then continue to obtain two images backward since image 7, namely obtain image 7
And whether include meeting the benchmark image of preset quality condition in detection image 7 and image 8 with image 8.
If detecting, image 7 meets the benchmark image of preset quality condition, by image 4, image 5, image 6 and image 8
As waiting for reconstructed image, and based on the image 7 as benchmark image, image weight is executed to image 4, image 5, image 6 and image 8
Structure processing.Here image reconstruction process is identical with the image reconstruction process in above-mentioned I, therefore repeats no more.
If detecting, image 7 and image 8 do not meet preset quality condition, continue with the next frame image of image 8,
I.e. image 9 is that new interim acquisition position then continues to obtain two images namely image 9 and image backward since image 9
10, and the detection process of the quasi- image of repeating group, until the benchmark image for meeting preset quality condition is found, and will be acquired before
All images without image reconstruction process as waiting for reconstructed image, treat reconstructed image using the benchmark image found and hold
Row image reconstruction process.S307：Using the interim position that obtains as the new current location of video.Jump to S101.
Image detection for next time and corresponding reconstruction processing, since this is to the figure before new interim acquisition position
As being detected, and corresponding reconstruction processing is carried out, in order to improve treatment effeciency, next image detection position can be with
Directly since new interim acquisition position.Such as：Assuming that until just detecting benchmark image in image 9,10 --- figure
As 10, then this original image 4, image 5, image 6, image 7, image 8 are unsatisfactory for quality requirement after testing, and
Image 10 has been used to reconstruct image 4, image 5, image 6, image 7, image 8, image 9, then what is handled next time rises
Point, can directly since image 9, rather than only from current location (i.e. image 4) sliding third preset frame number (such as：2),
In this way, treatment effeciency can be improved.
In addition, in this embodiment, it should be noted that it prevents from waiting for that reconstructed image may include the excessive content of scene,
If waiting for that the quantity of reconstructed image is excessive using what benchmark image executed image reconstruction process, the excessive nothing of scene may result in
Method is connected, it is therefore desirable to using what benchmark image executed image reconstruction process to wait for that the quantity of reconstructed image limits.
When specific implementation, if from current temporary position, the number of the consecutive image of the second default frame number is obtained
Reach preset times, then no longer reacquires the consecutive image of the second default frame number.Abandon the first default frame to current location
The picture quality enhancement of several consecutive images operates.Also, at this point, the image before from current temporary position has been carried out
Detection, next round operation It is not necessary to carry out similar detection and reconstruct again, then can also directly by current temporary position (such as
Image 9 in aforementioned exemplary) as new current location, and S301 is returned, to improve treatment effeciency.By above-described embodiment,
Do not including the reference map for meeting preset quality condition from the consecutive image that the current location of video obtains the first default frame number
When picture, will continue to the consecutive image for obtaining the second default frame number backward, and detect in the consecutive image of the second default frame number whether
Benchmark image including meeting preset quality condition until device finds the base condition for meeting preset quality, is realized to from working as
Front position is risen, and the reconstruct for waiting for reconstructed image in all consecutive images of acquisition in addition to benchmark image was avoided because of working as from video
It does not include caused frequency when meeting the benchmark image of preset quality condition in the consecutive image of the first default frame number of front position acquisition
The problem of picture quality can not enhance.
Shown in Figure 4, the embodiment of the present application four also provides a kind of video picture quality enhancement method, and this method includes：
S401：From the initial position of pending video or video clip, cycle executes described in the embodiment of the present application
Video image quality Enhancement Method, until pending video or video clip are disposed.
In video image quality Enhancement Method provided by the embodiments of the present application, image quality carries out quality enhancing in video
When, it is according to video image quality Enhancement Method provided by the embodiments of the present application, from pending video or video clip
Initial position rise, the image that cannot be satisfied preset quality condition to quality in video or video clip successively is reconstructed,
Until by pending video, either video clip is disposed and enables to figure second-rate in video or video clip
As more being influenced after reconstitution by the higher image of quality so that the image quality of video has better enhancing effect.
Based on same inventive concept, regard corresponding with video image quality Enhancement Method is additionally provided in the embodiment of the present application
Frequency picture quality intensifier, the principle solved the problems, such as due to the device in the embodiment of the present application with the embodiment of the present application is above-mentioned regards
Frequency method for enhancing image quality is similar, therefore the implementation of device may refer to the implementation of method, and overlaps will not be repeated.
Shown in Figure 5, the embodiment of the present application five also provides a kind of video image quality intensifier and includes：
Acquiring unit 51, the consecutive image for obtaining the first default frame number from video current location；
Whether detection unit 52 includes meeting preset quality condition in the consecutive image for detecting the first default frame number
Benchmark image；
Reconfiguration unit 53, for executing：If there are benchmark image, benchmark will be removed in the consecutive image of the first default frame number
Image outside image, which is used as, waits for reconstructed image, and treats reconstructed image based on benchmark image and execute image reconstruction process.
When the embodiment of the present application carries out quality enhancing to the image in video, first have to obtain first from current location
The consecutive image of default frame number detects whether to include the benchmark for meeting preset quality condition then from the consecutive image obtained
Image.If there are benchmark image, it is based on the benchmark image, image weight is carried out to other images in acquired consecutive image
Structure processing, the lower image of quality can be also reconstructed based on quality higher image, thus for quality itself compared with
For the image of difference, can largely be influenced by the higher image of quality after reconstruct, promotion that can be more this
The effect of the quality of parts of images, picture quality enhancement is more preferable, can reach use demand.
Optionally, reconfiguration unit 53 is used to execute image reconstruction process by following step：
By benchmark image and wait for that reconstructed image is input to convolutional neural networks, on the basis of image zooming-out fisrt feature reflect
It penetrates, and to wait for that reconstructed image extracts second feature mapping；
Weighted overlap-add procedure is carried out to fisrt feature mapping and second feature mapping, obtains reconstruct Feature Mapping；
Reconstructed image is obtained based on reconstruct Feature Mapping, reconstructed image is waited for using reconstructed image replacement.
Optionally, reconfiguration unit 53 is used to map fisrt feature by following step and second feature mapping is weighted
Overlap-add procedure：
According to preset specific gravity factor, processing, and first are weighted to fisrt feature mapping and second feature mapping
The specific gravity factor of Feature Mapping is more than the specific gravity factor of second feature mapping；
It the fisrt feature mapping that will be weighted that treated and is weighted that treated second feature is mapped into
Row superposition.
Optionally, further include：Training unit 54, for using the training image set for meeting preset quality condition to convolution
Neural network is trained.
Optionally, first movement unit 55, for reconfiguration unit 53 treat reconstructed image execute image reconstruction process it
Afterwards, current location movement third is preset to the position after frame number, the new current location as video；It is small that third presets frame number
In or equal to the first default frame number.
Optionally, further include：Unit 56 is reacquired, if not including in the consecutive image for executing the first default frame number
Meet the benchmark image of preset quality condition, continues the consecutive image for obtaining the second default frame number backward, until it is pre- to find satisfaction
If the benchmark image of quality requirements；
Further include the second reconfiguration unit 57, for reference map will to be removed from all consecutive images acquired current location
Image as outside treats reconstructed image based on benchmark image and executes image reconstruction process as reconstructed image is waited for.
Optionally, unit 56 is reacquired, is specifically used for through the continuous company for obtaining the second default frame number backward of following step
Continuous image：
With the next frame image of the consecutive image obtained be it is interim obtain position, from obtaining position temporarily, obtain the
The consecutive image of two default frame numbers；Second default frame number is less than or equal to the first default frame number；
Further include the second mobile unit 58, is executed for treating reconstructed image based on benchmark image in the second reconfiguration unit 57
After image reconstruction process, using the interim position that obtains as the new current location of video.
Optionally, detection unit 52 is specifically used for：Based on the first default frame number of non-reference picture method for evaluating quality pair
The quality of consecutive image is evaluated, and the benchmark image for meeting preset quality condition is found；Or
The consecutive image of first default frame number is input to advance trained two-value grader,
Judge whether to include meeting to preset according to the classification results of the consecutive image of the first default frame number of two-value grader pair
The benchmark image of quality requirements.
Corresponding to the video image quality Enhancement Method in Fig. 1, the embodiment of the present application six additionally provides a kind of computer and sets
It is standby, as shown in fig. 6, the equipment includes memory 1000, processor 2000 and is stored on the memory 1000 and can be at this
The computer program run on reason device 2000, wherein above-mentioned processor 2000 realizes above-mentioned regard when executing above computer program
The step of frequency method for enhancing image quality.
Specifically, above-mentioned memory 1000 and processor 2000 can be general memory and processor, not do here
It is specific to limit, when the computer program of 2000 run memory 1000 of processor storage, it is able to carry out above-mentioned video image matter
Enhancement Method is measured, it is caused for quality to solve to carry out picture quality enhancement processing to present image based on current image date
Poor present image does not have larger increased quality still yet, is unable to reach use demand after having carried out picture quality enhancement
The problem of, and then reach and the lower image of other quality is reconstructed based on the higher image of quality in video, after reconstruct
Image can largely be influenced by the higher image of quality, can be more promotion this second-rate image matter
Amount, the effect of picture quality enhancement is more preferable, can achieve the effect that use demand.
Corresponding to the video image quality Enhancement Method in Fig. 1, the embodiment of the present application seven additionally provides a kind of computer can
Storage medium is read, computer program is stored on the computer readable storage medium, when which is run by processor
The step of executing above-mentioned video image quality Enhancement Method.
Specifically, which can be general storage medium, such as mobile disk, hard disk, on the storage medium
Computer program when being run, above-mentioned video image quality Enhancement Method is able to carry out, to solve to be based on present image number
According to present image carry out picture quality enhancement processing, it is caused for second-rate present image after having carried out picture quality enhancement,
Not the problem of not having larger increased quality still, being unable to reach use demand yet, and then reach higher based on quality in video
Image the lower image of other quality is reconstructed, the image after reconstruct can be largely by the higher figure of quality
The influence of picture, can be more promotion this second-rate image quality, the effect of picture quality enhancement is more preferable, can reach use
The effect of demand.
A kind of video image quality Enhancement Method, device and the video picture quality enhancement method that the embodiment of the present application is provided
Computer program product, including store the computer readable storage medium of program code, the finger that said program code includes
It enables and can be used for executing the method described in previous methods embodiment, specific implementation can be found in embodiment of the method, and details are not described herein.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description
It with the specific work process of device, can refer to corresponding processes in the foregoing method embodiment, details are not described herein.
It, can be with if the function is realized in the form of SFU software functional unit and when sold or used as an independent product
It is stored in a computer read/write memory medium.Based on this understanding, the technical solution of the application is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be expressed in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be
People's computer, server or network equipment etc.) execute each embodiment the method for the application all or part of step.
And storage medium above-mentioned includes：USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
The various media that can store program code such as reservoir (RAM, Random Access Memory), magnetic disc or CD.
The above, the only specific implementation mode of the application, but the protection domain of the application is not limited thereto, it is any
Those familiar with the art can easily think of the change or the replacement in the technical scope that the application discloses, and should all contain
It covers within the protection domain of the application.Therefore, the protection domain of the application shall be subject to the protection scope of the claim.
Claims (10)
1. a kind of video image quality Enhancement Method, which is characterized in that including：
The consecutive image of the first default frame number is obtained from video current location；
Whether detect in the consecutive image of the described first default frame number includes the benchmark image for meeting preset quality condition；
If there are the benchmark image, using the image in the consecutive image of the described first default frame number in addition to benchmark image as
It waits for reconstructed image, waits for that reconstructed image executes image reconstruction process to described based on the benchmark image.
2. according to the method described in claim 1, it is characterised in that it includes：Described image reconstruction processing includes：
By the benchmark image and it is described wait for that reconstructed image is input to convolutional neural networks, for the benchmark image extract first
Feature Mapping, and wait for that reconstructed image extracts second feature mapping to be described；
Weighted overlap-add procedure is carried out to fisrt feature mapping and second feature mapping, obtains reconstruct Feature Mapping；
Reconstructed image is obtained based on the reconstruct Feature Mapping, reconstructed image is waited for using described in reconstructed image replacement.
3. according to the method described in claim 2, it is characterized in that, described to fisrt feature mapping and the second feature
Mapping carries out weighted overlap-add procedure, specifically includes：
According to preset specific gravity factor, processing is weighted to fisrt feature mapping and second feature mapping, and
The specific gravity factor of the fisrt feature mapping is more than the specific gravity factor of second feature mapping；
It the fisrt feature mapping that will be weighted that treated and is weighted that treated second feature mapping is folded
Add.
4. according to the method described in claim 2, it is characterized in that, this method further includes：
The convolutional neural networks are trained using the training image set for meeting preset quality condition.
5. according to the method described in claim 1, it is characterized in that, to it is described wait for reconstructed image execute image reconstruction process it
Afterwards, further include：Current location movement third is preset to the position after frame number, the new current location as the video；Institute
It states third and presets frame number less than or equal to the described first default frame number.
6. according to any methods of claim 1-4, which is characterized in that this method further includes：If the first default frame
Do not include the benchmark image for meeting preset quality condition in several consecutive images, continues to obtain the continuous of the second default frame number backward
Image, until find the benchmark image for meeting the preset quality condition, and will be acquired all continuous from current location
Image in image in addition to benchmark image waits for reconstructed image execution figure based on the benchmark image as reconstructed image is waited for described
As reconstruction processing.
7. according to the method described in claim 6, it is characterized in that, continue backward obtain the second default frame number consecutive image,
Including：
With the next frame image of the consecutive image obtained be it is interim obtain position, from the interim acquisition position, obtain the
The consecutive image of two default frame numbers；The second default frame number is less than or equal to the described first default frame number；And
To it is described wait for reconstructed image execute image reconstruction process after, further include：Using the interim acquisition position as described in
The new current location of video.
8. according to any methods of claim 1-5, which is characterized in that described to detect the continuous of the first default frame number
Whether include the benchmark image for meeting preset quality condition in image, specifically includes：
The quality of the consecutive image of the described first default frame number is evaluated based on non-reference picture method for evaluating quality, is found
Meet the benchmark image of preset quality condition；Or
The consecutive image of described first default frame number is input to advance trained two-value grader,
Including meeting to preset is judged whether to the classification results of the consecutive image of the described first default frame number according to two-value grader
The benchmark image of quality requirements.
9. a kind of video picture quality enhancement method, which is characterized in that including：
From the initial position of pending video or video clip, cycle perform claim requires any video figures of 1-8
As quality enhancement method, until pending video or video clip are disposed.
10. a kind of video image quality intensifier, which is characterized in that including：
Acquiring unit, the consecutive image for obtaining the first default frame number from video current location；
Whether detection unit includes the base for meeting preset quality condition in the consecutive image for detecting the described first default frame number
Quasi- image；
Reconfiguration unit, for executing：If there are the benchmark image, base will be removed in the consecutive image of the described first default frame number
Image outside quasi- image, which is used as, waits for reconstructed image, and waits for that reconstructed image executes at image reconstruction to described based on the benchmark image
Reason.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810403772.8A CN108596855A (en) | 2018-04-28 | 2018-04-28 | A kind of video image quality Enhancement Method, device and video picture quality enhancement method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810403772.8A CN108596855A (en) | 2018-04-28 | 2018-04-28 | A kind of video image quality Enhancement Method, device and video picture quality enhancement method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108596855A true CN108596855A (en) | 2018-09-28 |
Family
ID=63619365
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810403772.8A Pending CN108596855A (en) | 2018-04-28 | 2018-04-28 | A kind of video image quality Enhancement Method, device and video picture quality enhancement method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108596855A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109862389A (en) * | 2018-11-20 | 2019-06-07 | 北京奇艺世纪科技有限公司 | A kind of method for processing video frequency, device, server and storage medium |
| CN109993737A (en) * | 2019-03-29 | 2019-07-09 | 联想(北京)有限公司 | A kind of processing method, equipment and computer readable storage medium |
| CN110099280A (en) * | 2019-05-24 | 2019-08-06 | 浙江大学 | A kind of video service quality Enhancement Method under wireless self-organization network Bandwidth-Constrained |
| CN110956768A (en) * | 2019-12-05 | 2020-04-03 | 重庆电子工程职业学院 | Automatic anti-theft device of intelligence house |
| CN112261383A (en) * | 2020-10-27 | 2021-01-22 | 广西壮族自治区水产科学研究院 | Tilapia mossambica culture water real-time monitoring and early warning integration method and integration system |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1568011A (en) * | 2003-06-20 | 2005-01-19 | 松下电器产业株式会社 | A method of inter-frame reinforcement based on Moving Joint Picture Experts Group |
| CN105844630A (en) * | 2016-03-21 | 2016-08-10 | 西安电子科技大学 | Binocular visual image super-resolution fusion de-noising method |
| WO2016132152A1 (en) * | 2015-02-19 | 2016-08-25 | Magic Pony Technology Limited | Interpolating visual data |
| CN105931210A (en) * | 2016-04-15 | 2016-09-07 | 中国航空工业集团公司洛阳电光设备研究所 | High-resolution image reconstruction method |
| CN106251365A (en) * | 2016-07-22 | 2016-12-21 | 北京邮电大学 | Many exposure video fusion method and device |
| CN106550244A (en) * | 2015-09-16 | 2017-03-29 | 广州市动景计算机科技有限公司 | The picture quality enhancement method and device of video image |
| CN106709875A (en) * | 2016-12-30 | 2017-05-24 | 北京工业大学 | Compressed low-resolution image restoration method based on combined deep network |
| CN107403413A (en) * | 2017-04-14 | 2017-11-28 | 杭州当虹科技有限公司 | A kind of video multiframe denoising and Enhancement Method |
| CN107833184A (en) * | 2017-10-12 | 2018-03-23 | 北京大学深圳研究生院 | A kind of image enchancing method for merging framework again based on more exposure generations |
-
2018
- 2018-04-28 CN CN201810403772.8A patent/CN108596855A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1568011A (en) * | 2003-06-20 | 2005-01-19 | 松下电器产业株式会社 | A method of inter-frame reinforcement based on Moving Joint Picture Experts Group |
| WO2016132152A1 (en) * | 2015-02-19 | 2016-08-25 | Magic Pony Technology Limited | Interpolating visual data |
| CN106550244A (en) * | 2015-09-16 | 2017-03-29 | 广州市动景计算机科技有限公司 | The picture quality enhancement method and device of video image |
| CN105844630A (en) * | 2016-03-21 | 2016-08-10 | 西安电子科技大学 | Binocular visual image super-resolution fusion de-noising method |
| CN105931210A (en) * | 2016-04-15 | 2016-09-07 | 中国航空工业集团公司洛阳电光设备研究所 | High-resolution image reconstruction method |
| CN106251365A (en) * | 2016-07-22 | 2016-12-21 | 北京邮电大学 | Many exposure video fusion method and device |
| CN106709875A (en) * | 2016-12-30 | 2017-05-24 | 北京工业大学 | Compressed low-resolution image restoration method based on combined deep network |
| CN107403413A (en) * | 2017-04-14 | 2017-11-28 | 杭州当虹科技有限公司 | A kind of video multiframe denoising and Enhancement Method |
| CN107833184A (en) * | 2017-10-12 | 2018-03-23 | 北京大学深圳研究生院 | A kind of image enchancing method for merging framework again based on more exposure generations |
Non-Patent Citations (2)
| Title |
|---|
| REN YANG ET AL.: "Multi-Frame Quality Enhancement", 《ARXIV》 * |
| 徐岩 等: "基于多特征融合的卷积神经网络图像去雾算法", 《激光与光电子学进展》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109862389A (en) * | 2018-11-20 | 2019-06-07 | 北京奇艺世纪科技有限公司 | A kind of method for processing video frequency, device, server and storage medium |
| CN109862389B (en) * | 2018-11-20 | 2021-08-20 | 北京奇艺世纪科技有限公司 | Video processing method, device, server and storage medium |
| CN109993737A (en) * | 2019-03-29 | 2019-07-09 | 联想(北京)有限公司 | A kind of processing method, equipment and computer readable storage medium |
| CN110099280A (en) * | 2019-05-24 | 2019-08-06 | 浙江大学 | A kind of video service quality Enhancement Method under wireless self-organization network Bandwidth-Constrained |
| CN110099280B (en) * | 2019-05-24 | 2020-05-08 | 浙江大学 | Video service quality enhancement method under limitation of wireless self-organizing network bandwidth |
| CN110956768A (en) * | 2019-12-05 | 2020-04-03 | 重庆电子工程职业学院 | Automatic anti-theft device of intelligence house |
| CN112261383A (en) * | 2020-10-27 | 2021-01-22 | 广西壮族自治区水产科学研究院 | Tilapia mossambica culture water real-time monitoring and early warning integration method and integration system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108596855A (en) | A kind of video image quality Enhancement Method, device and video picture quality enhancement method | |
| Cavigelli et al. | CAS-CNN: A deep convolutional neural network for image compression artifact suppression | |
| Ma et al. | Image retargeting quality assessment: A study of subjective scores and objective metrics | |
| RU2546616C2 (en) | Image compression system and method | |
| Mainberger et al. | Edge-based compression of cartoon-like images with homogeneous diffusion | |
| CN109325928A (en) | A kind of image rebuilding method, device and equipment | |
| CN108805840A (en) | Method, apparatus, terminal and the computer readable storage medium of image denoising | |
| JP2003523587A (en) | Visual attention system | |
| CN103856829B (en) | A kind of insertion of video digital watermark, extracting method and device | |
| CN101690226A (en) | Statistic image improving method, image encoding method, and image decoding method | |
| Wang et al. | Image representation using block pattern models and its image processing applications | |
| WO2016033725A1 (en) | Block segmentation mode processing method in video coding and relevant apparatus | |
| CN105979283A (en) | Video transcoding method and device | |
| TW463133B (en) | Color image processing method and apparatus thereof | |
| Huang et al. | Anatomical‐functional image fusion based on deep convolution neural networks in local Laplacian pyramid domain | |
| JP2022508245A (en) | Block-based picture fusion for contextual partitioning and processing | |
| Fu | Color image quality measures and retrieval | |
| Cao et al. | Oodhdr-codec: Out-of-distribution generalization for hdr image compression | |
| CN116778539A (en) | Human face image super-resolution network model based on attention mechanism and processing method | |
| He et al. | High-fidelity data embedding for image annotation | |
| Sun et al. | Reduced-reference image quality assessment through SIFT intensity ratio | |
| US8908986B1 (en) | Systems and methods for selecting ink colors | |
| Prabukumar et al. | Compressed Domain Contrast and Brightness Improvement Algorithm for Color Images through Contrast Measuring and Mapping of DWT Coefficients | |
| Duan et al. | A natural image compression approach based on independent component analysis and visual saliency detection | |
| US10405002B2 (en) | Low complexity perceptual visual quality evaluation for JPEG2000 compressed streams |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 101-8, 1st floor, building 31, area 1, 188 South Fourth Ring Road West, Fengtai District, Beijing Applicant after: Guoxin Youyi Data Co., Ltd Address before: 100070, No. 188, building 31, headquarters square, South Fourth Ring Road West, Fengtai District, Beijing Applicant before: SIC YOUE DATA Co.,Ltd. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180928 |