CN107492085A - Stereo image quality evaluation method based on dual-tree complex wavelet transform - Google Patents
Stereo image quality evaluation method based on dual-tree complex wavelet transform Download PDFInfo
- Publication number
- CN107492085A CN107492085A CN201710597141.XA CN201710597141A CN107492085A CN 107492085 A CN107492085 A CN 107492085A CN 201710597141 A CN201710597141 A CN 201710597141A CN 107492085 A CN107492085 A CN 107492085A
- Authority
- CN
- China
- Prior art keywords
- image
- original
- wavelet
- stereo
- picture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000013441 quality evaluation Methods 0.000 title claims abstract description 29
- 238000000034 method Methods 0.000 title claims abstract description 28
- 239000002131 composite material Substances 0.000 claims abstract description 26
- 238000010586 diagram Methods 0.000 claims abstract description 26
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 10
- 238000004364 calculation method Methods 0.000 claims abstract description 7
- 239000000284 extract Substances 0.000 claims abstract description 3
- 238000011156 evaluation Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 230000004927 fusion Effects 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 3
- 238000012360 testing method Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000002156 mixing Methods 0.000 description 2
- 238000001303 quality assessment method Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000005562 fading Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 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
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
- Image Analysis (AREA)
Abstract
Stereo image quality evaluation method is referred to based on dual-tree complex wavelet transform entirely the present invention relates to a kind of, step is as follows：1) carries out left image respectively for original three-dimensional image and distortion stereo-picture and right image merges；2) does dual-tree complex wavelet transform for the original three-dimensional image and left image of distortion stereo-picture, right image and composite diagram；3) extracts contrast, structure and brightness from original graph and each wavelet sub-band of distortion map；4) makes energy calculation to each wavelet sub-band of original three-dimensional image, gain control method is used to make weight with each sub-belt energy after left image, right image and composite diagram the progress wavelet decomposition of original three-dimensional image, with reference to features such as brightness, contrast and the structures of original image and each wavelet sub-band of distorted image, stereo image quality evaluation is carried out, obtains final quality evaluation fraction.
Description
Technical field
The invention belongs to the objective evaluation system of image processing field, specifically stereo image quality, is related to using double trees
The objective image evaluation method of complex wavelet transform and stereo-picture composite diagram.
Background technology
With the arrival in multimedia messages epoch, people as the data signal active influence of representative and is changed using image and video
Live and work mode.Due to the fast development of imaging technique and display device, the popularization of stereo-picture, it is greatly enhanced
The visual quality experience of user.But stereo-picture can produce during acquisition, pretreatment, coding, transmission, decoding etc.
Degradation problems, therefore stereo image quality evaluation turns into an important research topic.Stereo image quality evaluation now
In method, most straightforward approach is exactly that left images are commented respectively using the statistical nature of plane picture quality evaluating method
Valency, finally prediction obtains the quality of stereo-picture, but such method does not have the visually-perceptible mechanism for considering that the mankind are complicated, because
This evaluation poor-performing.Composite diagram [1] incorporates the parallax information of visually-perceptible so that stereo image quality is evaluated and subjective assessment
Uniformity is higher.Therefore, the present invention proposes the traditional plane picture quality evaluating method of application in left image, right image and synthesis
Feature is extracted on each subband of small echo of figure, carries out stereo image quality evaluation.
[1]Chen M J,Su C C,Kwon D K,et al.Full-reference quality assessment
of stereopairs accounting for rivalry[J].Signal Processing:Image
Communication,2013,28(9):1143-1155.
[2]Wang Z,Bovik A C,Sheikh H R,et al.Image quality assessment:from
error visibility to structural similarity[J].IEEE transactions on image
processing,2004,13(4):600-612.
The content of the invention
It is an object of the invention to for three-dimensional distorted image quality evaluation problem, it is preferably vertical to propose that one kind can obtain
The complete of body image quality evaluation effect refers to stereo image quality evaluation method.The present invention is on original three-dimensional image left image, the right side
Wavelet decomposition is carried out on image and composite diagram and distortion stereo-picture left image, right image and composite diagram, in each wavelet sub-band
Upper extraction is than degree, structure and brightness, with each wavelet sub-band energy of original three-dimensional image left image, right image and composite diagram
Make weight fusion character pair, the method for carrying out stereo image quality evaluation.Technical scheme is as follows：
It is a kind of that stereo image quality evaluation method is referred to based on dual-tree complex wavelet transform entirely, utilize the left figure of stereo-picture
Picture, right image and each wavelet sub-band energy of composite diagram three weight each wavelet sub-band feature and carry out quality to distortion stereo-picture
Evaluation, step are as follows：
1) carries out left image respectively for original three-dimensional image and distortion stereo-picture and right image merges
For original three-dimensional image and distortion stereo-picture, it is poor that left view is calculated by left image and right image, uses left image
Left disparity map is added to obtain with left view difference, then weight is calculated to synthesize composite diagram by left image and left disparity map, passes through normalization
Gabor filtered energies respond to obtain image weights；
2) it is multiple small to do double trees for the original three-dimensional image and left image of distortion stereo-picture, right image and composite diagram by
Wave conversion
Each image is resolved into 3 layers, every layer of 6 direction, i.e.,：They distinguish table
Show n-th layer ± 15 °, ± 45 °, the wavelet coefficient of ± 75 ° of subbands, n=1,2,3；Obtain 18 wavelet sub-bands；
3) extracts contrast, structure and brightness from original graph and each wavelet sub-band of distortion map
By the contrast of original three-dimensional image and the left image of distortion stereo-picture, right image and each wavelet sub-band of composite diagram
Degree, structure and brightness are as feature extraction；
4) makes energy calculation to each wavelet sub-band of original three-dimensional image, and Fusion Features
Each wavelet sub-band of original three-dimensional image is made energy calculation, uses gain control method with original stereo figure
The left image of picture, right image and composite diagram carry out each sub-belt energy after wavelet decomposition and make weight, with reference to original image and distortion
The features such as brightness, contrast and the structure of each wavelet sub-band of image, stereo image quality evaluation is carried out, final quality is obtained and comments
Valency fraction.
The present invention carries out wavelet decomposition in the left image, right image and composite diagram of original three-dimensional image, after wavelet decomposition
Each sub-belt energy as gain control weight, brightness, contrast and the architectural feature of each image are merged, to distortion stereo-picture
Quality evaluation, obtain quality evaluation fraction.The present invention and subjective picture quality evaluation uniformity are strong, and performance is better than what be presently, there are
Most of stereo image quality evaluation algorithms.
Brief description of the drawings
The flow chart of Fig. 1 present invention
Tri- layers of dual-tree complex wavelet transforms of Fig. 2
Fig. 3 uniformity
Embodiment
The present invention proposes a kind of based on extracting wavelet decomposition in the left image, right image and composite diagram of original three-dimensional image
The full reference image quality appraisement method of weight that is controlled as gain of each sub-belt energy.To make technical scheme more
Add clear, with reference to figure 1, of the invention comprises the following steps that.
1. left images merge
After different degrees of alteration occurs in stereo-picture, due to the presence of parallax, the subjective quality of image can not only lead to
The average of left images quality is crossed to obtain, the stereo-picture for merging parallax information is different from plane picture, left images fusion
Obtain composite diagram and mainly consider parallax information, therefore the present invention, to being handled as follows to left and right figure, left images carry out Gabor filters
Ripple, the multiple Gabor filtering of two dimension are defined as follows：
Wherein R1=xcos θ+ysin θ, R2=-sin θ+ycon θ.σx,σyIt is standard deviation, ζx,ζyIt is spatial frequency, θ is filter
Ripple direction.Energy response value of the left images in all yardsticks and direction is respectively GEL, GER。
The calculating of left and right weight is obtained by normalizing Gabor filtered energies response assignment, is defined as：
Composite diagram is defined as by the present invention：
C (x, y)=wL(x,y)·IL(x,y)+wR(x+d,y)·IR(x+d,y)
(4)
Wherein, C represents composite diagram, ILAnd IRLeft images are represented, d is parallax, wLAnd wRFor left and right weight, for solid
The quality evaluation of image, present invention utilizes the parallax information of left images to synthesize composite diagram, is prepared for extraction feature.
2. dual-tree complex wavelet transform
The present invention carries out wavelet decomposition on left image, right image and composite diagram, because the peak value of wavelet coefficient can be anti-
Reflect the definition of image.Dual-tree complex wavelet transform is developed by traditional wavelet, and traditional wavelet transformation is broken down into water
Flat, vertical and 3 directions of tilted direction detailed information, dual-tree complex wavelet transform inherit the good characteristic of traditional wavelet,
The directivity information of image can also be described more, Complex Wavelet Transform is realized by real DWT, by real part and
Imaginary part is separated, and the wavelet conversion coefficient of real and imaginary parts is obtained by two groups of parallel real filter groups.The present invention is made
Three layers of dual-tree complex wavelet transform such as Fig. 2.
Each image is resolved into 3 layers by the present invention, every layer of 6 direction, i.e.,：They
± 15 ° of expression n-th layer respectively, ± 45 °, the wavelet coefficient of ± 75 ° of subbands, n=1,2,3.Each subband of small echo is carried out
Energy balane, as the final weight for calculating mass fraction
3. character representation
These three aspects of contrast, structure and brightness all directly affect picture quality, between them again independently of each other.This hair
It is bright using contrast, structure and the brightness of the left images of stereo-picture and each wavelet sub-band of composite diagram as feature extraction.
Picture contrast, structure and brightness are extracted separately below：
1) contrast is defined as the standard deviation sigma of signalx, the similarity of contrast is defined as：
C1It is the constant of a very little.
To two signals x and y, brightness is defined as base average, i.e.,：
In formula, N is the length of signal.
2) structural similarity is defined as：
3) similarity of brightness is defined as：
In formula, C2It is the constant of a very little, ensuresWith 0 very close to when numerical value stablize it is qualitative.
Image structure similarity (SSIM) [2] is defined as：
SSIM (x, y)=[L (x, y)]α[C(x,y)]β[F(x,y)]γ (9)
In formula, α, beta, gamma is the number more than 0, and the size by adjusting this 3 parameters can change 3 components in quality
Importance in fraction.
4. Fusion Features
The present invention uses gain control method to carry out each sub-belt energy work after wavelet decomposition with left and right image and composite diagram
Weight, with reference to features such as the brightness of each wavelet sub-band of image, contrast and architectural features, carry out stereo image quality evaluation.Its
Effect is that simple linear combination mode is not reached.Formula is as follows：
WhereinEL、ERIt is similar,Represent that left image carries out i-th layer of j-th of direction after wavelet decomposition
Energy：
η,Respectively Li,jThe real and imaginary parts in (i-th layer of j-th of direction of left image),Formula it is similar.The structural similarity of original left image and distortion right image is represented,
Q is final mass fraction.
5. database selects
Two open test storehouses of present invention selection, they are the asymmetric stereo-picture test libraries that LIVE laboratories provide
LIVE-3D II and symmetrical stereo-picture test library LIVE-3D I.In LIVE 3D rendering quality evaluations storehouse II, totally 360 width lose
True stereo-picture and 8 width original images, the storehouse includes symmetrical and asymmetric 2 kinds of distortions, also comprising 5 kinds of type of distortion:JPEG is pressed
Contracting, JPEG2000 (JP2K), Gaussian Blur (Gaussian blur, GBLUR), white noise (white noise, WN) and fast
Weak (fast fading, FF), and provide the subjective scoring difference and parallax value of every group of distortion stereo-picture.Scheme in LIVE3D
As in quality evaluation storehouse I, totally 365 width distortion stereo-pictures and 20 width original images, are that left and right distortion level identical is symmetrical
Distortion, include 5 kinds of type of distortion:JPEG,JP2K,GBLUR,WN,FF.
For prove image prediction objective quality scores that the inventive method obtains and subjective quality scores have it is very high consistent
Property, prediction objective quality scores can accurately reflect the quality of image, the inventive method is surveyed in symmetrical and asymmetric stereo-picture
Tested on examination storehouse LIVE-3D II and LIVE-3D I, take 3 measurement Objective image quality evaluation algorithms commonly used in the world
Index evaluation the inventive method performance, 3 indexs are respectively Spearman sequence coefficient correlation (Spearman rank-
Order correlation coefficient, SRCC), Pearson's linearly dependent coefficient (Pearson linear
Correlation coefficient, PLCC) and root-mean-square error (Root Mean Squared Error, RMSE), wherein,
PLCC and RMSE indexs weigh the forecasting accuracy of objective algorithm, and SRCC indexs weigh the prediction monotonicity of objective algorithm.PLCC
It is smaller closer to 1, RMSE value with SRCC value, illustrate that algorithm performance is better, prediction objective quality scores and subjective quality point
Several correlations is higher.
6. compare and parser performance
Present invention verification algorithm performance on stereo-picture test library LIVE-3D I and LIVE-3D II, from following contrast
In understand that the inventive method achieves good effect, the inventive method vertical stops image quality evaluating method uniformity with subjective
It is high.Table 1 represents performance of the present invention on LIVE 3D Phase I test libraries, it can be seen that the present invention is for right from table 1
Claim Gaussian Blur distortion PLCC to reach 0.944,0.937 is reached to white Gaussian noise distortion SPRCC.It can see by table 2, this hair
It is bright to reach 0.971, SPRCC for asymmetric Gaussian Blur mixing distortion PLCC on LIVE 3D Phase II test libraries and reach
0.947.Tables 1 and 2 has carried out contrast of the present invention on LIVE storehouses with other methods, as can be seen from the table, the present invention
Method and the degree of correlation of subjective assessment, order of accuarcy are significantly improved, and Fig. 3 represents the correlation of subjective and objective fraction, says
The prediction objective quality scores of bright the inventive method are high with subjective quality scores correlation.
Table 1 is in LIVE 3D Phase I to the uniformity result of calculation of 5 kinds of distortions
Table 2 is in LIVE 3D Phase II to the uniformity result of calculation of 5 kinds of mixing distortions
The inventive method has advantages below：
(1) the inventive method and subjective stereo image quality evaluation uniformity are high.
(2) different from plane picture quality evaluation, the present invention considers the distinctive parallax information of stereo-picture, proposes from conjunction
Brightness of image, structure and contrast metric are extracted on into each subband of small echo of figure, test result indicates that the inventive method performance is excellent
In the most of stereo image quality evaluation algorithms that presently, there are.
Claims (1)
1. a kind of refer to stereo image quality evaluation method entirely based on dual-tree complex wavelet transform, the left figure of stereo-picture is utilized
Picture, right image and each wavelet sub-band energy of composite diagram three weight each wavelet sub-band feature and carry out quality to distortion stereo-picture
Evaluation.Step is as follows：
1) carries out left image respectively for original three-dimensional image and distortion stereo-picture and right image merges
For original three-dimensional image and distortion stereo-picture, it is poor that left view is calculated by left image and right image, with left image and a left side
Parallax is added to obtain left disparity map, then calculates weight by left image and left disparity map to synthesize composite diagram, by normalizing Gabor filters
Wave energy responds to obtain image weights；
2) does dual-tree complex wavelet change for the original three-dimensional image and left image of distortion stereo-picture, right image and composite diagram
Change
Each image is resolved into 3 layers, every layer of 6 direction, i.e.,：They represent n-th respectively
± 15 ° of layer, ± 45 °, the wavelet coefficient of ± 75 ° of subbands, n=1,2,3；Obtain 18 wavelet sub-bands；
3) extracts contrast, structure and brightness from original graph and each wavelet sub-band of distortion map
By original three-dimensional image and the left image of distortion stereo-picture, the contrast of right image and each wavelet sub-band of composite diagram, knot
Structure and brightness are as feature extraction；
4) makes energy calculation to each wavelet sub-band of original three-dimensional image, and Fusion Features
Each wavelet sub-band of original three-dimensional image is made energy calculation, uses gain control method with original three-dimensional image
Each sub-belt energy after left image, right image and composite diagram progress wavelet decomposition makees weight, with reference to original image and distorted image
The features such as brightness, contrast and the structure of each wavelet sub-band, stereo image quality evaluation is carried out, obtain final quality evaluation point
Number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710597141.XA CN107492085B (en) | 2017-07-20 | 2017-07-20 | Stereo image quality evaluation method based on dual-tree complex wavelet transform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710597141.XA CN107492085B (en) | 2017-07-20 | 2017-07-20 | Stereo image quality evaluation method based on dual-tree complex wavelet transform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107492085A true CN107492085A (en) | 2017-12-19 |
| CN107492085B CN107492085B (en) | 2021-01-15 |
Family
ID=60643480
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710597141.XA Active CN107492085B (en) | 2017-07-20 | 2017-07-20 | Stereo image quality evaluation method based on dual-tree complex wavelet transform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107492085B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108269253A (en) * | 2018-01-11 | 2018-07-10 | 天津大学 | Stereo image quality evaluation method based on wavelet transformation and local structure feature |
| CN110363763A (en) * | 2019-07-23 | 2019-10-22 | 上饶师范学院 | Image quality evaluating method, device, electronic equipment and readable storage medium storing program for executing |
| CN112330757A (en) * | 2019-08-05 | 2021-02-05 | 复旦大学 | Complementary color wavelet measurement for evaluating color image automatic focusing definition |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103152600A (en) * | 2013-03-08 | 2013-06-12 | 天津大学 | Three-dimensional video quality evaluation method |
| CN104918039A (en) * | 2015-05-05 | 2015-09-16 | 四川九洲电器集团有限责任公司 | Image quality evaluation method and image quality evaluation system |
| CN105959684A (en) * | 2016-05-26 | 2016-09-21 | 天津大学 | Stereo image quality evaluation method based on binocular fusion |
-
2017
- 2017-07-20 CN CN201710597141.XA patent/CN107492085B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103152600A (en) * | 2013-03-08 | 2013-06-12 | 天津大学 | Three-dimensional video quality evaluation method |
| CN104918039A (en) * | 2015-05-05 | 2015-09-16 | 四川九洲电器集团有限责任公司 | Image quality evaluation method and image quality evaluation system |
| CN105959684A (en) * | 2016-05-26 | 2016-09-21 | 天津大学 | Stereo image quality evaluation method based on binocular fusion |
Non-Patent Citations (2)
| Title |
|---|
| 张达奇: "全参考图像主观质量客观评价算法研究", 《中国优秀硕士学位论文全文数据库（电子期刊）》 * |
| 熊润生: "基于小波变换的无参考立体图像质量评价", 《计算机科学》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108269253A (en) * | 2018-01-11 | 2018-07-10 | 天津大学 | Stereo image quality evaluation method based on wavelet transformation and local structure feature |
| CN110363763A (en) * | 2019-07-23 | 2019-10-22 | 上饶师范学院 | Image quality evaluating method, device, electronic equipment and readable storage medium storing program for executing |
| CN112330757A (en) * | 2019-08-05 | 2021-02-05 | 复旦大学 | Complementary color wavelet measurement for evaluating color image automatic focusing definition |
| CN112330757B (en) * | 2019-08-05 | 2022-11-29 | 复旦大学 | Complementary color wavelet measurement for evaluating color image automatic focusing definition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107492085B (en) | 2021-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108876735B (en) | Real image blind denoising method based on depth residual error network | |
| CN105959684B (en) | Stereo image quality evaluation method based on binocular fusion | |
| CN107393542B (en) | Bird species identification method based on two-channel neural network | |
| CN110969124A (en) | Two-dimensional human body posture estimation method and system based on lightweight multi-branch network | |
| CN108765414B (en) | No-reference stereo image quality evaluation method based on wavelet decomposition and natural scene statistics | |
| CN107341786A (en) | The infrared and visible light image fusion method that wavelet transformation represents with joint sparse | |
| CN107123091A (en) | A kind of near-infrared face image super-resolution reconstruction method based on deep learning | |
| CN109167996B (en) | It is a kind of based on convolutional neural networks without reference stereo image quality evaluation method | |
| Li et al. | Deep learning methods in real-time image super-resolution: a survey | |
| CN110070574B (en) | Binocular vision stereo matching method based on improved PSMAT net | |
| CN105844635B (en) | A kind of rarefaction representation depth image method for reconstructing based on structure dictionary | |
| CN101562675B (en) | No-reference image quality evaluation method based on Contourlet transform | |
| CN112288668B (en) | Infrared and visible light image fusion method based on depth unsupervised dense convolution network | |
| CN104376565B (en) | Based on discrete cosine transform and the non-reference picture quality appraisement method of rarefaction representation | |
| CN104199627B (en) | Gradable video encoding system based on multiple dimensioned online dictionary learning | |
| CN109242834A (en) | It is a kind of based on convolutional neural networks without reference stereo image quality evaluation method | |
| CN107492085A (en) | Stereo image quality evaluation method based on dual-tree complex wavelet transform | |
| CN109523513A (en) | Based on the sparse stereo image quality evaluation method for rebuilding color fusion image | |
| CN111539247A (en) | Hyper-spectrum face recognition method and device, electronic equipment and storage medium thereof | |
| CN108259893B (en) | Virtual reality video quality evaluation method based on double-current convolutional neural network | |
| CN107071423A (en) | Application process of the vision multi-channel model in stereoscopic video quality objective evaluation | |
| CN108257098A (en) | Video denoising method based on maximum posteriori decoding and three-dimensional bits matched filtering | |
| Li et al. | Recent advances and challenges in video quality assessment | |
| CN114266957A (en) | Hyperspectral image super-resolution restoration method based on multi-degradation mode data augmentation | |
| CN114926337A (en) | Single image super-resolution reconstruction method and system based on CNN and Transformer hybrid network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |