CN109544504A - Screen picture quality evaluating method based on rarefaction representation - Google Patents

Screen picture quality evaluating method based on rarefaction representation Download PDF

Info

Publication number
CN109544504A
CN109544504A CN201811203757.5A CN201811203757A CN109544504A CN 109544504 A CN109544504 A CN 109544504A CN 201811203757 A CN201811203757 A CN 201811203757A CN 109544504 A CN109544504 A CN 109544504A
Authority
CN
China
Prior art keywords
gradient
screen picture
screen
image
rarefaction representation
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
Application number
CN201811203757.5A
Other languages
Chinese (zh)
Inventor
杨嘉琛
刘佳成
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tianjin University
Original Assignee
Tianjin University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tianjin University filed Critical Tianjin University
Priority to CN201811203757.5A priority Critical patent/CN109544504A/en
Publication of CN109544504A publication Critical patent/CN109544504A/en
Pending legal-status Critical Current

Links

Classifications

    • 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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/28—Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
    • G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • 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
    • G06T2207/20081—Training; Learning
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30168—Image quality inspection

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Quality & Reliability (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Image Analysis (AREA)

Abstract

The present invention relates to a kind of screen picture quality evaluating method based on rarefaction representation, comprising the following steps: step 1: three kinds of gradient maps of calculated distortion screen picture, including absolute gradient figure, relative gradient directional diagram and relative gradient figure；Step 2: extracting the textural characteristics of screen picture, the absolute gradient figure obtained from second step extracts gradient orientation histogram feature in relative gradient directional diagram and relative gradient figure.Learn the building of dictionary；The rarefaction representation of test screen image feature vector.Step 5: the mass fraction pond stage.Screen picture quality evaluating method proposed by the invention accurately predicts the mass fraction after screen picture distortion using image second order local derviation information and rarefaction representation.

Description

Screen picture quality evaluating method based on rarefaction representation
Technical field
The invention belongs to digital image processing field, it is related to the screen picture quality evaluating method based on rarefaction representation.
Background technique
With the rapid development of internet and various electronic equipments, people not only touch natural image in daily life, There are also various screen pictures.Natural image is often to be shot by digital camera, however screen picture is by computer Or the acquisition of mobile device screenshotss.It, can be inevitably by various images in the transmission of screen picture, reception and cataloged procedure The interference of type of distortion, to influence the mankind to the visual perception of screen picture.If for example, teleconference and online cloud video It is influenced by unfavorable factors such as transmission distortion, network delays, requires to assess online real time screen picture quality, with Just service provider's dynamic adjustment source position strategy, to meet the needs of service quality.
Since the assessment technique of online screen picture quality has just emerged in recent years, it is directed to distorted screen image number at present According to library to objectively evaluate system fewer.It is different from the main natural image shot comprising natural scene and by video camera, Screen picture can be natural image and computer generates the mixing of content (such as figure, text, chart, symbol).Therefore, shield The feature and natural image of curtain image have very big difference.For example, an advertising poster image may include one or more images, With text description together insertion, etc..This kind of image often has sharp edge, has high comparison in some regions of image Degree and less color category.The algorithm of many assessment natural image quality can not be suitable for the quality evaluation of screen picture. Therefore, it is necessary to propose a kind of effective screen picture quality evaluating method by the present invention.
Summary of the invention
It is an object of the invention to establish a kind of quality evaluating method for screen picture, screen map proposed by the present invention Image quality evaluation method, based on sensitive theoretical and screen picture spy of texture information of the human visual system to screen picture itself Property, the second order for extracting distorted screen image leads feature and represents texture information and the subjective quality scores progress of its original screen image Dictionary learning and rarefaction representation, and then make more comprehensive and accurate objectively evaluate.Technical solution is as follows:
A kind of screen picture quality evaluating method based on rarefaction representation, the following steps are included:
Step 1: three kinds of gradient maps of calculated distortion screen picture, including absolute gradient figure, relative gradient directional diagram and phase To gradient map
(1) by distorted screen image gray processing, using Gauss local derviation difference operator respectively along the level of its gray level image Direction and vertical direction carry out Gaussian convolution filtering, obtain horizontal direction gradient map and vertical gradient figure, are taken absolutely It is added after value and obtains absolute gradient figure.
(2) average gradient value of horizontal direction gradient map and vertical gradient figure is calculated separately, formula is as follows:
Wherein, I (x, y) represents the pixel position of gradient map, and (p, q) is relative coordinate offset, and Ω is relative coordinate The set of offset, it is the regional area for calculating local derviation numerical value, calculates separately water outlet square using the part P=Q=3 small cube To the average gradient figure of gradient map and vertical gradient figure；
(3) using horizontally and vertically average gradient figure the arctan function of the ratio of a certain pixel can To find out the relative gradient direction of the pixel, entire relative gradient directional diagram is calculated.
(4) horizontal direction gradient map and vertical gradient figure are subtracted into its corresponding direction by pixel corresponding rule Relative gradient figure can be obtained in average gradient figure, the addition that takes absolute value.
Step 2: extracting the textural characteristics of screen picture: the absolute gradient figure obtained from second step, relative gradient directional diagram With extraction gradient orientation histogram feature in relative gradient figure: carrying out gamma correction to gradient map, then divide the image into several A 8 × 8 cell cell, every 2 × 2 cell constitutes 1 block；The gradient direction of each cell is calculated, then by gradient Direction is divided into 9 parts, and each part is 40 degree, and then the modulus value of the gradient in each part is added, obtains the part Component, thus obtain the feature of a cell, i.e., the vector of one 9 dimension.Finally all features of the cell in block are connected Get up, constitutes the feature of the block, i.e., the vector of one 36 dimension；One width distorted screen image is after coefficient is 2 down-sampling With 216 dimensional feature vectors.
Step 3: the building of study dictionary: for screen picture training set, extracting the spy of wherein each width screen picture Vector is levied, the subjective quality scores value that it is corresponded to each width training screen picture constitutes a matrix, the instruction finally constructed Practice collection dictionary H are as follows:
Wherein, G represents m × n eigenvectors matrix, and m represents feature vector dimension, and n represents the quantity of screen picture, DMOS For the subjective quality scores set of the corresponding 1 × n dimension of whole distorted screen images, f represents screen picture in training set and extracts Feature vector, dmos represents certain corresponding subjective quality scores of width distorted screen image；
Step 4: the rarefaction representation of test screen image feature vector.In screen picture test set, finds out a certain width and survey After trying the corresponding feature vector of screen picture, then find out by the dictionary H that previous step training obtains the spy of the width test screen image Rarefaction representation coefficient corresponding to vector is levied, formula is as follows:
Wherein αiRepresent the expression coefficient of the i-th width training screen picture feature vector, HiRepresent the i-th column in dictionary H, α table Show the rarefaction representation coefficient of test screen image feature vector；
Above-mentioned rarefaction representation coefficient be converted into more convenient solution without constraint linear optimization problem, formula is as follows:
Wherein, RNR dimension real number space is represented, λ is a normal number.
Step 5: the mass fraction pond stage: the mass fraction of test screen image can be according to its corresponding subjective matter Measure score and sparse coefficient α*It acquires, score convergence strategy is as follows:
Wherein, Q represents forecast quality score, and range is 0 to 100, dmosiRepresent the corresponding subjectivity of the i-th width screen picture Mass fraction.
Screen picture quality evaluating method proposed by the invention is accurate using image second order local derviation information and rarefaction representation Ground predicts the mass fraction after screen picture distortion.The screen picture preprocess method that the present invention takes is simple, has relatively strong Practicability, the test model time-consuming proposed is small, easily operated.The screen picture Objective Quality Assessment result that this method obtains There is very high consistency with subjective evaluation result, can accurately reflect the quality of screen picture.
Detailed description of the invention
The mentioned method flow diagram of Fig. 1
Specific embodiment
Screen picture quality evaluating method based on rarefaction representation of the invention, if distorted image is F, including following step It is rapid:
Step 1: the absolute gradient figure of calculated distortion screen picture.First by distorted screen image gray processing, then find out Absolute gradient value F (x, y) in pixel position (x, y):
F (x, y)=| Fh(x, y) |+| Fv(x, y) |
Wherein
I (x, y) represents the brightness layer of distorted screen image in formula,Linear convolution core is represented, | Fh(x, y) | and | Fv (x, y) | it respectively represents along absolute gradient value horizontally and vertically, Gauss local derviation filter Tγ, γ ∈'s (h, v) Calculation formula are as follows:
Wherein σ represents the parameter of Gaussian function g (x, y | σ).
Step 2: the relative gradient directional diagram and relative gradient figure of calculated distortion screen picture.In pixel position (x, y) Relative gradient direction value be respectively FRD(x, y) and FRM(x, y):
Wherein
Ω represents relative coordinate variable quantity in formula, is defined as seeking the regional area of second-order partial differential coefficient.The present invention takes 3 × 3 regional area.
Step 3: extracting the textural characteristics of screen picture.Since the second-order partial differential coefficient of screen picture characterizes its texture Structural information.So being extracted in the absolute gradient figure that the present invention is obtained from second step, relative gradient directional diagram and relative gradient figure Gradient orientation histogram feature.Gamma correction is carried out to gradient map first, then divides the image into several 8 × 8 cells Cell, every 2 × 2 cell constitute 1 block；The gradient direction of each cell is calculated, gradient direction is then divided into 9 portions Point, each part is 40 degree.Then the modulus value of the gradient in each part is added, has just obtained the component of the part.In this way The feature of a cell, i.e., the vector of one 9 dimension are just obtained.Finally all features of the cell in block are together in series, Constitute the feature of the block, i.e., the vector of one 36 dimension；One width distorted screen image has after coefficient is 2 down-sampling 216 dimensional feature vectors.
Step 4: the building of study dictionary.For screen picture training set, the spy of wherein each width screen picture is extracted Vector is levied, the subjective quality scores value that it is corresponded to each width training screen picture constitutes a matrix, the instruction finally constructed Practice collection dictionary H are as follows:
Wherein, G represents m × n eigenvectors matrix (m represents feature vector dimension, and n represents the quantity of screen picture), DMOS is the subjective quality scores set of the corresponding 1 × n dimension of whole distorted screen images.F represents screen picture in training set and mentions The feature vector of taking-up, dmos represent certain corresponding subjective quality scores of width distorted screen image.
Step 5: the rarefaction representation of test screen image feature vector.In screen picture test set, finds out a certain width and survey After trying the corresponding feature vector of screen picture, then find out by the dictionary H that previous step training obtains the spy of the width test screen image Levy rarefaction representation coefficient corresponding to vector.Formula is as follows:
Wherein αiRepresent the expression coefficient of the i-th width training screen picture feature vector, HiRepresent the i-th column in dictionary H, α table Show the rarefaction representation coefficient of test screen image feature vector.
Above-mentioned expression formula can be converted into more convenient solution without constraint linear optimization problem, formula is as follows:
Wherein, RNR dimension real number space is represented, λ is a normal number.
Step 6: the mass fraction pond stage.The mass fraction of test screen image can be according to its corresponding subjective matter Measure score and sparse coefficient α*It acquires.Score convergence strategy is as follows:
Wherein, Q represents forecast quality score, and range is 0 to 100, dmosiRepresent the corresponding subjectivity of the i-th width screen picture Mass fraction.

Claims (1)

1. a kind of screen picture quality evaluating method based on rarefaction representation, the following steps are included:
Step 1: three kinds of gradient maps of calculated distortion screen picture, including absolute gradient figure, relative gradient directional diagram and relatively ladder Degree figure
(1) by distorted screen image gray processing, using Gauss local derviation difference operator respectively along the horizontal direction of its gray level image Gaussian convolution filtering is carried out with vertical direction, horizontal direction gradient map and vertical gradient figure are obtained, after being taken absolute value Addition obtains absolute gradient figure；
(2) average gradient value of horizontal direction gradient map and vertical gradient figure is calculated separately, formula is as follows:
Wherein, I (x, y) represents the pixel position of gradient map, and (p, q) is relative coordinate offset, and Ω is relative coordinate offset The set of amount, it is the regional area for calculating local derviation numerical value, calculates separately out horizontal direction ladder using the part P=Q=3 small cube The average gradient figure of degree figure and vertical gradient figure；
(3) using average gradient figure horizontally and vertically the arctan function of the ratio of a certain pixel can be in the hope of The relative gradient direction of the pixel out calculates entire relative gradient directional diagram；
(4) horizontal direction gradient map and vertical gradient figure are subtracted into being averaged for its corresponding direction by pixel corresponding rule Relative gradient figure can be obtained in gradient map, the addition that takes absolute value；
Step 2: extracting the textural characteristics of screen picture, the absolute gradient figure obtained from second step, relative gradient directional diagram and phase To in gradient map extract gradient orientation histogram feature: to gradient map carry out gamma correction, then divide the image into several 8 × 8 cell cell, every 2 × 2 cell constitutes 1 block；The gradient direction of each cell is calculated, then by gradient direction It is divided into 9 parts, each part is 40 degree, then the modulus value of the gradient in each part is added, obtains the component of the part, To obtain the feature of a cell, i.e., the vector of one 9 dimension；Finally all features of the cell in block are together in series, Constitute the feature of the block, i.e., the vector of one 36 dimension；One width distorted screen image has after coefficient is 2 down-sampling 216 dimensional feature vectors；
Step 3: the building of study dictionary: to screen picture training set, the feature vector of wherein each width screen picture is extracted, The subjective quality scores value that it is corresponded to each width training screen picture constitutes a matrix, the training set dictionary finally constructed H are as follows:
Wherein, G represents m × n eigenvectors matrix, and m represents feature vector dimension, and n represents the quantity of screen picture, and DMOS is complete The subjective quality scores set of the corresponding 1 × n dimension of portion's distorted screen image, f represent the spy that screen picture in training set extracts Vector is levied, dmos represents certain corresponding subjective quality scores of width distorted screen image；
Step 4: the rarefaction representation of test screen image feature vector；In screen picture test set, a certain width test panel is found out After the corresponding feature vector of curtain image, then from the dictionary H that previous step training obtains find out the feature of the width test screen image to The corresponding rarefaction representation coefficient of amount, formula are as follows:
Wherein αiRepresent the expression coefficient of the i-th width training screen picture feature vector, HiThe i-th column in dictionary H are represented, α indicates to survey Try the rarefaction representation coefficient of screen picture feature vector；
Above-mentioned rarefaction representation coefficient be converted into more convenient solution without constraint linear optimization problem, formula is as follows:
Wherein, RNR dimension real number space is represented, λ is a normal number；
Step 5: the mass fraction pond stage: the mass fraction of test screen image can be according to its corresponding subjective quality point Several and sparse coefficient α*It acquires, score convergence strategy is as follows:
Wherein, Q represents forecast quality score, and range is 0 to 100, dmosiRepresent the corresponding subjective quality point of the i-th width screen picture Number.
CN201811203757.5A 2018-10-16 2018-10-16 Screen picture quality evaluating method based on rarefaction representation Pending CN109544504A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811203757.5A CN109544504A (en) 2018-10-16 2018-10-16 Screen picture quality evaluating method based on rarefaction representation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811203757.5A CN109544504A (en) 2018-10-16 2018-10-16 Screen picture quality evaluating method based on rarefaction representation

Publications (1)

Publication Number Publication Date
CN109544504A true CN109544504A (en) 2019-03-29

Family

ID=65843992

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811203757.5A Pending CN109544504A (en) 2018-10-16 2018-10-16 Screen picture quality evaluating method based on rarefaction representation

Country Status (1)

Country Link
CN (1) CN109544504A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110473181A (en) * 2019-07-31 2019-11-19 天津大学 Screen content image based on edge feature information without ginseng quality evaluating method
CN111325685A (en) * 2020-02-04 2020-06-23 北京锐影医疗技术有限公司 Image enhancement algorithm based on multi-scale relative gradient histogram equalization
CN111369531A (en) * 2020-03-04 2020-07-03 浙江大华技术股份有限公司 Image definition grading method, equipment and storage device
US20200327582A1 (en) * 2019-04-15 2020-10-15 Yandex Europe Ag Method and system for determining result for task executed in crowd-sourced environment
CN113658130A (en) * 2021-08-16 2021-11-16 福州大学 No-reference screen content image quality evaluation method based on dual twin network
US11727329B2 (en) 2020-02-14 2023-08-15 Yandex Europe Ag Method and system for receiving label for digital task executed within crowd-sourced environment

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104581143A (en) * 2015-01-14 2015-04-29 宁波大学 Reference-free three-dimensional picture quality objective evaluation method based on machine learning
CN106504236A (en) * 2016-10-20 2017-03-15 天津大学 Based on rarefaction representation without referring to objective evaluation method for quality of stereo images
CN107292331A (en) * 2017-05-22 2017-10-24 浙江科技学院 Based on unsupervised feature learning without with reference to screen image quality evaluating method
CN107481236A (en) * 2017-08-28 2017-12-15 方玉明 A kind of quality evaluating method of screen picture
CN108038852A (en) * 2017-12-14 2018-05-15 天津师范大学 A kind of Quality Measures for Image Fusion represented based on joint sparse

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104581143A (en) * 2015-01-14 2015-04-29 宁波大学 Reference-free three-dimensional picture quality objective evaluation method based on machine learning
CN106504236A (en) * 2016-10-20 2017-03-15 天津大学 Based on rarefaction representation without referring to objective evaluation method for quality of stereo images
CN107292331A (en) * 2017-05-22 2017-10-24 浙江科技学院 Based on unsupervised feature learning without with reference to screen image quality evaluating method
CN107481236A (en) * 2017-08-28 2017-12-15 方玉明 A kind of quality evaluating method of screen picture
CN108038852A (en) * 2017-12-14 2018-05-15 天津师范大学 A kind of Quality Measures for Image Fusion represented based on joint sparse

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
JIACHEN YANG ET AL: "No Reference Quality Evaluation for Screen Content Images Considering Texture Feature Based on Sparse Representation", 《SIGNAL PROCESSING》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200327582A1 (en) * 2019-04-15 2020-10-15 Yandex Europe Ag Method and system for determining result for task executed in crowd-sourced environment
US11727336B2 (en) * 2019-04-15 2023-08-15 Yandex Europe Ag Method and system for determining result for task executed in crowd-sourced environment
CN110473181A (en) * 2019-07-31 2019-11-19 天津大学 Screen content image based on edge feature information without ginseng quality evaluating method
CN111325685A (en) * 2020-02-04 2020-06-23 北京锐影医疗技术有限公司 Image enhancement algorithm based on multi-scale relative gradient histogram equalization
US11727329B2 (en) 2020-02-14 2023-08-15 Yandex Europe Ag Method and system for receiving label for digital task executed within crowd-sourced environment
CN111369531A (en) * 2020-03-04 2020-07-03 浙江大华技术股份有限公司 Image definition grading method, equipment and storage device
CN111369531B (en) * 2020-03-04 2023-09-01 浙江大华技术股份有限公司 Image definition scoring method, device and storage device
CN113658130A (en) * 2021-08-16 2021-11-16 福州大学 No-reference screen content image quality evaluation method based on dual twin network
CN113658130B (en) * 2021-08-16 2023-07-28 福州大学 Dual-twin-network-based reference-free screen content image quality evaluation method

Similar Documents

Publication Publication Date Title
CN109544504A (en) Screen picture quality evaluating method based on rarefaction representation
CN110046673B (en) No-reference tone mapping image quality evaluation method based on multi-feature fusion
CN107767413B (en) Image depth estimation method based on convolutional neural network
Kundu et al. No-reference quality assessment of tone-mapped HDR pictures
Tian et al. A multi-order derivative feature-based quality assessment model for light field image
CN109978854B (en) Screen content image quality evaluation method based on edge and structural features
CN109598794A (en) The construction method of three-dimension GIS dynamic model
CN103971340A (en) High-bit-width digital image dynamic range compression and detail enhancement method
CN106548169A (en) Fuzzy literal Enhancement Method and device based on deep neural network
CN104751407A (en) Method and device used for blurring image
CN108470178B (en) Depth map significance detection method combined with depth credibility evaluation factor
CN108550146A (en) A kind of image quality evaluating method based on ROI
CN107343196A (en) One kind mixing distortion non-reference picture quality appraisement method
CN108460833A (en) A kind of information platform building traditional architecture digital protection and reparation based on BIM
CN104751406A (en) Method and device used for blurring image
CN110910365A (en) Quality evaluation method for multi-exposure fusion image of dynamic scene and static scene simultaneously
CN114202491B (en) Method and system for enhancing optical image
CN108848365B (en) A kind of reorientation stereo image quality evaluation method
CN111641822A (en) Method for evaluating quality of repositioning stereo image
CN113132700B (en) Method and system for adjusting contrast of projector
CN113935917A (en) Optical remote sensing image thin cloud removing method based on cloud picture operation and multi-scale generation countermeasure network
CN110136105B (en) Method for evaluating definition of same content image based on variance and smoothness
CN112508847A (en) Image quality evaluation method based on depth feature and structure weighted LBP feature
CN113706400A (en) Image correction method, image correction device, microscope image correction method, and electronic apparatus
CN110796635B (en) Light field image quality evaluation method based on shear wave transformation

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20190329