CN107371015A - One kind is without with reference to contrast modified-image quality evaluating method - Google Patents
One kind is without with reference to contrast modified-image quality evaluating method Download PDFInfo
- Publication number
- CN107371015A CN107371015A CN201710600600.5A CN201710600600A CN107371015A CN 107371015 A CN107371015 A CN 107371015A CN 201710600600 A CN201710600600 A CN 201710600600A CN 107371015 A CN107371015 A CN 107371015A
- Authority
- CN
- China
- Prior art keywords
- image
- contrast
- modified
- contrast modified
- evaluating method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N17/00—Diagnosis, testing or measuring for television systems or their details
- H04N17/004—Diagnosis, testing or measuring for television systems or their details for digital television systems
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to one kind without contrast modified-image quality evaluating method is referred to, subjective perception characteristic of the human visual system for contrast modified-image is taken into full account, extracts four characteristics of image respectively:Definition, comentropy, brightness and contrast;And combination supporting vector machine SVM (Support Vector Machine) is trained to obtain the mapping relations MODEL C CQAM of the characteristic vector of contrast modified-image and subjective quality scores to extracted characteristics of image;The mass fraction of contrast modified-image is finally evaluated using the mapping relations MODEL C CQAM of training gained.Method of the present invention calculates simply, and practicality is stronger, and close with human eye subjective quality assessment, can evaluate the mass fraction of contrast modified-image well.
Description
Technical field
The present invention relates to image processing field, is commented more specifically to one kind without contrast modified-image quality is referred to
Valency method.
Background technology
With the fast development of cloud computing, big data, multimedia communication and consumer electronics, image quality evaluation turns into science
Boundary and the important research content of industrial quarters.Because image is readily incorporated various noises, example when obtaining, encode, transmitting and showing
Such as Gaussian noise, fuzzy distortion, compression artefacts, color distortion, contrast change etc., how accurate evaluation picture quality is to promoting
Image processing techniques development has important theory significance and actual application value.
Image quality evaluating method is broadly divided into subjective quality assessment and evaluating objective quality.
Subjective picture quality evaluation is that final score is judged image and provided taking human as observer.It is but subjective
Image quality evaluation is not often suitable for practical application because it is time-consuming, laborious and is easily influenceed by factors such as environment.
Objective image quality evaluation imitates human visual system with evaluation result and subjectivity by founding mathematical models
Quality evaluation is consistent, and simple to operate, with more actual application value.
Objective image quality evaluation method according to reference picture provide degree can be divided into full reference image quality appraisement method,
Half reference image quality appraisement method and non-reference picture quality appraisement method.
Full reference image quality appraisement method utilizes the full detail of reference picture, by calculating reference picture and distortion map
The similarity of picture, obtain the quality evaluation value of distorted image.Such as classical mean square error (Mean Square Error, MSE)
With Y-PSNR (Peak Signal to Noise Ratio, PSNR).
Half reference image quality appraisement only uses the perceived quality that reference picture Partial Feature information carrys out calculated distortion image.
But in actual applications often without reference to image information, both approaches have certain limitation.Non-reference picture matter
Amount evaluation method can make corresponding evaluation score without reference to image to image, have higher practical value.
Typical non-reference picture quality appraisement method, there is the non-reference picture quality appraisement model based on two-stage framework
(Blind Image Quality Index, BIQI), the non-reference picture matter based on spatial domain normalization coefficient extraction regression model
Evaluation model (Blind/Referenceless Image Spatial Quality Evaluator, BRISQUE) etc. is measured, all
Most of type of distortion (e.g., compression, JPEG distortions) can be assessed well.
However, existing many image quality evaluating methods are all unable to distortion caused by the change of effective evaluation contrast.At present
Lack the non-reference picture quality appraisement method to deteriroation of image quality caused by contrast change in image processing field.
The content of the invention
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of practical and easy calculating without reference
Contrast modified-image quality evaluating method.
Technical scheme is as follows:
For one kind without contrast modified-image quality evaluating method is referred to, step is as follows:
1) the contrast modified-image I for training pattern is inputtedk, k=1,2 ..., K, K be training process in contrast
Modified-image sum;
2) to contrast modified-image IkPre-processed to obtain pretreatment image I'k;
Calculate contrast modified-image IkCharacteristics of image:Contrast modified-image IkDefinition and comentropy, and in advance
Handle image I 'kBrightness and contrast, the characteristic vector using these characteristics of image as current contrast modified-image;
3) corresponding subjective quality scores are combined to the characteristic vector of all contrast modified-images using SVMs
It is trained, obtains the characteristic vector of contrast modified-image and the mapping relations model of subjective quality scores;
4) input contrast modified-image to be measured, according to the method for step 2) be calculated the feature of image to be tested to
Amount;
5) characteristic vector of contrast modified-image to be measured is inputted into mapping relations model, exports contrast variation diagram to be measured
The mass fraction of picture.
Preferably, to contrast modified-image IkCarrying out pretreatment includes gray processing processing, filtering process.
Preferably, to contrast modified-image IkThe step of being pre-processed is as follows:
2.1) gray processing contrast modified-image Ik, gray level image is obtained, only retains monochrome information;
2.2) gray level image is filtered, step is as follows:
2.2.1 Fourier transformation) is carried out to gray level image, obtains frequency domain figure as Fk;
2.2.2) utilize Contrast sensitivity functionTo frequency domain figure as FkCarry out
Filtering process, obtain filtered image
2.2.3) to filtered imageCarry out Fourier inversion and obtain pretreatment image I 'k;
Wherein,U and v be corresponding pixel points frequency domain variable, k=1,2 ..., K.
Preferably, in step 2), definition
Wherein,Represent respectively on x and y directions
Difference, M and N represent contrast modified-image I respectivelykSize in the horizontal and vertical directions.
Preferably, in step 2), comentropy
Wherein,Represent contrast modified-image IkMiddle gray value is the ratio shared by i pixel, and f (i) is
Contrast modified-image IkMiddle gray value is the frequency that i occurs, and (0≤i≤255), M and N represent image horizontal and vertical respectively
The upward size of Nogata.
Preferably, in step 2), monochrome information
Preferably, in step 2), contrast information
Wherein, P and Q represents pretreatment image I ' respectivelykSize in the horizontal and vertical directions.
Preferably, all contrast modified-image I will be based onkThe eigenvalue cluster of gained is calculated into eigenvectors matrixUsing SVMs to all contrast modified-image IkEigenvectors matrix R combine it is subjective
Mass fraction is trained, and obtains the characteristic vector of contrast modified-image and the mapping relations model of subjective quality scores.
Preferably, the training process of mapping relations model is off-line training.
Beneficial effects of the present invention are as follows:
It is of the present invention without refer to contrast modified-image quality evaluating method, take into full account human visual system for
The subjective perception characteristic of contrast modified-image, four characteristics of image are extracted respectively:Definition, comentropy, brightness and contrast;
And combination supporting vector machine SVM (Support Vector Machine) is trained to obtain contrast to extracted characteristics of image
The characteristic vector of modified-image and the mapping relations MODEL C CQAM of subjective quality scores;Finally closed using the mapping of training gained
It is the mass fraction of MODEL C CQAM evaluation contrast modified-images.
Method of the present invention calculates simply, and practicality is stronger, and close with human eye subjective quality assessment, can be well
Evaluate the mass fraction of contrast modified-image.
Brief description of the drawings
Fig. 1 is the FB(flow block) of training step of the present invention;
Fig. 2 is the FB(flow block) of evaluation procedure of the present invention.
Embodiment
The present invention is described in further detail below in conjunction with drawings and Examples.
The present invention in order to solve not being suitable for existing for prior art contrast change caused by image quality distortion evaluation
Phenomenon, make up deficiency of the algorithm to image quality distortion evaluation caused by contrast change of prior art, there is provided one kind is without ginseng
Contrast modified-image quality evaluating method is examined, as shown in Figure 1 and Figure 2, method of the present invention includes training step and evaluation
Step, wherein, step 1), step 2), step 3) they are training step, and the training process of mapping relations model is off-line training;Step
It is rapid 4), step 5) be evaluation procedure.
Comprise the following steps that:
1) the contrast modified-image I for training pattern is inputtedk, k=1,2 ..., K, K be training process in contrast
Modified-image sum.
2) to contrast modified-image IkPre-processed to obtain pretreatment image I'k, calculate contrast modified-image Ik's
Characteristics of image:Contrast modified-image IkDefinition and comentropy, and pretreatment image I'kBrightness and contrast, by this
A little characteristic vectors of the characteristics of image as current contrast modified-image.
In the present embodiment, to contrast modified-image IkCarrying out pretreatment includes gray processing processing, filtering process.
To contrast modified-image IkThe step of being pre-processed is as follows:
2.1) gray processing contrast modified-image Ik, gray level image is obtained, only retains monochrome information;
2.2) using Contrast sensitivity function (Contrast Sensitivity Filtering, CSF) to gray processing figure
As being filtered, step is as follows:
2.2.1 Fourier transformation) is carried out to gray level image, obtains frequency domain figure as Fk;
2.2.2) utilize Contrast sensitivity functionTo frequency domain figure as FkCarry out
Filtering process, obtain filtered image
2.2.3) to filtered imageCarry out Fourier inversion and obtain pretreatment image I 'k;
Wherein,U and v is the frequency domain variable of corresponding pixel points, k=1,2 ..., K, K.
In the present embodiment, definition
Wherein,Represent respectively on x and y directions
Difference, M and N represent contrast modified-image I respectivelykSize in the horizontal and vertical directions.
Comentropy
Wherein,Represent contrast modified-image IkMiddle gray value is the ratio shared by i pixel, and f (i) is
Contrast modified-image IkMiddle gray value is the frequency that i occurs, and (0≤i≤255), M and N represent image horizontal and vertical respectively
The upward size of Nogata.
Monochrome information
Contrast information
Wherein, P and Q represents pretreatment image I ' respectivelykSize in the horizontal and vertical directions.
3) using support vector machines (Support Vector Machine) to the feature of all contrast modified-images
Vector is trained with reference to subjective quality scores, will be based on all contrast modified-image IkCalculate the eigenvalue cluster Cheng Te of gained
Levy vector matrixUsing SVMs to all contrast modified-image IkCharacteristic vector square
Battle array R combination subjective quality scores are trained, and obtain the characteristic vector of contrast modified-image and the mapping of subjective quality scores
Relational model CCQAM.
4) the contrast modified-image of quality evaluation is treated in input, and contrast to be measured, which is calculated, according to the method for step 2) becomes
Change the characteristic vector of image;
5) characteristic vector of contrast modified-image to be measured is inputted into mapping relations model, exports contrast variation diagram to be measured
The mass fraction of picture.
Above-described embodiment is intended merely to the explanation present invention, and is not used as limitation of the invention.As long as according to this hair
Bright technical spirit, above-described embodiment is changed, modification etc. will all fall in the range of the claim of the present invention.
Claims (9)
- It is 1. a kind of without with reference to contrast modified-image quality evaluating method, it is characterised in that step is as follows:1) the contrast modified-image I for training pattern is inputtedk, k=1,2 ..., K, K be training process in contrast change Total number of images;2) to contrast modified-image IkPre-processed to obtain pretreatment image I'k;Calculate contrast modified-image IkCharacteristics of image:Contrast modified-image IkDefinition and comentropy, and pretreatment Image I 'kBrightness and contrast, the characteristic vector using these characteristics of image as current contrast modified-image;3) the characteristic vector combination subjective quality scores of all contrast modified-images are trained using SVMs, obtained To the characteristic vector of contrast modified-image and the mapping relations model of subjective quality scores;4) contrast modified-image to be measured is inputted, the feature of contrast modified-image to be measured is calculated according to the method for step 2) Vector;5) characteristic vector of contrast modified-image to be measured is inputted into mapping relations model, exports contrast modified-image to be measured Mass fraction.
- It is 2. according to claim 1 without with reference to contrast modified-image quality evaluating method, it is characterised in that step 2) In, to contrast modified-image IkCarrying out pretreatment includes gray processing processing, filtering process.
- It is 3. according to claim 2 without with reference to contrast modified-image quality evaluating method, it is characterised in that to contrast Modified-image IkThe step of being pre-processed is as follows:2.1) gray processing contrast modified-image Ik, gray level image is obtained, only retains monochrome information;2.2) gray level image is filtered, step is as follows:2.2.1 Fourier transformation) is carried out to gray level image, obtains frequency domain figure as Fk;2.2.2) utilize Contrast sensitivity functionTo frequency domain figure as FkIt is filtered Processing, obtains filtered image2.2.3) to filtered imageCarry out Fourier inversion and obtain pretreatment image I 'k;Wherein,U and v be corresponding pixel points frequency domain variable, k=1,2 ..., K.
- 4. existing without reference contrast modified-image quality evaluating method, its feature according to any one of claims 1 to 3 In, in step 2), definitionWherein,The difference on x and y directions is represented respectively Point, M and N represent contrast modified-image I respectivelykSize in the horizontal and vertical directions.
- 5. existing without reference contrast modified-image quality evaluating method, its feature according to any one of claims 1 to 3 In, in step 2), comentropyWherein,Represent contrast modified-image IkMiddle gray value is the ratio shared by i pixel, and f (i) is contrast Modified-image IkMiddle gray value is the frequency that i occurs, and (0≤i≤255), M and N represent image both horizontally and vertically respectively On size.
- 6. existing without reference contrast modified-image quality evaluating method, its feature according to any one of claims 1 to 3 In, in step 2), monochrome informationWherein, P and Q represents pretreatment image I ' respectivelykSize in the horizontal and vertical directions.
- 7. existing without reference contrast modified-image quality evaluating method, its feature according to any one of claims 1 to 3 In, in step 2), contrast informationWherein, P and Q represents pretreatment image I ' respectivelykSize in the horizontal and vertical directions.
- It is 8. according to claim 1 without with reference to contrast modified-image quality evaluating method, it is characterised in that institute will be based on There is contrast modified-image IkThe eigenvalue cluster of gained is calculated into eigenvectors matrixUtilize support Vector machine is to all contrast modified-image IkEigenvectors matrix R combination subjective quality scores be trained, contrasted Spend the characteristic vector of modified-image and the mapping relations model of subjective quality scores.
- It is 9. according to claim 1 without with reference to contrast modified-image quality evaluating method, it is characterised in that mapping relations The training process of model is off-line training.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710600600.5A CN107371015A (en) | 2017-07-21 | 2017-07-21 | One kind is without with reference to contrast modified-image quality evaluating method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710600600.5A CN107371015A (en) | 2017-07-21 | 2017-07-21 | One kind is without with reference to contrast modified-image quality evaluating method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107371015A true CN107371015A (en) | 2017-11-21 |
Family
ID=60307186
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710600600.5A Pending CN107371015A (en) | 2017-07-21 | 2017-07-21 | One kind is without with reference to contrast modified-image quality evaluating method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107371015A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108769671A (en) * | 2018-06-13 | 2018-11-06 | 天津大学 | Stereo image quality evaluation method based on adaptive blending image |
CN111354048A (en) * | 2020-02-24 | 2020-06-30 | 清华大学深圳国际研究生院 | Quality evaluation method and device for camera-oriented acquired pictures |
CN111489333A (en) * | 2020-03-31 | 2020-08-04 | 天津大学 | No-reference night natural image quality evaluation method |
CN111753873A (en) * | 2020-05-12 | 2020-10-09 | 北京捷通华声科技股份有限公司 | Image detection method and device |
CN115187519A (en) * | 2022-06-21 | 2022-10-14 | 上海市计量测试技术研究院 | Image quality evaluation method, system and computer readable medium |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7289679B2 (en) * | 2003-08-12 | 2007-10-30 | International Business Machines Corporation | System and method for measuring image quality using compressed image data |
CN102202227A (en) * | 2011-06-21 | 2011-09-28 | 珠海世纪鼎利通信科技股份有限公司 | No-reference objective video quality assessment method |
CN102254185A (en) * | 2011-07-21 | 2011-11-23 | 西安电子科技大学 | Background clutter quantizing method based on contrast ratio function |
CN102740114A (en) * | 2012-07-16 | 2012-10-17 | 公安部第三研究所 | Non-parameter evaluation method for subjective quality of video |
CN102881010A (en) * | 2012-08-28 | 2013-01-16 | 北京理工大学 | Method for evaluating perception sharpness of fused image based on human visual characteristics |
CN103049915A (en) * | 2013-01-12 | 2013-04-17 | 深圳市华星光电技术有限公司 | Establishment method of evaluation criterion parameters and method for evaluating image quality of display screen |
CN103051922A (en) * | 2013-01-12 | 2013-04-17 | 深圳市华星光电技术有限公司 | Establishment method of evaluation criterion parameters and method for evaluating image quality of display screen |
CN106303507A (en) * | 2015-06-05 | 2017-01-04 | 江苏惠纬讯信息科技有限公司 | Video quality evaluation without reference method based on space-time united information |
CN106600597A (en) * | 2016-12-22 | 2017-04-26 | 华中科技大学 | Non-reference color image quality evaluation method based on local binary pattern |
-
2017
- 2017-07-21 CN CN201710600600.5A patent/CN107371015A/en active Pending
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7289679B2 (en) * | 2003-08-12 | 2007-10-30 | International Business Machines Corporation | System and method for measuring image quality using compressed image data |
CN102202227A (en) * | 2011-06-21 | 2011-09-28 | 珠海世纪鼎利通信科技股份有限公司 | No-reference objective video quality assessment method |
CN102254185A (en) * | 2011-07-21 | 2011-11-23 | 西安电子科技大学 | Background clutter quantizing method based on contrast ratio function |
CN102740114A (en) * | 2012-07-16 | 2012-10-17 | 公安部第三研究所 | Non-parameter evaluation method for subjective quality of video |
CN102881010A (en) * | 2012-08-28 | 2013-01-16 | 北京理工大学 | Method for evaluating perception sharpness of fused image based on human visual characteristics |
CN103049915A (en) * | 2013-01-12 | 2013-04-17 | 深圳市华星光电技术有限公司 | Establishment method of evaluation criterion parameters and method for evaluating image quality of display screen |
CN103051922A (en) * | 2013-01-12 | 2013-04-17 | 深圳市华星光电技术有限公司 | Establishment method of evaluation criterion parameters and method for evaluating image quality of display screen |
CN106303507A (en) * | 2015-06-05 | 2017-01-04 | 江苏惠纬讯信息科技有限公司 | Video quality evaluation without reference method based on space-time united information |
CN106600597A (en) * | 2016-12-22 | 2017-04-26 | 华中科技大学 | Non-reference color image quality evaluation method based on local binary pattern |
Non-Patent Citations (2)
Title |
---|
范媛媛等: "对比敏感度下的图像对比度评价", 《沈阳建筑大学学报(自然科学版)》 * |
陈振跃等: "基于人眼视觉系统的假彩色融合图像质量的评价方法", 《光子学报》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108769671A (en) * | 2018-06-13 | 2018-11-06 | 天津大学 | Stereo image quality evaluation method based on adaptive blending image |
CN111354048A (en) * | 2020-02-24 | 2020-06-30 | 清华大学深圳国际研究生院 | Quality evaluation method and device for camera-oriented acquired pictures |
CN111489333A (en) * | 2020-03-31 | 2020-08-04 | 天津大学 | No-reference night natural image quality evaluation method |
CN111489333B (en) * | 2020-03-31 | 2022-06-03 | 天津大学 | No-reference night natural image quality evaluation method |
CN111753873A (en) * | 2020-05-12 | 2020-10-09 | 北京捷通华声科技股份有限公司 | Image detection method and device |
CN115187519A (en) * | 2022-06-21 | 2022-10-14 | 上海市计量测试技术研究院 | Image quality evaluation method, system and computer readable medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107371015A (en) | One kind is without with reference to contrast modified-image quality evaluating method | |
CN105744256B (en) | Based on the significant objective evaluation method for quality of stereo images of collection of illustrative plates vision | |
CN102333233B (en) | Stereo image quality objective evaluation method based on visual perception | |
CN101882304B (en) | Self-adaptive de-noising and characteristic enhancing method of SAR (Synthetic Aperture Radar) image | |
CN109191428B (en) | Masking texture feature-based full-reference image quality evaluation method | |
CN107194872A (en) | Remote sensed image super-resolution reconstruction method based on perception of content deep learning network | |
CN107481236A (en) | A kind of quality evaluating method of screen picture | |
Yue et al. | Blind stereoscopic 3D image quality assessment via analysis of naturalness, structure, and binocular asymmetry | |
CN109831664B (en) | Rapid compressed stereo video quality evaluation method based on deep learning | |
CN107146220B (en) | A kind of universal non-reference picture quality appraisement method | |
CN104811691B (en) | A kind of stereoscopic video quality method for objectively evaluating based on wavelet transformation | |
CN104036493B (en) | No-reference image quality evaluation method based on multifractal spectrum | |
CN104318545B (en) | A kind of quality evaluating method for greasy weather polarization image | |
CN107464222B (en) | Based on tensor space without reference high dynamic range images method for evaluating objective quality | |
CN109523513A (en) | Based on the sparse stereo image quality evaluation method for rebuilding color fusion image | |
CN104240248B (en) | Method for objectively evaluating quality of three-dimensional image without reference | |
CN104574381B (en) | A kind of full reference image quality appraisement method based on local binary patterns | |
CN101950422A (en) | Singular value decomposition(SVD)-based image quality evaluation method | |
CN107610093B (en) | Full-reference image quality evaluation method based on similarity feature fusion | |
CN109788275A (en) | Naturality, structure and binocular asymmetry are without reference stereo image quality evaluation method | |
CN102547368A (en) | Objective evaluation method for quality of stereo images | |
CN109429051B (en) | Non-reference stereo video quality objective evaluation method based on multi-view feature learning | |
CN110120034B (en) | Image quality evaluation method related to visual perception | |
CN112184672A (en) | No-reference image quality evaluation method and system | |
Yang et al. | A no-reference optical flow-based quality evaluator for stereoscopic videos in curvelet domain |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171121 |
|
RJ01 | Rejection of invention patent application after publication |