CN107610101B - Method for measuring visual balance quality of digital image - Google Patents
Method for measuring visual balance quality of digital image Download PDFInfo
- Publication number
- CN107610101B CN107610101B CN201710732837.9A CN201710732837A CN107610101B CN 107610101 B CN107610101 B CN 107610101B CN 201710732837 A CN201710732837 A CN 201710732837A CN 107610101 B CN107610101 B CN 107610101B
- Authority
- CN
- China
- Prior art keywords
- visual
- image
- digital image
- visual balance
- balance
- 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.)
- Active
Links
Images
Abstract
The invention relates to a method for measuring the visual balance quality of a digital image, belonging to the technical field of digital image processing and data analysis. The invention adopts image mining to preprocess a digital image and acquire data for analysis, then establishes a model for quantitatively measuring the visual balance quality of the digital image based on the moment balance principle and the visual balance theory, selects the image which can reach the visual balance for characteristic extraction and calculation and obtains a threshold value K for judging whether the digital image is in visual balance or not, and further compares the visual balance quality index Q of the digital image to be measuredrAnd the size of the corresponding threshold value K, thereby judging whether the image visual balance quality is good or not. The invention converts the qualitative evaluation of the visual balance quality of the digital image into quantitative evaluation, can further carry out visual balance quality sequencing on the digital image, can also give the weight bias condition of the visual weight of the non-visual balance image along the transverse axis and the longitudinal axis, and has the advantages of convenient operation, simple method and good use effect.
Description
Technical Field
The invention relates to a method for measuring the visual balance quality of a digital image, belonging to the technical field of digital image processing and data analysis.
Background
In recent years, with the continuous development of computer technology, the application field of digital images is more and more extensive, and it becomes a convenient and effective way to acquire a large amount of useful information from images for research. Visual balance, one of the criteria for measuring the aesthetic quality of digital images, plays a guiding role in artistic design, and is also significant in maintaining the legibility and the aesthetic appearance of image works. When the works are designed, the visual balance degree is well mastered, the relationship among all elements in the images can be reasonable and coordinated, and unexpected visual effects are realized. However, because the human psychology is difficult to have a uniform standard, the visual balance quality of the image is difficult to be quantized all the time, so that a simple and feasible model for quantitatively describing the visual balance is lacked, and how to make the design and use of the digital image more effective makes the image more beautiful and easier to read on the premise of meeting the availability of the digital image, the visual balance quality of the whole digital image must be better.
At present, some defects exist in the evaluation of visual works and the existing visual balance quality measurement model, and the defects are as follows:
(1) most digital image designers make qualitative assessment on the visual balance quality of the work mainly by means of long-term experience of the designers in the design, and the method has strong subjectivity, often has the phenomenon that different designers draw inconsistent conclusions, and inevitably has serious influence on the design process and results;
(2) sequencing often fails when a designer judges whether the visual balance quality of a work is good or bad, and the designer can evaluate but cannot find a basis;
(3) the visual balance of the image work is of poor quality and it is unknown how to modify it.
Disclosure of Invention
The invention overcomes the defects of the method and provides a simple and feasible new method with higher application value for quantitatively measuring the visual balance quality of the digital image. The method effectively solves the problem that the visual balance quality of the digital image cannot be quantitatively measured in the design, acquires information from the digital image, converts the qualitative measurement into the quantitative measurement based on the moment balance theory, has very important practical value in the aspect of sequencing the visual balance quality of the image, and has practical theoretical significance for the effective use of the image works.
The technical scheme of the invention is as follows: a method for measuring the visual balance quality of digital image includes such steps as preprocessing the digital image by image mining, analyzing the acquired data, creating model based on moment balance principle and visual balance theory, extracting features of the image, calculating to obtain the threshold value K for judging if the digital image is visual balance, and comparing the visual balance quality index Q of the digital imagerAnd the size of the corresponding threshold value K, thereby judging whether the image visual balance quality is good or not.
The method comprises the following specific steps:
step1, obtaining a digital image by digitalizing the analog image;
step2, acquiring image data of the digital image, acquiring an original image data set for data mining, and carrying out gray processing on the image; analyzing the image data to obtain image characteristics;
step3, according to a moment balance principle and a visual balance theory, determining corresponding visual weight, visual gravity moment and visual balance mass, and establishing an image visual balance mass quantitative measurement model:
wherein m and n are respectively the line number and column number of the gray matrix of the pixel block of the map image, and gijIs the visual weight, x, of the pixel blocks in the ith row and j columns of the imageijIs the abscissa value of the center point of the ith row and j column of pixel blocks, wherein the origin point is the geometric center of the image, gijxijThe visual gravity moment of the pixel blocks in the ith row and the j column on the horizontal axis; in the same way, yijIs the ordinate value of the center point of the pixel block in the ith row and j columnijyijThe visual gravity moment of the pixel blocks in the ith row and the jth column on the longitudinal axis; q1、Q2The visual balance quality indexes of the image in the directions of the horizontal axis and the vertical axis respectively, the signs of the visual balance quality indexes reflect the weight bias of the visual weight of the whole image in the directions of the horizontal axis and the vertical axis, Q is the visual balance quality index of the whole image, and in addition, as the visual center point of the image is positioned above the geometric center, the best visual balance quality index Q which is achieved in the rectangular range of 0.2 × 0.2.2 around the geometric center in the visual attention center area is selectedrAs an index for judging whether the image reaches visual balance, QrIs the minimum value of Q in the central region of visual attention;
step4, performing statistical calculation of a threshold value K, selecting K images which can reach visual balance according to the digital image type limiting conditions in the formula (1), performing feature extraction and calculation on all images which meet the conditions, and obtaining Q corresponding to each imagerValue, then take all QkThe average value of the values is used as the threshold value K of the digital image, and the formula is as follows:
step5, comparing visual balance quality index Q of digital imagerA threshold value K, if Q corresponding to the class of imagerMore than or equal to K, the image reaches visual balance and is according to QrSequencing the visual balance quality of the images; if Qr<K, the image does not reach the visual balance, namely the visual balance quality is poor, and the index Q is divided according to the calculated visual balance quality1And Q2Judging the unbalanced condition of the visual weight of the map image along the horizontal axis and the vertical axis;
when | Q1< K and Q1<0, the map image in the horizontal axis direction has heavier visual weight in the area at the left side of the visual attention center;
when | Q1Less than K and less than or equal to 01If < K, the weight bias is large on the right side;
when | Q1If | > K, the map image has no obvious weight bias in the direction of the horizontal axis;
when | Q2< K and Q2<0, the visual weight of the map image in the area above the visual attention center in the vertical axis direction is heavier;
when | Q2Less than K and less than or equal to 02If < K, the weight is larger below the K;
when | Q2And | ≧ K, it is known that the map image has no obvious weight bias in the longitudinal axis direction.
In Step3, the visual gravity moments of all pixel blocks are key factors for determining whether the digital image as a whole reaches visual balance, and the visual gravity moments of the pixel blocks are the visual gravity multiplied by the distance from the center point of the pixel block to the visual center of the image, where the visual gravity is the weight of each pixel block and is composed of its gray value, white is recorded as 255, black is recorded as 0, and the period is other gray values.
The elements in the model of the method are pixels of the digital image, and the factors influencing the visual weight in the image are considered: position, form, size, hue, color brightness, contrast, proximity to the pivot, etc., so that an electronic image obtained by copying a paper photograph with a camera or a mobile phone is not within the applicable scope of the method.
The image data acquisition and acquisition is to acquire and acquire an original image data set for data mining from an existing digital image work; after an original image is obtained, the image is subjected to pixel unified processing by using the existing Photoshop software, so that the original image can be subjected to early-stage pixel unified processing and controlled in a small range for increasing the visibility of an experiment and simple calculation (the quantity of screenshot pixel blocks does not influence the measurement of the visual balance quality through experimental tests, but the numerical values measured by the experiment are different in precision); in the existing Matlab software, the function rgb2gray () is used to perform the graying process on the image, and the function imread () is used to read the grayscale data in the grayscale image file and present the grayscale data in the form of a matrix.
The invention has the beneficial effects that:
1. the method is simple and feasible, convenient to operate, good in using effect, supported by a visual balance theory, and high in practical value;
2. the method converts the qualitative evaluation of the visual balance quality of the digital image into quantitative evaluation;
3. the method can sequence the visual balance quality of the digital image;
4. the digital image model which does not reach the visual balance standard can also give the weight bias condition of the visual weight in the horizontal and vertical directions, thereby providing a basis for further improvement;
5. the invention is applied to the quantitative measurement of the visual balance quality of the digital image, provides a reliable and practical method for quantitatively measuring the visual balance effect of the image, and can also be applied to the aspects of the design and creation of artistic works, the content positioning of digital advertisements, the cutting of photos and the like;
6. the method obtains the relevant information by processing the image processing technology, and relevant parameters in the model are obtained by mathematical statistics, so that the method is easy to operate and objective in result; the method can not only convert qualitative evaluation of the digital image visual balance quality into quantitative evaluation, and the numerical indexes of the quantitative evaluation are used for sequencing the digital image visual balance degree, but also provide modified reference suggestions for non-visual balance images.
Drawings
FIG. 1 is a north American map of the present invention;
FIG. 2 is a gray scale map of a north American map image in a two-dimensional coordinate system in accordance with the present invention;
FIG. 3 is an enlarged view of the visual center point for determining the best visual balance mass according to the present invention;
FIG. 4 is a bar graph of data obtained from a set of maps in accordance with the present invention.
Detailed Description
Example 1: as shown in fig. 1-4, a method for measuring visual balance quality of a digital image comprises the following specific steps:
(1) acquiring a digital image north america map as shown in fig. 1;
(2) the map image is processed to obtain a gray image and is placed in a two-dimensional coordinate as shown in fig. 2, and the length and the width of one pixel are taken as a scale unit. Wherein, the visual center area is 0.2 × 0.2 unit areas around the geometric center (i.e. the origin O of the coordinate system) as shown in fig. 3.
Obtaining the gray matrix of the pixel block of the north American map as
(3) Reading data information in the matrix and applying the matrix to a model, and calculating visual balance quality judgment indexes Q of the North American map image along the horizontal axis and the vertical axis respectively1And Q2Respectively-0.196 and 0.167, and the best visual balance quality index Q of the map imager81.85%;
further, in step3, the visual gravity moments of all the pixel blocks are key factors for determining whether the digital image as a whole reaches visual balance, and the visual gravity moments of the pixel blocks are the visual gravity multiplied by the distance from the center point of the pixel blocks to the visual center of the image, where the visual gravity is the weight of each pixel block and is composed of its gray value, white is recorded as 255, black is recorded as 0, and the period is other gray values.
(4) 50 standard map images are downloaded from a national mapping geographic information bureau standard map service website (http:// bzdt. nasg. gov. cn /), the image aspect ratios of the 50 maps are not identical and are in RGB colors, and the image sizes are different from 973 x 1201 to 3348 x 2347. Selecting 46 images which are considered by all related professionals to achieve visual balance, respectively carrying out early-stage pixel reduction on the images by using Photoshop software and controlling the pixels to be between 16 multiplied by 22 and 20 multiplied by 32px for increasing the visibility of an experiment and simple calculation, then further processing the images by using the conventional Matlab software, using the obtained data for model calculation and summarizing the data into a bar chart shown in FIG. 4, and obtaining a threshold value K for judging whether the images achieve visual balance to be 58.5%;
(5) will QrThe value is compared with the K value to know Qr>K, the north american map, meets the visual balance criteria. In addition, the value 81.85% can also be used for quantifying the visual balance quality of the map image and for ordering the visual balance quality of different digital images; index Q of calculation result1And Q2The signs of the values can respectively reflect the weight bias of the digital map image visual weight along the horizontal axis and the vertical axis, i.e. if Q1<0, indicating that the visual weight of the map image on the left side of the visual attention center area along the horizontal direction is heavier; if Q1>0, indicating that the right side has heavier weight; by the same token, Q can be known2<0 and Q2>0, so that in the event that the digital map image does not meet the visual balance criteria, analysis or improvement can be made based on this determination.
While the present invention has been described in detail with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, and various changes can be made without departing from the spirit of the present invention within the knowledge of those skilled in the art.
Claims (2)
1. A method of measuring visual balance quality of a digital image, comprising: preprocessing a digital image by adopting image mining, acquiring data and analyzing the data, and then establishing a model for quantitatively measuring the vision of the digital image based on a moment balance principle and a visual balance theoryBalancing quality, selecting image with visual balance, extracting and calculating characteristics, obtaining threshold K for judging whether digital image is visual balance, and comparing visual balance quality index Q of digital image to be measuredrThe image quality is judged according to the size of the corresponding threshold value K;
the method comprises the following specific steps:
step1, obtaining a digital image by digitalizing the analog image;
step2, acquiring image data of the digital image, acquiring an original image data set for data mining, and carrying out gray processing on the image; analyzing the image data to obtain image characteristics;
step3, according to a moment balance principle and a visual balance theory, determining corresponding visual weight, visual gravity moment and visual balance mass, and establishing an image visual balance mass quantitative measurement model:
wherein m and n are respectively the line number and column number of the gray matrix of the pixel block of the map image, and gijIs the visual weight, x, of the pixel blocks in the ith row and j columns of the imageijIs the abscissa value of the center point of the ith row and j column of pixel blocks, wherein the origin point is the geometric center of the image, gijxijThe visual gravity moment of the pixel blocks in the ith row and the j column on the horizontal axis; in the same way, yijIs the ordinate value of the center point of the pixel block in the ith row and j columnijyijThe visual gravity moment of the pixel blocks in the ith row and the jth column on the longitudinal axis; q1、Q2The visual balance quality indexes of the image in the directions of the horizontal axis and the vertical axis respectively, the signs of the visual balance quality indexes reflect the weight bias of the visual weight of the whole image in the directions of the horizontal axis and the vertical axis, Q is the visual balance quality index of the whole image, and in addition, as the visual center point of the image is positioned above the geometric center, the best visual balance quality index Q which is achieved in the rectangular range of 0.2 × 0.2.2 around the geometric center in the visual attention center area is selectedrAs a means for judging whether the image reachesIndex of visual balance, QrIs the minimum value of Q in the central region of visual attention;
step4, performing statistical calculation of a threshold value K, selecting K images which can reach visual balance according to the digital image type limiting conditions in the formula (1), performing feature extraction and calculation on all images which meet the conditions, and obtaining Q corresponding to each imagerValue, then take all QkThe average value of the values is used as a threshold K for judging whether the digital image is in visual balance, and the formula is as follows:
step5, comparing visual balance quality index Q of digital imagerAnd a threshold K for judging whether the digital image is in visual balance, if QrMore than or equal to K, the image reaches visual balance and is according to QrSequencing the visual balance quality of the images; if Qr<K, the image does not reach the visual balance, namely the visual balance quality is poor, and the index Q is divided according to the calculated visual balance quality1And Q2Judging the unbalanced condition of the visual weight of the map image along the horizontal axis and the vertical axis;
when | Q1< K and Q1<0, the map image in the horizontal axis direction has heavier visual weight in the area at the left side of the visual attention center;
when | Q1Less than K and less than or equal to 01If < K, the weight bias is large on the right side;
when | Q1If | > K, the map image has no obvious weight bias in the direction of the horizontal axis;
when | Q2< K and Q2<0, the visual weight of the map image in the area above the visual attention center in the vertical axis direction is heavier;
when | Q2Less than K and less than or equal to 02If < K, the weight is larger below the K;
when | Q2And | ≧ K, it is known that the map image has no obvious weight bias in the longitudinal axis direction.
2. The method of measuring the visual balance quality of a digital image according to claim 1, wherein: in Step3, the visual gravity moment of the pixel block is the visual gravity multiplied by the distance from the center point of the pixel block to the visual center of the image, where the visual gravity is the weight of each pixel block and is composed of its gray value, white is recorded as 255, black is recorded as 0, and the period is other gray values.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710732837.9A CN107610101B (en) | 2017-08-24 | 2017-08-24 | Method for measuring visual balance quality of digital image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710732837.9A CN107610101B (en) | 2017-08-24 | 2017-08-24 | Method for measuring visual balance quality of digital image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107610101A CN107610101A (en) | 2018-01-19 |
| CN107610101B true CN107610101B (en) | 2020-08-25 |
Family
ID=61064164
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710732837.9A Active CN107610101B (en) | 2017-08-24 | 2017-08-24 | Method for measuring visual balance quality of digital image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107610101B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111798457B (en) * | 2020-06-10 | 2021-04-06 | 上海众言网络科技有限公司 | Image visual weight determining method and device and image evaluation method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102789737A (en) * | 2012-08-11 | 2012-11-21 | 中国人民解放军信息工程大学 | Method for establishing visual balance model of map on basis of moment balance principle |
| CN104700416A (en) * | 2015-03-23 | 2015-06-10 | 河海大学常州校区 | Image segmentation threshold determination method based on visual analysis |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050254727A1 (en) * | 2004-05-14 | 2005-11-17 | Eastman Kodak Company | Method, apparatus and computer program product for determining image quality |
-
2017
- 2017-08-24 CN CN201710732837.9A patent/CN107610101B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102789737A (en) * | 2012-08-11 | 2012-11-21 | 中国人民解放军信息工程大学 | Method for establishing visual balance model of map on basis of moment balance principle |
| CN104700416A (en) * | 2015-03-23 | 2015-06-10 | 河海大学常州校区 | Image segmentation threshold determination method based on visual analysis |
Non-Patent Citations (1)
| Title |
|---|
| 马俊 等.采用力矩平衡原理建立地图视觉平衡模型.《武汉大学学报•信息科学版》.2013,第38卷(第1期), * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107610101A (en) | 2018-01-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110059694B (en) | Intelligent identification method for character data in complex scene of power industry | |
| WO2021000524A1 (en) | Hole protection cap detection method and apparatus, computer device and storage medium | |
| CN108052980B (en) | Image-based air quality grade detection method | |
| RU2008105393A (en) | COMPUTER-BASED APPLICATION FORMATION AND VERIFICATION OF TRAINING IMAGES INTENDED FOR MULTI-POINT THEOSTATISTIC ANALYSIS | |
| CN107157447B (en) | Skin surface roughness detection method based on image RGB color space | |
| CN108564085B (en) | Method for automatically reading of pointer type instrument | |
| CN109978854B (en) | Screen content image quality evaluation method based on edge and structural features | |
| CN110910343A (en) | Method and device for detecting pavement cracks and computer equipment | |
| CN112465748A (en) | Neural network based crack identification method, device, equipment and storage medium | |
| CN105528776B (en) | The quality evaluating method kept for the conspicuousness details of jpeg image format | |
| CN110210448B (en) | Intelligent face skin aging degree identification and evaluation method | |
| CN107743225B (en) | A method of it is characterized using multilayer depth and carries out non-reference picture prediction of quality | |
| CN111598869B (en) | Method, equipment and storage medium for detecting Mura of display screen | |
| CN104036493A (en) | No-reference image quality evaluation method based on multifractal spectrum | |
| JP2023511869A (en) | Method and system for automatic identification and grading of low multiplication acid etching defects by machine vision | |
| CN109191386B (en) | BPNN-based rapid Gamma correction method and device | |
| CN112712519A (en) | Non-contact type machine-made sandstone powder content intelligent real-time detection method and device | |
| CN114066857A (en) | Infrared image quality evaluation method and device, electronic equipment and readable storage medium | |
| CN115272826A (en) | Image identification method, device and system based on convolutional neural network | |
| CN111626358A (en) | Tunnel surrounding rock grading method based on BIM picture recognition | |
| CN114428110A (en) | Method and system for detecting defects of fluorescent magnetic powder inspection image of bearing ring | |
| CN107610101B (en) | Method for measuring visual balance quality of digital image | |
| CN114882002A (en) | Target defect detection method and detection device, computer equipment and storage medium | |
| CN110751170A (en) | Panel quality detection method, system, terminal device and computer readable medium | |
| CN110188662A (en) | A kind of AI intelligent identification Method of water meter number |
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 |