CN108389173A - A kind of parameter optimization method based on opportunity cost - Google Patents
A kind of parameter optimization method based on opportunity cost Download PDFInfo
- Publication number
- CN108389173A CN108389173A CN201810248394.0A CN201810248394A CN108389173A CN 108389173 A CN108389173 A CN 108389173A CN 201810248394 A CN201810248394 A CN 201810248394A CN 108389173 A CN108389173 A CN 108389173A
- Authority
- CN
- China
- Prior art keywords
- psnr
- image
- opportunity cost
- parameter
- cost
- 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
- 238000000034 method Methods 0.000 title claims abstract description 11
- 238000005457 optimization Methods 0.000 title claims abstract description 10
- 238000004422 calculation algorithm Methods 0.000 description 14
- 230000002708 enhancing effect Effects 0.000 description 9
- 238000010586 diagram Methods 0.000 description 5
- 244000144977 poultry Species 0.000 description 5
- 235000013594 poultry meat Nutrition 0.000 description 5
- 238000005553 drilling Methods 0.000 description 4
- 241000272525 Anas platyrhynchos Species 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 238000012258 culturing Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 241000282887 Suidae Species 0.000 description 2
- 238000009795 derivation Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000002566 Capsicum Nutrition 0.000 description 1
- 241000287828 Gallus gallus Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 241000758706 Piperaceae Species 0.000 description 1
- 241000287531 Psittacidae Species 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 235000013330 chicken meat Nutrition 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- 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
- G06T2207/20172—Image enhancement details
Abstract
The present invention relates to a kind of parameter optimization methods based on opportunity cost, n width images are inputted first, then the PSNR values of different images under all parameter combinations are calculated, calculate the opportunity cost of value of each image, then the opportunity cost total value of all images is calculated, finally chooses the parameter combination of opportunity cost total value minimum as best parameter group.Opportunity cost thought in economics is referred to field of image enhancement by the present invention, selects best parameter group.
Description
Technical field
The present invention relates to the non-linear enhancing field of image, especially a kind of parameter optimization methods based on opportunity cost.
Background technology
Opportunity cost refer to when make some selection after, caused by have to abandon other analyses that may be selected
Journey.When both be forced to make take the decision of one when it is necessary to pay opportunity cost, the opportunity cost of a decision is that another can
The value of obtained best decision.Analysis on opportunity cost method focuses on the cost for analyzing selection.In life, some chances at
This usable currency is weighed.For example, peasant when obtaining more soils, cannot select to raise chickens if selection is raised pigs, pig raising
Opportunity cost is just to give up the income of poultry.But some opportunity costs can not often be weighed with currency, for example, reading a book in library
Study or enjoy TV play bring it is happy between selected.
And opportunity cost refers to all one of them maximum losses after making a choice, opportunity cost can be with the generation paid
Price-reform becomes and makes changes, such as the favorable rating of option that is rejected or is worth when making change, and obtained value is
Opportunity cost will not be enabled to change.Using following example as example calculation:Peasant is when obtaining more soils, if selection is supported
Pig cannot select to support other poultry, and the opportunity cost of pig raising is just to give up the income of poultry or duck culturing etc..Assuming that pig raising can be with
90,000 yuan are obtained, poultry can obtain 70,000 yuan, and duck culturing can obtain 80,000 yuan, then the opportunity cost raised pigs is 80,000 yuan, poultry
Opportunity cost be 90,000 yuan, the opportunity cost of duck culturing is also 90,000 yuan.
In information technology field, for the algorithm of non-one-parameter (number of parameters is two or more), all kinds of parameters
The selection of value is often a kind of tradeoff.The cost of selection optimized parameter is to give up the potential income of the second choice of opimization.For this reason
By being commonly used in economics and related fields.In information technology field, algorithm or framework involved by each link are often multi-parameter
Or multinomial technical indicator.Such as in the non-linear enhancing field of image, the low-frequency component that piece image can be expressed as image adds
Upper corresponding radio-frequency component, if then want to get to the enhanced image of original image, it can be by image G to be reinforced0
Regard low-frequency image as, if corresponding higher frequency image L can be constructed-1, superposition G0And L-1It will obtain comparing G0Clearer image
G-1(i.e. enhanced image), basic fundamental flow is as shown in Figure of description 1.To G0In high fdrequency component L0It carries out non-thread
Property filtering, can extrapolated and L0The consistent higher frequency image L of phase-1, to realize the enhancing to blurred picture high fdrequency component.
To G0In high frequency element L0Carry out secondary derivation, four derivations obtain L0" and L0" ", so by adjusting L0" with
L0" " ideal high fdrequency component L then can be more approached by formula (3-1)-1'sWherein k1With k2For amplitude adjusted parameter, use
This method can generate the high fdrequency component of needs
Next, the high fdrequency component that can will then obtainWith input picture G0It is added the result images that can be obtained enhancingI.e.:
When carrying out enhancing processing to two dimensional image, classical linear Laplacian high-pass filters can be used to height
Frequency component L0Ask second dervative, Fourth-Derivative.In experiment simulation, the filter can use Figure of description 2 shown in 3 ×
3Laplacian templates replace.
In this algorithm for image enhancement, parameter k1With k2Value final enhancing effect can be had an impact.To image
When quality carries out objective evaluation, reflect picture quality indeed through the size of the objective parameter image-related with this
Quality, the value for being used herein as Y-PSNR (Peak Signal Noise Ratio, PSNR) illustrate.In general,
When being compared to different enhancing algorithms, PSNR values are bigger, and the effect of image is better.PSNR's is defined as follows:
Wherein 2n- 1 is the pixel value of maximum possible in image, and n is the bit number needed for each pixel, and 8bit is indicated
Image, the value are 255.
By formula (3-3) if it can be seen from gm,n≈fm,n, then the value of PSNR level off to infinity, caused by after handling at this time
Distortion is smaller, and picture quality is preferable.
Invention content
In view of this, the purpose of the present invention is to propose to a kind of parameter optimization method based on opportunity cost, it will be in economics
Opportunity cost thought refer to field of image enhancement, select best parameter group.
The present invention is realized using following scheme:A kind of parameter optimization method based on opportunity cost, specifically includes following step
Suddenly:
Step S1:Given n width input pictures set the parameter for influencing image high fdrequency component as k1With k2;For each width figure
As for, each group of k1With k2Parameter combination can generate corresponding PSNR values, if the k that shared m groups are different1With k2Parameter
Combination;
Step S2:Calculate all k1With k2Combination under different images PSNR values, that is, calculate PSNR (i, j);Wherein, i tables
Show k1With k2Combination serial number, j indicates the serial number of current width image, 1≤i≤m, 1≤j≤n;
Step S3:Calculate the maximum PSNR values of each image, the wherein maximum PSNR values Optimal_PSNR of jth width image
(j) following formula is used to calculate:
Optimal_PSNR (j)=Max (PSNR (1:m,j));
Step S4:The opportunity cost of value of each image is calculated, the opportunity cost of value Cost (j) of wherein jth width image is adopted
It is calculated with following formula:
Cost (j)=| Optimal_PSNR (j)-PSNR (i, j) |;
Step S5:The opportunity cost total value Tcost of all images is calculated using following formula:
Step S6:Choose the k corresponding to i values and the i values corresponding when Tcost minimums1With k2Combination, and choosing should
k1With k2Combination best parameter group the most.
Further, in step S1, m≤n.
Compared with prior art, the present invention has following advantageous effect:
1, the opportunity cost in economics is applied to field of image enhancement by the present invention, can for multi-parameter algorithm
Obtain the optimized parameter on definite meaning.
2, algorithm of the invention can realize that the initialization of automatically parameter was chosen for the algorithm of multi-parameter class
Journey can finally carry out the subtle tuning of parameter according to specific algorithm.
Description of the drawings
Fig. 1 is the fundamental block diagram of nonlinear images enhanced scheme in background of invention.
Fig. 2 is linear high pass Laplacian filter template M schematic diagrames in background of invention.
Fig. 3 is the algorithm principle schematic diagram in the embodiment of the present invention.
Fig. 4 is test image schematic diagram in the embodiment of the present invention.
Fig. 5 is application scenario schematic diagram in the embodiment of the present invention.
Specific implementation mode
The present invention will be further described with reference to the accompanying drawings and embodiments.
A kind of parameter optimization method based on opportunity cost is present embodiments provided, in order to meet specific objective, when more
The advantage that other solutions are brought will be lost after making a choice in the different solutions of kind, and opportunity cost refers to selecting to be somebody's turn to do
Caused loss when scheme.For example, X is enabled1With X2Two images are indicated respectively, are selected for X2Best enhanced scheme
Corresponding opportunity cost cost (X will be will produce22), it also will produce for X under program situation1Opportunity cost cost
(X21).Similarly, selection is for X1Best enhanced scheme will will produce corresponding opportunity cost cost (X11), in the party
It also will produce for X under case situation2Opportunity cost cost (X12).By the opportunity cost phase of different images under identical solution
Add and can be obtained for image X1With X2For total opportunity cost, i.e. the chance totle drilling cost Tcost of scheme 11And scheme 2
Chance totle drilling cost Tcost2, compare their size, select numerical value therein smaller one group can be obtained for enhancing figure
As X1With X2For the optimal solution based on opportunity cost.In upper example, opportunity cost cost (Xij) in subscript release
Justice is:First subscript i indicates the optimal solution for the width image, second subscript j indicate under this scenario pair
The image answered.
Assuming that given n (n > 1) width input picture, for every piece image, each group of k1With k2Parameter combination value
Corresponding PSNR values can be generated, we can easily find out optimal corresponding to maximum PSNR values in these PSNR values
Parameter combination value.Assuming that there is m group parameter combination values, it is therefore apparent that m≤n, this is because the best parameter group of some images is
The same.In this m group parameter combination, m opportunity cost of value will be had per piece image, it will be different under each group of parameter combination
The opportunity cost addition of image can then obtain m opportunity cost total value Tcost, corresponding ginseng when selecting to take Tcost minimum values
Number combined value is best parameter group value, and algorithm specific example is as shown in Figure 3.
Specifically include following steps:
Step S1:Given n width input pictures set the parameter for influencing image high fdrequency component as k1With k2;For each width figure
As for, each group of k1With k2Parameter combination can generate corresponding PSNR values, if the k that shared m groups are different1With k2Parameter
Combination;
Step S2:Calculate all k1With k2Combination under different images PSNR values, that is, calculate PSNR (i, j);Wherein, i tables
Show k1With k2Combination serial number, j indicates the serial number of current width image, 1≤i≤m, 1≤j≤n;
Step S3:Calculate the maximum PSNR values of each image, the wherein maximum PSNR values Optimal_PSNR of jth width image
(j) following formula is used to calculate:
Optimal_PSNR (j)=Max (PSNR (1:m,j));
Step S4:The opportunity cost of value of each image is calculated, the opportunity cost of value Cost (j) of wherein jth width image is adopted
It is calculated with following formula:
Cost (j)=| Optimal_PSNR (j)-PSNR (i, j) |;
Step S5:The opportunity cost total value Tcost of all images is calculated using following formula:
Step S6:Choose the k corresponding to i values and the i values corresponding when Tcost minimums1With k2Combination, and choosing should
k1With k2Combination best parameter group the most.
In the present embodiment, in step S1, m≤n.
Since optimal parameter combination is difficult to intuitively obtain, we by virtue of experience define k1With k2Parameter value range.
In emulation experiment, k1With k2Value condition be respectively k1=0.05,0.10,0.15 ..., 1 and k2=0.1,0.2,
0.3,...,1.The normal grayscale image that 12 width sizes are 512 × 512 is chosen, as shown in Figure 4.
PSNR (dB) value analysis under 1 different images different parameters of table
2 difference k of table1、k2Opportunity cost corresponding to parameter combination and totle drilling cost
X1~X2 difference correspondence images Aerial, Airplane, Bike, Couple, Crowd in table 1 and table 2,
Flintstones, Elaine, Lena, Parrot, Peppers, Stream and bridge and Tank.Table 1 show 12 width
PSNR value of the image under respective best parameter group value, it can be seen that selected 12 width image shares 3 groups of best parameter groups
It is worth option, respectively k1=0.05 and k2=0.3, k1=0.05 and k2=0.4 and k1=0.05 and k2=0.5.Table 2 is shown
Different k1With k2Opportunity cost corresponding to best parameter group and totle drilling cost, it can be clearly seen that parameter combination value k1=0.05
With k2Opportunity cost total value Tcost corresponding to=0.4 is minimum, therefore this group of parameter combination is optimal.
PSNR (dB) performance of 3 enhancing algorithms of table compares
For the present embodiment algorithm, when using the parameter optimization algorithm of opportunity cost, it can be seen from Table 3 that k1=
0.05,k2=0.4 is adapted for the best parameter group of all test images, the opportunity cost total value Tcost corresponding to the combination
Not only minimum (being shown in Table 2), and it is also more excellent for the PSNR values of test image.
The present embodiment can be applied to non-linear enhancing algorithm field, as shown in Figure 5.
The foregoing is merely presently preferred embodiments of the present invention, all equivalent changes done according to scope of the present invention patent with
Modification should all belong to the covering scope of the present invention.
Claims (2)
1. a kind of parameter optimization method based on opportunity cost, it is characterised in that:Include the following steps:
Step S1:Given n width input pictures set the parameter for influencing image high fdrequency component as k1With k2;Every piece image is come
It says, each group of k1With k2Parameter combination can generate corresponding PSNR values, if the k that shared m groups are different1With k2Parameter combination;
Step S2:Calculate all k1With k2Combination under different images PSNR values, that is, calculate PSNR (i, j);Wherein, i indicates k1
With k2Combination serial number, j indicates the serial number of current width image, 1≤i≤m, 1≤j≤n;
Step S3:Calculate the maximum PSNR values of each image, the wherein maximum PSNR value Optimal_PSNR (j) of jth width image
It is calculated using following formula:
Optimal_PSNR (j)=Max (PSNR (1:m,j));
Step S4:The opportunity cost of value of each image is calculated, under wherein the opportunity cost of value Cost (j) of jth width image is used
Formula calculates:
Cost (j)=| Optimal_PSNR (j)-PSNR (i, j) |;
Step S5:The opportunity cost total value Tcost of all images is calculated using following formula:
Step S6:Choose the k corresponding to i values and the i values corresponding when Tcost minimums1With k2Combination, and choose the k1With
k2Combination best parameter group the most.
2. a kind of parameter optimization method based on opportunity cost according to claim 1, it is characterised in that:In step S1, m
≤n。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810248394.0A CN108389173B (en) | 2018-03-24 | 2018-03-24 | Opportunity cost-based parameter optimization method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810248394.0A CN108389173B (en) | 2018-03-24 | 2018-03-24 | Opportunity cost-based parameter optimization method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108389173A true CN108389173A (en) | 2018-08-10 |
CN108389173B CN108389173B (en) | 2021-08-31 |
Family
ID=63066788
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810248394.0A Active CN108389173B (en) | 2018-03-24 | 2018-03-24 | Opportunity cost-based parameter optimization method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108389173B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104200434A (en) * | 2014-08-28 | 2014-12-10 | 哈尔滨工程大学 | Non-local mean image denoising method based on noise variance estimation |
CN104392446A (en) * | 2014-11-20 | 2015-03-04 | 江南大学 | Improved PSNR (Peak Signal to Noise Ratio)-based DCT (Discrete Cosine Transformation) domain non-reference blurred image quality evaluation method |
CN104598987A (en) * | 2014-12-16 | 2015-05-06 | 南京华苏科技股份有限公司 | Method for predicating offline tendency and probability of mobile subscriber through learning and network effects of social networking services |
US9456086B1 (en) * | 2003-03-07 | 2016-09-27 | Wai Wu | Method and system for matching entities in an auction |
CN106972488A (en) * | 2017-05-12 | 2017-07-21 | 国网江西省电力公司经济技术研究院 | A kind of life cycle management programming screening method for considering essential safety |
CN107563995A (en) * | 2017-08-14 | 2018-01-09 | 华南理工大学 | A kind of confrontation network method of more arbiter error-duration models |
-
2018
- 2018-03-24 CN CN201810248394.0A patent/CN108389173B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9456086B1 (en) * | 2003-03-07 | 2016-09-27 | Wai Wu | Method and system for matching entities in an auction |
CN104200434A (en) * | 2014-08-28 | 2014-12-10 | 哈尔滨工程大学 | Non-local mean image denoising method based on noise variance estimation |
CN104392446A (en) * | 2014-11-20 | 2015-03-04 | 江南大学 | Improved PSNR (Peak Signal to Noise Ratio)-based DCT (Discrete Cosine Transformation) domain non-reference blurred image quality evaluation method |
CN104598987A (en) * | 2014-12-16 | 2015-05-06 | 南京华苏科技股份有限公司 | Method for predicating offline tendency and probability of mobile subscriber through learning and network effects of social networking services |
CN106972488A (en) * | 2017-05-12 | 2017-07-21 | 国网江西省电力公司经济技术研究院 | A kind of life cycle management programming screening method for considering essential safety |
CN107563995A (en) * | 2017-08-14 | 2018-01-09 | 华南理工大学 | A kind of confrontation network method of more arbiter error-duration models |
Non-Patent Citations (3)
Title |
---|
DANIEL A等: "Opportunity Costs and Non-Scale Free Capabilities: Profit Maximization, Corporate Scope, and Profit margins", 《UNIVERSITY OF PENNSYLVANIA SCHOLARLY COMMONS》 * |
杨鹏: "基于运动矢量相位角和卷积码的大容量视频隐写算法", 《计算机应用》 * |
简佩茹: "农村民间借贷的形成机制和外部性影响分析及监管", 《财政监督》 * |
Also Published As
Publication number | Publication date |
---|---|
CN108389173B (en) | 2021-08-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108932697B (en) | Distortion removing method and device for distorted image and electronic equipment | |
Xiao et al. | Brightness and contrast controllable image enhancement based on histogram specification | |
Ni et al. | Gradient direction for screen content image quality assessment | |
CN104658001B (en) | Non-reference asymmetric distorted stereo image objective quality assessment method | |
CN106295682A (en) | A kind of judge the method for the picture quality factor, device and calculating equipment | |
CN105574901B (en) | A kind of general non-reference picture quality appraisement method based on local contrast pattern | |
CN102063701A (en) | Image enhancement method and apparatuses | |
CN102547368B (en) | Objective evaluation method for quality of stereo images | |
CN107481209B (en) | Image or video quality enhancement method based on convolutional neural network | |
CN105894522B (en) | A kind of more distortion objective evaluation method for quality of stereo images | |
CN111383173B (en) | Baseline-based image super-resolution reconstruction method and system | |
Kim et al. | Multiple level feature-based universal blind image quality assessment model | |
US20160247259A1 (en) | Image processing apparatus and image processing method | |
CN110120034B (en) | Image quality evaluation method related to visual perception | |
CN109726732A (en) | The system and method for stream for processes data values | |
CN109729261A (en) | The system and method for stream for processes data values | |
Esmaeilzehi et al. | SRNSSI: a deep light-weight network for single image super resolution using spatial and spectral information | |
CN102306378A (en) | Image enhancement method | |
CN109727182A (en) | The system and method for stream for processes data values | |
CN111986275A (en) | Inverse halftoning method for multi-modal halftone image | |
CN103871035B (en) | Image denoising method and device | |
Wang et al. | Adaptor: Improving the robustness and imperceptibility of watermarking by the adaptive strength factor | |
CN108389173A (en) | A kind of parameter optimization method based on opportunity cost | |
CN114067006B (en) | Screen content image quality evaluation method based on discrete cosine transform | |
Fu et al. | Image quality assessment using edge and contrast similarity |
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 |