CN110111326A - Reconstructed image quality evaluation method based on ERT system - Google Patents
Reconstructed image quality evaluation method based on ERT system Download PDFInfo
- Publication number
- CN110111326A CN110111326A CN201910405133.XA CN201910405133A CN110111326A CN 110111326 A CN110111326 A CN 110111326A CN 201910405133 A CN201910405133 A CN 201910405133A CN 110111326 A CN110111326 A CN 110111326A
- Authority
- CN
- China
- Prior art keywords
- image
- entropy
- ert
- value
- gray level
- 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
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
- 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
- G06T2207/20048—Transform domain processing
- G06T2207/20052—Discrete cosine transform [DCT]
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of reconstructed image quality evaluation methods based on ERT system, comprising steps of the one, extraction to ERT image progress feature；Two, the flow pattern information of ERT image is extracted；Three, general space pixel grey scale entropy feature, pixel gradient entropy feature, frequency spectrum entropy feature and flow pattern value tag are put into training in KSVD model.Method and step of the invention is simple, it is convenient to realize, for ERT imaging system image generated, after extraction space pixel gray level entropy value, pixel gradient entropy and frequency spectrum entropy feature, the flow pattern features of ERT image are obtained using CNN model again, and all features are input to KSVD model and carry out quality evaluation prediction, precision of prediction can be greatly improved, it is practical, using effect is good, convenient for promoting the use of.
Description
Technical field
The invention belongs to image assessment techniques fields, and in particular to a kind of reconstructed image quality evaluation based on ERT system
Method.
Background technique
Image assessment techniques develop at home advances by leaps and bounds, many fields can apply well and effect also very
Good, most of existing technology is all based on existing image library, then (is based on spatial information entropy and spectrum information using SSEQ
The non-reference picture quality evaluation algorithm of entropy), SSGEQ (comment by the non-reference picture quality based on space, frequency spectrum and gradient information entropy
Valence algorithm) or other algorithms extract feature, then using SVM (Support Vector Machine, support vector machines) or
Person's rarefaction representation algorithm assesses image；But these methods can not be directed to ERT (Electrical Resistance
Tomography, Electrical Resistance Tomography) system reconstruction image generated, carry out on-line evaluation.It has been primarily present following
Problem:
1, the reconstruction image of ERT system, ununified standard picture library；For different applications, establishes data set and deposit
In larger difficulty.
2, there are apparent differences in the feature of frequency domain or spatial domain for ERT image and existing other types of image, increase
The difficulty of image quality measure is added.
3, using existing prediction homing method such as SVM or rarefaction representation algorithm, the ERT that SSEQ, SSGEQ algorithm are extracted
It is unsatisfactory that characteristics of image is trained prediction effect.
Summary of the invention
In view of the above-mentioned deficiencies in the prior art, the technical problem to be solved by the present invention is that providing a kind of based on ERT
The reconstructed image quality evaluation method of system, method and step is simple, and it is convenient to realize, for ERT imaging system figure generated
Picture obtains ERT after extraction space pixel gray level entropy value, pixel gradient entropy and frequency spectrum entropy feature, then using CNN model
The flow pattern features of image, and all features are input to KSVD model and carry out quality evaluation prediction, prediction essence can be greatly improved
Degree, practical, using effect is good, convenient for promoting the use of.
In order to solve the above technical problems, the technical solution adopted by the present invention is that: a kind of reconstruction image matter based on ERT system
Evaluation method is measured, method includes the following steps:
Step 1: carrying out the extraction of feature to ERT image: converting images into gray level image, and carry out piecemeal and adopted under
Sample processing, calculates the space pixel gray level entropy value of image, pixel gradient entropy and frequency spectrum entropy are as feature；
Step 2: extracting the flow pattern information of ERT image: writing CNN model using python language, ERT image is inputted
Into CNN model, the extraction of ERT image flow pattern features is carried out；
Step 3: general space pixel grey scale entropy feature, pixel gradient entropy feature, frequency spectrum entropy feature and flow pattern value
Feature is put into training in KSVD model: by 4 kinds of latent structures at an excessively complete dictionary, being put into KSVD model and is instructed
Practice, and calculates the forecast quality score of image using sparse coefficient.
The above-mentioned reconstructed image quality evaluation method based on ERT system, progress piecemeal and down-sampling described in step 1
When processing, it is the down-sampling for carrying out down-sampling rate and being 1/4 and 1/8,10 × 10 image block is divided into the ERT image after sampling.
The above-mentioned reconstructed image quality evaluation method based on ERT system calculates the aerial image of image described in step 1
When plain gray scale entropy, the calculation formula used is Es=-∑xp(x)log2P (x), wherein x is the pixel value of image block, p (x)
For the experienced probability distribution of x, EsFor the space pixel gray level entropy value of image；The space pixel ash of image is calculated described in step 1
It is to calculate the down-sampled images that gray level image, down-sampling rate are 1/4 and the down-sampled images that down-sampling rate is 1/8 when spending entropy
Space pixel gray level entropy mean value and the degree of bias, form 6 characteristic values of space space pixel grey scale entropy, i.e. 3 mean values and 3
Degree of bias value.
The above-mentioned reconstructed image quality evaluation method based on ERT system calculates pixel gradient entropy described in step 1
When, first using the method for laplacian spectral radius according to formula g (x, y)=f (x, y)+c [▽2F (x, y)] to original grayscale image
Picture f (x, y) is enhanced, and obtains enhanced image g (x, y), then use calculation formula Eg=-∑x∑yp[gm(x,y)]log2
p[gm(x, y)] calculate pixel gradient entropy Eg, wherein c is the coefficient of laplacian spectral radius enhancing, p [gm(x, y)] it is image ladder
Spend the experienced probability distribution of information, gm(x, y) be image block gradient magnitude andgx
(x, y) is the image gradient size after x orientation enhancement, gy(x, y) is the image gradient size after y orientation enhancement.
The above-mentioned reconstructed image quality evaluation method based on ERT system, when calculating frequency spectrum entropy described in step 1, first
It solves to obtain the experienced probability distribution p (i, j) of the DCT coefficient statistics of gray level image using dct2 function in MATLAB software,
Further according to formula Ef=-∑i∑jp(i,j)log2Frequency spectrum entropy E is calculated in p (i, j)f, wherein i is picture traverse pixel value
And 1 < i≤10, j is picture altitude pixel value and 1 j≤10 <, is to calculate gray scale when calculating frequency spectrum entropy described in step 1
The mean value of the frequency spectrum entropy for the down-sampled images that the down-sampled images and down-sampling rate that image, down-sampling rate are 1/4 are 1/8 and partially
Degree forms 6 characteristic values of frequency spectrum entropy, i.e. 3 mean values and 3 degree of bias values.
The above-mentioned reconstructed image quality evaluation method based on ERT system, CNN model described in step 2 is using 2 volumes
Lamination, convolution kernel are respectively 7 × 7 and 3 × 3,2 pond layers, and pond layer center size is unified for 2 × 2, then plus 2 full articulamentums,
Finally output result is 3 class of circulation, laminar flow and bubble flow.
The above-mentioned reconstructed image quality evaluation method based on ERT system is calculated described in step 3 using sparse coefficient
The forecast quality score of image out, the calculation formula of use are as follows:Wherein, domskFor k-th of word
The corresponding scoring of allusion quotation element, XkFor the product of k-th of dictionary element corresponding sparse coefficient and itself absolute value, N was complete
The total number of dictionary element, doms in dictionarypFor test image forecast quality score.
Compared with the prior art, the present invention has the following advantages:
1, method and step of the invention is simple, and it is convenient to realize.
2, design of the invention is extracting space pixel gray level entropy value for ERT imaging system image generated, as
After plain gradient entropy and frequency spectrum entropy feature, then the flow pattern features of ERT image are obtained using CNN model, and by all features
It is input to KSVD model and carries out quality evaluation prediction, precision of prediction can be greatly improved.
3, the construction method of excessively complete dictionary of the invention, atomic energy can be better with the feature structure of testing image
Match, more accurate forecast image quality score can be obtained.
4, of the invention practical, using effect is good, convenient for promoting the use of.
Below by drawings and examples, technical scheme of the present invention will be described in further detail.
Detailed description of the invention
Fig. 1 is method flow block diagram of the invention.
Fig. 2 is the simulation and prediction mass fraction matched curve comparison diagram of the method for the present invention with method in the prior art.
Specific embodiment
As shown in Figure 1, the reconstructed image quality evaluation method of the invention based on ERT system, comprising the following steps:
Step 1: carrying out the extraction of feature to ERT image: converting images into gray level image, and carry out piecemeal and adopted under
Sample processing, calculates the space pixel gray level entropy value of image, pixel gradient entropy and frequency spectrum entropy are as feature；
In the present embodiment, described in step 1 progress piecemeal and down-sampling handle when, be carry out down-sampling rate be 1/4 and 1/
8 down-sampling is divided into 10 × 10 image block to the ERT image after sampling.
In the present embodiment, when calculating the space pixel gray level entropy value of image described in step 1, the calculation formula that uses for
Es=-∑xp(x)log2P (x), wherein x is the pixel value of image block, and p (x) is the experienced probability distribution of x, EsFor the sky of image
Between pixel grey scale entropy；It is to calculate gray level image, down-sampling when calculating the space pixel gray level entropy value of image described in step 1
The mean value and the degree of bias of the space pixel gray level entropy for the down-sampled images that the down-sampled images and down-sampling rate that rate is 1/4 are 1/8, group
At 6 characteristic values of space space pixel grey scale entropy, i.e. 3 mean values and 3 degree of bias values.
When calculating pixel gradient entropy in the present embodiment, described in step 1, the method root of laplacian spectral radius is first used
According to formula g (x, y)=f (x, y)+c [▽2F (x, y)] original gray level image f (x, y) is enhanced, it obtains enhanced
Image g (x, y), then use calculation formula Eg=-∑x∑yp[gm(x,y)]log2 p[gm(x, y)] calculate pixel gradient entropy
Eg, wherein c is the coefficient of laplacian spectral radius enhancing, p [gm(x, y)] be image gradient information experienced probability distribution, gm(x,
Y) for image block gradient magnitude andgx(x, y) is the image ladder after x orientation enhancement
Spend size, gy(x, y) is the image gradient size after y orientation enhancement.
When it is implemented, described in step 1 calculate pixel gradient entropy when, be calculate gray level image, down-sampling rate be 1/
The mean value and the degree of bias of the pixel gradient entropy for the down-sampled images that 4 down-sampled images and down-sampling rate are 1/8 form pixel gradient
6 characteristic values of entropy, i.e. 3 mean values and 3 degree of bias values.
When calculating frequency spectrum entropy in the present embodiment, described in step 1, first asked in MATLAB software using dct2 function
Solution obtains the experienced probability distribution p (i, j) of the DCT coefficient statistics of gray level image, further according to formula Ef=-∑i∑jp(i,j)
log2Frequency spectrum entropy E is calculated in p (i, j)f, wherein i is picture traverse pixel value and 1 < i≤10, j is picture altitude pixel
Value and 1 j≤10 < are to calculate gray level image, down-sampling rate as 1/4 down-sampling when calculating frequency spectrum entropy described in step 1
The mean value and the degree of bias of the frequency spectrum entropy for the down-sampled images that image and down-sampling rate are 1/8,6 characteristic values of composition frequency spectrum entropy, i.e., 3
A mean value and 3 degree of bias values.
Step 2: extracting the flow pattern information of ERT image: writing CNN model using python language, ERT image is inputted
Into CNN model, the extraction of ERT image flow pattern features is carried out；
This model uses convolutional layer and dropout technology, can be very good to obtain the flow pattern features of ERT image；
In the present embodiment, CNN model described in step 2 uses 2 convolutional layers, and convolution kernel is respectively 7 × 7 and 3 × 3, and 2
A pond layer, pond layer center size are unified for 2 × 2, then plus 2 full articulamentums, finally exporting result is circulation, laminar flow and blister
Flow 3 classes.
Convolutional neural networks are one of most classic models in deep learning, it, which can use seldom weight and reaches, compares
Good effect.The main distinction of convolutional neural networks and general neural network is that convolutional neural networks include multiple by convolutional layer
The feature extractor constituted with pond layer.A neuron and part adjacent bed neuron in convolutional network, in convolutional layer
Connection, generally comprises several characteristic planes, these characteristic planes are the neural tuples by some rectangular arrangeds in convolutional layer
At in the same characteristic plane, neuron shares weight, and what shared weight referred to here is exactly convolution kernel, general volume
Product core is initialized in the form of random decimal matrix, and convolution kernel is reasonably weighed study in the training process of network
Value.Convolution kernel can be very good to reduce the connection between each layer of network, while can also reduce over-fitting risk.In convolutional Neural net
Further include sub-sampling in network, be also pond layer, effect is can to greatly simplify model complexity, reduces model parameter.
Step 3: general space pixel grey scale entropy feature, pixel gradient entropy feature, frequency spectrum entropy feature and flow pattern value
Feature is put into training in KSVD model: by 4 kinds of latent structures at an excessively complete dictionary, being put into KSVD model and is instructed
Practice, and calculates the forecast quality score of image using sparse coefficient.
KSVD model is the extension of K-means algorithm, mainly includes two steps: sparse coding and dictionary updating.
When it is implemented, the sparse coding of KSVD model uses OMP algorithm, dictionary updating uses SVD singular value decomposition method.Dictionary is more
It is newly that the primary update of dictionary is completed by K iteration using dictionary atom vector is gradually updated.It is excessively complete in the present embodiment
The constructive method of dictionary are as follows: space pixel gray level entropy value tag, pixel gradient entropy feature and each 6 dimension of frequency spectrum entropy feature, then
In addition the one-dimensional flow pattern features extracted constitute a dictionary element, each element is by row by each image feature by column arrangement
Complete dictionary was rearranged, the picture number of training set is much larger than the number of feature in excessively complete dictionary.Of the invention is excessively complete
The construction method of dictionary, atomic energy can preferably be matched with the feature structure of testing image, can obtain more accurate prediction
Picture quality scoring.
In the present embodiment, the forecast quality score of image, the meter of use are calculated described in step 3 using sparse coefficient
Calculate formula are as follows:Wherein, domskFor the corresponding scoring of k-th of dictionary element, XkFor k-th of dictionary
The product of element corresponding sparse coefficient and itself absolute value, N were the total number of dictionary element in complete dictionary, domspFor
Test image forecast quality score.
In order to verify the effect that the present invention can generate, in MATLAB software, to method and the prior art of the invention
In image evaluation method emulated, predict based on ERT system reconstructed image quality related coefficient fitting effect comparison
As shown in Figure 2；Obviously, figure it is seen that the prediction of the mentioned method (KSVD has flow pattern) of the present invention is more accurate.Completely may be used
To meet the requirement of the reconstructed image quality evaluation based on ERT system without reference.
It should be understood by those skilled in the art that, embodiments herein can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the application
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the application, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The application is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present application
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
The aforementioned description to specific exemplary embodiment of the invention is in order to illustrate and illustration purpose.These descriptions
It is not wishing to limit the invention to disclosed precise forms, and it will be apparent that according to the above instruction, can much be changed
And variation.The purpose of selecting and describing the exemplary embodiment is that explaining specific principle of the invention and its actually answering
With so that those skilled in the art can be realized and utilize a variety of different exemplary implementation schemes of the invention and
Various chooses and changes.The scope of the present invention is intended to be limited by claims and its equivalents.
Claims (7)
1. a kind of reconstructed image quality evaluation method based on ERT system, which is characterized in that method includes the following steps:
Step 1: carrying out the extraction of feature to ERT image: converting images into gray level image, and carry out at piecemeal and down-sampling
Reason, calculates the space pixel gray level entropy value of image, pixel gradient entropy and frequency spectrum entropy are as feature；
Step 2: extracting the flow pattern information of ERT image: writing CNN model using python language, ERT image is input to CNN
In model, the extraction of ERT image flow pattern features is carried out；
Step 3: general space pixel grey scale entropy feature, pixel gradient entropy feature, frequency spectrum entropy feature and flow pattern value tag
It is put into training in KSVD model: by 4 kinds of latent structures at an excessively complete dictionary, being put into KSVD model and be trained, and
The forecast quality score of image is calculated using sparse coefficient.
2. the reconstructed image quality evaluation method described in accordance with the claim 1 based on ERT system, it is characterised in that: step 1
Described in progress piecemeal and down-sampling handle when, be carry out down-sampling rate be 1/4 and 1/8 down-sampling, to after sampling ERT scheme
Image block as being divided into 10 × 10.
3. according to claim 2 based on the reconstructed image quality evaluation method of ERT system, it is characterised in that: step 1
Described in calculating image space pixel gray level entropy value when, the calculation formula used is Es=-∑xp(x)log2P (x), wherein x
For the pixel value of image block, p (x) is the experienced probability distribution of x, EsFor the space pixel gray level entropy value of image；Institute in step 1
State calculate image space pixel gray level entropy value when, be calculate gray level image, down-sampling rate be 1/4 down-sampled images and under adopt
The mean value and the degree of bias of the space pixel gray level entropy for the down-sampled images that sample rate is 1/8,6 of composition space space pixel grey scale entropy
Characteristic value, i.e. 3 mean values and 3 degree of bias values.
4. according to the reconstructed image quality evaluation method described in claim 1,2 or 3 based on ERT system, it is characterised in that: step
When calculating pixel gradient entropy described in rapid one, first using the method for laplacian spectral radius according to formula g (x, y)=f (x, y)+c
[▽2F (x, y)] original gray level image f (x, y) is enhanced, enhanced image g (x, y) is obtained, then public using calculating
Formula Eg=-∑x∑yp[gm(x,y)]log2p[gm(x, y)] calculate pixel gradient entropy Eg, wherein c is laplacian spectral radius increasing
Strong coefficient, p [gm(x, y)] be image gradient information experienced probability distribution, gm(x, y) be image block gradient magnitude andgx(x, y) is the image gradient size after x orientation enhancement, gy(x, y) is the increasing of the direction y
Image gradient size after strong.
5. according to the reconstructed image quality evaluation method described in claim 1,2 or 3 based on ERT system, it is characterised in that: step
When calculating frequency spectrum entropy described in rapid one, first solve to obtain the DCT system of gray level image using dct2 function in MATLAB software
The experienced probability distribution p (i, j) of number statistics, further according to formula Ef=-∑i∑jp(i,j)log2Frequency spectrum entropy is calculated in p (i, j)
Value Ef, wherein i is picture traverse pixel value and 1 < i≤10, j is picture altitude pixel value and 1 j≤10 <, institute in step 1
It states when calculating frequency spectrum entropy, is to calculate under down-sampled images that gray level image, down-sampling rate are 1/4 and down-sampling rate are 1/8
The mean value and the degree of bias of the frequency spectrum entropy of sampled images form 6 characteristic values of frequency spectrum entropy, i.e. 3 mean values and 3 degree of bias values.
6. the reconstructed image quality evaluation method described in accordance with the claim 1 based on ERT system, it is characterised in that: step 2
Described in CNN model use 2 convolutional layers, convolution kernel is respectively 7 × 7 and 3 × 3,2 pond layers, and pond layer center size is united
One be 2 × 2, then plus 2 full articulamentums, finally export result be 3 class of circulation, laminar flow and bubble flow.
7. the reconstructed image quality evaluation method described in accordance with the claim 1 based on ERT system, it is characterised in that: step 3
Described in the forecast quality score of image, the calculation formula of use are calculated using sparse coefficient are as follows:Wherein, domskFor the corresponding scoring of k-th of dictionary element, XkIt is corresponding for k-th of dictionary element
Sparse coefficient and itself absolute value product, N was the total number of dictionary element in complete dictionary, domspFor test image
Forecast quality score.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910405133.XA CN110111326B (en) | 2019-05-15 | 2019-05-15 | Reconstructed image quality evaluation method based on ERT system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910405133.XA CN110111326B (en) | 2019-05-15 | 2019-05-15 | Reconstructed image quality evaluation method based on ERT system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110111326A true CN110111326A (en) | 2019-08-09 |
| CN110111326B CN110111326B (en) | 2021-01-15 |
Family
ID=67490414
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910405133.XA Active CN110111326B (en) | 2019-05-15 | 2019-05-15 | Reconstructed image quality evaluation method based on ERT system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110111326B (en) |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI280516B (en) * | 2005-02-02 | 2007-05-01 | Mitac Int Corp | Method of measurement for image quality |
| US20100106473A1 (en) * | 2008-10-24 | 2010-04-29 | Amicus Therapeutics, Inc. | Multiple compartment dosing model |
| WO2011113169A1 (en) * | 2010-03-16 | 2011-09-22 | Swisstom Ag | Electrode for a scanning electrical impedance tomography device and a scanning electrical impedance tomography device |
| CN103487234A (en) * | 2013-10-07 | 2014-01-01 | 青岛科技大学 | Gas-liquid two-phase flow dynamics representation and identification method based on multi-scale arrangement entropy |
| CN103810704A (en) * | 2014-01-23 | 2014-05-21 | 西安电子科技大学 | SAR (synthetic aperture radar) image change detection method based on support vector machine and discriminative random field |
| CN104182962A (en) * | 2013-05-28 | 2014-12-03 | 腾讯科技（深圳）有限公司 | Picture definition evaluation method and device |
| US9406110B2 (en) * | 2014-01-30 | 2016-08-02 | Adobe Systems Incorporated | Cropping boundary simplicity |
| CN106934766A (en) * | 2017-03-15 | 2017-07-07 | 西安理工大学 | A kind of infrared image super resolution ratio reconstruction method based on rarefaction representation |
| CN107402116A (en) * | 2017-08-24 | 2017-11-28 | 中南大学 | A kind of Method for Discriminating Gas-liquid Two Phase Flow and detection means |
| US9894285B1 (en) * | 2015-10-01 | 2018-02-13 | Hrl Laboratories, Llc | Real-time auto exposure adjustment of camera using contrast entropy |
| CN108769671A (en) * | 2018-06-13 | 2018-11-06 | 天津大学 | Stereo image quality evaluation method based on adaptive blending image |
| CN109523513A (en) * | 2018-10-18 | 2019-03-26 | 天津大学 | Based on the sparse stereo image quality evaluation method for rebuilding color fusion image |
| WO2019070577A1 (en) * | 2017-10-02 | 2019-04-11 | Denali Therapeutics Inc. | Fusion proteins comprising enzyme replacement therapy enzymes |
| CN109685117A (en) * | 2018-12-04 | 2019-04-26 | 浙江大学 | The image measuring method of flow parameter in a kind of gas liquid particle three phase system |
| CN109727246A (en) * | 2019-01-26 | 2019-05-07 | 福州大学 | Comparative learning image quality evaluation method based on twin network |
-
2019
- 2019-05-15 CN CN201910405133.XA patent/CN110111326B/en active Active
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI280516B (en) * | 2005-02-02 | 2007-05-01 | Mitac Int Corp | Method of measurement for image quality |
| US20100106473A1 (en) * | 2008-10-24 | 2010-04-29 | Amicus Therapeutics, Inc. | Multiple compartment dosing model |
| WO2011113169A1 (en) * | 2010-03-16 | 2011-09-22 | Swisstom Ag | Electrode for a scanning electrical impedance tomography device and a scanning electrical impedance tomography device |
| CN104182962A (en) * | 2013-05-28 | 2014-12-03 | 腾讯科技（深圳）有限公司 | Picture definition evaluation method and device |
| CN103487234A (en) * | 2013-10-07 | 2014-01-01 | 青岛科技大学 | Gas-liquid two-phase flow dynamics representation and identification method based on multi-scale arrangement entropy |
| CN103810704A (en) * | 2014-01-23 | 2014-05-21 | 西安电子科技大学 | SAR (synthetic aperture radar) image change detection method based on support vector machine and discriminative random field |
| US9406110B2 (en) * | 2014-01-30 | 2016-08-02 | Adobe Systems Incorporated | Cropping boundary simplicity |
| US9894285B1 (en) * | 2015-10-01 | 2018-02-13 | Hrl Laboratories, Llc | Real-time auto exposure adjustment of camera using contrast entropy |
| CN106934766B (en) * | 2017-03-15 | 2020-04-21 | 西安理工大学 | Infrared image super-resolution reconstruction method based on sparse representation |
| CN106934766A (en) * | 2017-03-15 | 2017-07-07 | 西安理工大学 | A kind of infrared image super resolution ratio reconstruction method based on rarefaction representation |
| CN107402116A (en) * | 2017-08-24 | 2017-11-28 | 中南大学 | A kind of Method for Discriminating Gas-liquid Two Phase Flow and detection means |
| WO2019070577A1 (en) * | 2017-10-02 | 2019-04-11 | Denali Therapeutics Inc. | Fusion proteins comprising enzyme replacement therapy enzymes |
| CN108769671A (en) * | 2018-06-13 | 2018-11-06 | 天津大学 | Stereo image quality evaluation method based on adaptive blending image |
| CN109523513A (en) * | 2018-10-18 | 2019-03-26 | 天津大学 | Based on the sparse stereo image quality evaluation method for rebuilding color fusion image |
| CN109685117A (en) * | 2018-12-04 | 2019-04-26 | 浙江大学 | The image measuring method of flow parameter in a kind of gas liquid particle three phase system |
| CN109727246A (en) * | 2019-01-26 | 2019-05-07 | 福州大学 | Comparative learning image quality evaluation method based on twin network |
Non-Patent Citations (3)
| Title |
|---|
| WANG PAI,ET AL: "《An Adaptive electrical resistance tomography sensor with flow pattern recognition capability》", 《SPRINGER》 * |
| ZIQIANG CUI,ET AL: "《A review on image reconstruction algorithms for electrical capacitance/resistance tomography》", 《SENSOR REVIEW》 * |
| 赵进创，等: "《电容层析成像系统重建图像质量评价方法探讨》", 《广西大学学报（自然科学版）》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110111326B (en) | 2021-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Luus et al. | Multiview deep learning for land-use classification | |
| Zhong et al. | Squeeze-and-excitation wide residual networks in image classification | |
| CN109840322B (en) | Complete shape filling type reading understanding analysis model and method based on reinforcement learning | |
| CN108108751B (en) | Scene recognition method based on convolution multi-feature and deep random forest | |
| Yang et al. | Sex Determination of Three‐Dimensional Skull Based on Improved Backpropagation Neural Network | |
| CN107944556A (en) | Deep neural network compression method based on block item tensor resolution | |
| CN114841257B (en) | Small sample target detection method based on self-supervision comparison constraint | |
| CN107194378B (en) | Face recognition method and device based on mixed dictionary learning | |
| CN104200203B (en) | A kind of human action detection method based on action dictionary learning | |
| CN105550649B (en) | Extremely low resolution ratio face identification method and system based on unity couping local constraint representation | |
| CN106203625A (en) | A kind of deep-neural-network training method based on multiple pre-training | |
| CN109711401A (en) | A kind of Method for text detection in natural scene image based on Faster Rcnn | |
| CN106034264B (en) | Method for acquiring video abstract based on collaborative model | |
| CN109598676A (en) | A kind of single image super-resolution method based on Hadamard transform | |
| CN110309835A (en) | A kind of image local feature extracting method and device | |
| CN113066037B (en) | Multispectral and full-color image fusion method and system based on graph attention machine system | |
| CN110516724A (en) | Visualize the high-performance multilayer dictionary learning characteristic image processing method of operation scene | |
| Yu et al. | Modeling spatial extremes via ensemble-of-trees of pairwise copulas | |
| CN109918542A (en) | A kind of convolution classification method and system for relationship diagram data | |
| Shang et al. | Spatiotemporal reflectance fusion using a generative adversarial network | |
| CN110188621B (en) | Three-dimensional facial expression recognition method based on SSF-IL-CNN | |
| Manikandan et al. | Deepq: Residue analysis of localization images in large scale solid state physical environments | |
| Khattak et al. | High energy physics calorimeter detector simulation using generative adversarial networks with domain related constraints | |
| CN104573726B (en) | Facial image recognition method based on the quartering and each ingredient reconstructed error optimum combination | |
| Hu et al. | A noising-denoising framework for point cloud upsampling via normalizing flows |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Wang Pai Inventor after: Song Bo Inventor after: Li Jiaqing Inventor after: Qin Xuebin Inventor after: Li Yangbo Inventor after: Liu Lang Inventor after: Zhang Bo Inventor before: Wang Pai Inventor before: Song Bo Inventor before: Li Yangbo Inventor before: Qin Xuebin Inventor before: Li Jiaqing Inventor before: Liu Lang Inventor before: Zhang Bo |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |