CN109146856A - Picture quality assessment method, device, computer equipment and storage medium - Google Patents
Picture quality assessment method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN109146856A CN109146856A CN201810869434.3A CN201810869434A CN109146856A CN 109146856 A CN109146856 A CN 109146856A CN 201810869434 A CN201810869434 A CN 201810869434A CN 109146856 A CN109146856 A CN 109146856A
- Authority
- CN
- China
- Prior art keywords
- image
- score
- training
- deep learning
- picture quality
- 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
- 238000001303 quality assessment method Methods 0.000 title claims abstract description 27
- 238000012549 training Methods 0.000 claims abstract description 130
- 238000013136 deep learning model Methods 0.000 claims abstract description 46
- 239000013598 vector Substances 0.000 claims abstract description 46
- 238000011156 evaluation Methods 0.000 claims abstract description 38
- 238000013135 deep learning Methods 0.000 claims abstract description 31
- 238000004458 analytical method Methods 0.000 claims abstract description 12
- 238000012360 testing method Methods 0.000 claims description 44
- 238000004590 computer program Methods 0.000 claims description 17
- 238000013441 quality evaluation Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 10
- 238000010191 image analysis Methods 0.000 claims 1
- 238000000034 method Methods 0.000 abstract description 19
- 239000000284 extract Substances 0.000 abstract description 8
- 238000005286 illumination Methods 0.000 abstract description 6
- 239000000203 mixture Substances 0.000 abstract description 6
- 230000015572 biosynthetic process Effects 0.000 abstract description 3
- 238000003786 synthesis reaction Methods 0.000 abstract description 3
- 238000013527 convolutional neural network Methods 0.000 description 28
- 230000006870 function Effects 0.000 description 17
- 238000010586 diagram Methods 0.000 description 11
- 230000001815 facial effect Effects 0.000 description 8
- 235000013399 edible fruits Nutrition 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 241001156002 Anthonomus pomorum Species 0.000 description 1
- 101100517192 Arabidopsis thaliana NRPD1 gene Proteins 0.000 description 1
- 101150094905 SMD2 gene Proteins 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to picture quality assessment method, device, computer equipment and storage medium, this method includes obtaining image；Clarity analysis is carried out to image, to obtain image definition score；Deep learning is carried out to image using deep learning model, and Score on Prediction is carried out to learning outcome, to obtain image quality score；Synthetic image articulation score and image quality score form evaluation result；Export evaluation result.The present invention is by first scoring to training image from illumination, size and deflection angle, training image is subjected to quality score, utilize the training set for the feature vector composition that training image extracts, the mass fraction of training image is trained adjustment to CNN sorter network, to obtain deep learning mode, when use, image is subjected to clarity scoring and deep learning model scores, synthesis is weighted to two scores again, obtain evaluation result, accuracy is determined to improve picture quality, improves the accuracy rate of true environment human face identification.
Description
Technical field
The present invention relates to image evaluation methods, more specifically refer to picture quality assessment method, device, computer equipment
And storage medium.
Background technique
With the continuous development of deep learning and artificial intelligence technology, face identification system is increasingly used in pacifying
Anti-, intelligence new retail, finance, subway, hotel, airport etc. need in the scene of authentication, and such as bank remotely opens an account, nobody
Super quotient is automatically performed that payment, access control system, subway brush face is paid, airport carries out testimony of a witness veritification etc. automatically by brush face.
Existing recognition of face is directly to obtain the face captured from camera to carry out recognition of face mostly, is directly captured
Picture contains many motion blurs or focuses unsharp face, and also many facial angles, scale and illumination is improper,
Recognition accuracy is leveraged, also face recognition application is caused to be a greater impact, although existing minority face identification system
Also there is the judgement of face picture quality, but be substantially and judged in terms of picture clarity, to facial angle, illumination, rotation
Turn, scale etc. is not assessed, while the judgement of clarity contains many subjective assessments, and face recognition accuracy rate is not
Height, therefore existing face picture Environmental Evaluation Model is poor to the evaluation effect of face picture quality.
Therefore, it is necessary to design a kind of new method, the accuracy for improving picture quality judgement is realized, greatly improve true
The accuracy rate of real environment human face identification.
Summary of the invention
It is an object of the invention to overcome the deficiencies of existing technologies, picture quality assessment method, device, computer are provided and set
Standby and storage medium.
To achieve the above object, the invention adopts the following technical scheme: picture quality assessment method, comprising:
Obtain image；
Clarity analysis is carried out to image, to obtain image definition score；
Deep learning is carried out to image using deep learning model, and Score on Prediction is carried out to learning outcome, to obtain figure
As mass fraction；
Synthetic image articulation score and image quality score form evaluation result；
Export evaluation result.
Its further technical solution are as follows: it is described that deep learning is carried out to image using deep learning model, to obtain image
Before mass fraction, further includes:
Obtain several training images；
Quality score is carried out to training image and is marked；
Training image after mark is trained, to obtain deep learning model.
Its further technical solution are as follows: the training image after described pair of mark is trained, to obtain deep learning model,
Include:
Whether the size of training of judgement image meets preset requirement；
If it is not, then carrying out size processing to training image, and enter next step, if so, into next step；
Several training image blocks are screened from training image；
The corresponding feature vector of training image blocks is extracted, and establishes training set；
Using training set and training image training CNN sorter network, to obtain deep learning model.
Its further technical solution are as follows: it is described that clarity analysis is carried out to image, to obtain image definition score, packet
It includes:
Energy gradient sharpness computation is carried out to image, to obtain calculated result；
The calculated result is normalized, to obtain image definition score.
Its further technical solution are as follows: it is described that deep learning is carried out to image using deep learning model, and study is tied
Fruit carries out Score on Prediction, to obtain image quality score, comprising:
Several testing image blocks are extracted to image；
Several testing image blocks are inputted into deep learning model, to extract the feature vector of several testing image blocks；
Predict to the feature vector of several testing image blocks the mass fraction of several testing image blocks；
The average value of the mass fraction of several testing image blocks is obtained, to obtain image quality score.
Its further technical solution are as follows: the synthetic image articulation score and image quality score form evaluation knot
Fruit, comprising:
Image definition score and image quality score are weighted summation according to specified weight, to form evaluation knot
Fruit.
The present invention also provides picture quality evaluation devices, comprising:
Image acquisition unit, for obtaining image；
Clarity analytical unit, for carrying out clarity analysis to image, to obtain image definition score；
Deep learning unit for carrying out deep learning to image using deep learning model, and carries out learning outcome
Score on Prediction, to obtain image quality score；
Comprehensive unit is used for synthetic image articulation score and image quality score, forms evaluation result；
As a result output unit, for exporting evaluation result.
Its further technical solution are as follows: the clarity analytical unit includes:
Sharpness computation subelement, for carrying out energy gradient sharpness computation to image, to obtain calculated result；
Normalizing subelement, for normalizing the calculated result, to obtain image definition score.
The present invention also provides a kind of computer equipment, including memory, processor and it is stored on the memory simultaneously
The computer program that can be run on the processor, the processor realize above-mentioned image when executing the computer program
Performance rating method.
The present invention also provides a kind of storage medium, the storage medium is stored with computer program, the computer journey
Sequence includes program instruction, and described program instruction makes the processor execute above-mentioned picture quality evaluation when being executed by a processor
Method.
Compared with the prior art, the invention has the advantages that: the present invention by first to training image from illumination, size and
Deflection angle scores, and training image is carried out quality score, utilizes the training for the feature vector composition that training image extracts
Collection, the mass fraction of training image is trained adjustment to CNN sorter network, to obtain deep learning mode, in use, will figure
It scores as carrying out clarity scoring and deep learning model, then synthesis is weighted to two scores, commented to obtain
Determine as a result, greatly improving the accuracy rate of true environment human face identification to improve the accuracy of picture quality judgement.
The invention will be further described in the following with reference to the drawings and specific embodiments.
Detailed description of the invention
Fig. 1 is the application scenario diagram for the picture quality assessment method that the specific embodiment of the invention provides；
Fig. 2 is the flow diagram for the picture quality assessment method that the specific embodiment of the invention provides；
Fig. 3 is the sub-process schematic diagram for the picture quality assessment method that the specific embodiment of the invention provides；
Fig. 4 is the sub-process schematic diagram for the picture quality assessment method that the specific embodiment of the invention provides；
Fig. 5 is the sub-process schematic diagram for the picture quality assessment method that the specific embodiment of the invention provides；
Fig. 6 is the schematic block diagram for the picture quality evaluation device that the specific embodiment of the invention provides；
Fig. 7 is the schematic block diagram for the picture quality evaluation device that another specific embodiment of the present invention provides；
Fig. 8 is the schematic frame of the model acquiring unit for the picture quality evaluation device that the specific embodiment of the invention provides
Figure；
Fig. 9 is the schematic frame of the clarity analytical unit for the picture quality evaluation device that the specific embodiment of the invention provides
Figure；
Figure 10 is the schematic frame of the deep learning unit for the picture quality evaluation device that the specific embodiment of the invention provides
Figure；
Figure 11 is a kind of schematic block diagram for computer equipment that the specific embodiment of the invention provides.
Specific embodiment
In order to more fully understand technology contents of the invention, combined with specific embodiments below to technical solution of the present invention into
One step introduction and explanation, but not limited to this.
It should be appreciated that ought use in this specification and in the appended claims, term " includes " and "comprising" instruction
Described feature, entirety, step, operation, the presence of element and/or component, but one or more of the other feature, whole is not precluded
Body, step, operation, the presence or addition of element, component and/or its set.
It is also understood that mesh of the term used in this present specification merely for the sake of description specific embodiment
And be not intended to limit the application.As present specification and it is used in the attached claims, unless on
Other situations are hereafter clearly indicated, otherwise " one " of singular, "one" and "the" are intended to include plural form.
It will be further appreciated that the term "and/or" used in present specification and the appended claims is
Refer to any combination and all possible combinations of one or more of associated item listed, and including these combinations.
Fig. 1 and Fig. 2 are please referred to, Fig. 1 is that the application scenarios of picture quality assessment method provided in an embodiment of the present invention are illustrated
Figure.Fig. 2 is the schematic flow chart of picture quality assessment method provided in an embodiment of the present invention.The picture quality assessment method can
Applied in server 20, exist in the form of picture quality evaluates platform, which can be counted with user terminal 10
According to interaction.Wherein, the user of user terminal 10 is usually a certain user using platform, such as the user of payment platform,
Facial image can be sent to server 20 by the judgement APP of user terminal 10, the face figure based on the received of server 20
As and to terminal feeds back corresponding evaluation result.
In addition, above-mentioned picture quality assessment method can also be applied to mobile terminal, with the shape of picture quality evaluation APP
Formula exists, and after user evaluates APP acquisition facial image by picture quality, is lived inside mobile terminal according to facial image
Physical examination is surveyed, and shows corresponding evaluation result.
As shown in Fig. 2, the picture quality assessment method of the embodiment of the present invention comprising steps of
S110, several training images are obtained.
In the present embodiment, above-mentioned training image is referred to for training to obtain the data of deep learning model, one
As using camera obtain image.
S120, quality score is carried out to training image and is marked.
In the present embodiment, specifically the face deflection angle in training image, human face light, facial size are beaten
Point, it is divided into five classes: Bad [0,20], Poor [20,40], Fair [40,60], Good in the section 0-100 according to quality when marking
[60,80], Excellent [80,100].First give a mark in the section 0-100 according to corresponding five grades to face deflection angle,
It gives a mark in the section 0-100 according to corresponding five grades to human face light, to facial size in the section 0-100 according to corresponding five
A grade marking, finally calculate according to marking the quality score and mark of training image, which is also classified into Bad
[0,20], Poor [20,40], Fair [40,60], Good [60,80], five classifications of Excellent [80,100].
It gives a mark to training image from illumination, size and deflection angle, this three aspect is also included in image evaluation process
In, it realizes the accuracy for improving picture quality judgement, greatly improves the accuracy rate of true environment human face identification.
S130, the training image after mark is trained, to obtain deep learning model.
In the present embodiment, deep learning model refer to using training image training CNN (convolutional neural networks,
Convolutional Neural Network) sorter network, a flight data recorder is obtained with this.
In one embodiment, as shown in figure 3, above-mentioned step S130 may include S131~S135.
S131, training of judgement image size whether meet preset requirement.
In the present embodiment, preset requirement refers to that the size of image is 224mm*224mm.
S132, if it is not, then carrying out size processing to training image, and S133 is entered step, if so, entering step
S133。
The size of training image is not 224mm*224mm, it is necessary to be processed to input picture, that is, from training image
In take the training image blocks of 224mm*224mm at random as the training image for input.
S133, several training image blocks are screened from training image.
In the present embodiment, training image blocks refer to that, when the size of training image meets preset requirement, training image is
It after then being taken to training image, is formed several for training image blocks when the size of training image is unsatisfactory for preset requirement
Image block, several image blocks i.e. training image blocks screened.
In order to guarantee trained performance, the training of multiple 224mm*224mm is selected at random from every training image at random
Image block, in the present embodiment, setting select 10.
S134, the corresponding feature vector of training image blocks is extracted, and establishes training set.
Specifically, it is carried out using CNN (convolutional neural networks, Convolutional Neural Network) sorter network
The model that deep learning training obtains can automatically extract feature.
In the present embodiment, feature vector refer to a certain class object be different from other class objects corresponding (essence) feature or
The set of characteristic or these features and characteristic.It is characterized in by measuring or handling the data that can be extracted.For image
Speech, every piece image all have the unique characteristics that can be different from other class images, some are oneself that can be perceive intuitively that
Right feature, such as brightness, edge, texture and color；Some are then that needs are just getable by transformation or processing, such as square, directly
Side's figure and main composition etc..The multiple or multifrequency nature of a certain class object is combined, a feature vector is formed and carrys out generation
The table class object, if only single number feature, feature vector is an one-dimensional vector, if it is the combination of n characteristic,
It is then a n dimensional feature vector.Input of such feature vector often as identifying system.In fact, a n dimensional feature is exactly
Point in n-dimensional space, and identify that the task of classification is exactly a kind of division found to this n-dimensional space.
The feature vector of each training image blocks is obtained, extracts maximum, minimum and the average value group of 10 randomized blocks respectively
A new vector f eature pooling (min, average, max) is synthesized, i.e., (maximum value, average value are minimum for feature pool
Value), all feature pools are combined into a vector, for indicating full figure.
One feature vector is corresponding with for the quality comprehensive score of each training image, is established with this feature vector
Training set can be used as the feature vector of a standard, for training CNN sorter network.
S135, CNN sorter network is trained using training set and training image, to obtain deep learning model.
In the present embodiment, CNN sorter network refers to the network that image classification is carried out using convolutional neural networks.
Above-mentioned Bad [0,20], Poor [20,40], Fair [40,60], Good [60,80], Excellent [80,100]
Training image in five classifications has the feature vector of standard.
When training CNN sorter network, training image is input to CNN sorter network, this training image is classified by CNN
Each one group of vector of layer process of convolution output quantity of network, but for completely by CCN sorter network that stochastic filtering device constructs and
Speech, output think the training image equiprobability be the corresponding a certain standard of training image of above-mentioned five classifications feature to
Amount, then can define a loss function, such as common MSE (mean square error, mean squared error), it is assumed that L at this time
L output valve is constantly fed back into entire CCN when training set being inputted CNN sorter network every time for the output of the loss function
Sorter network, so that penalty values L does small, that is, allows each convolutional layer to modify the weight of the filter of each layer convolution
Filter can combine the specific mode of detection of optimization, so as to form deep learning model, so that each training
Image by CCN sorter network extract feature vector and its corresponding to standard feature vector gap meet set want
It asks.
Training CNN sorter network is specifically the filter of each convolutional layer in training CNN sorter network, these is allowed to filter
Wave device group has high activation to specific mode, to reach the classification purpose of CNN sorter network.
Feature is extracted with trained CNN sorter network, CNN deep learning network automatically extracts feature, also
It is to say that deep learning is equivalent to a flight data recorder, sample is thrown into, is obtained in output as a result, obtaining the spy for belonging to the sample
Vector is levied, intermediate process is automatically performed by deep learning.
S140, image is obtained.
In the present embodiment, the facial image for needing to determine from the evaluation APP input of the picture quality of user terminal 10, with this
Image to be determined as input carries out scoring judgement to the image.
S150, clarity analysis is carried out to image, to obtain image definition score.
In the quality evaluation of non-reference picture, the clarity of image is to measure the important indicator of picture quality superiority and inferiority, it
Can be preferably corresponding with the subjective feeling of people, the clarity of image is not high to show the fuzzy of image.
In one embodiment, as shown in figure 4, above-mentioned steps S150 may include having S151~S152.
S151, energy gradient sharpness computation is carried out to image, to obtain calculated result；
S152, the normalization calculated result, to obtain image definition score.
In the present embodiment, above-mentioned image definition score refers to that the clarity according to image is given a mark, and is formed
Score.Specifically, Energy energy gradient definition judgment carried out to the image of input, Energy gradient function such as:Here D (f) is to be based on
The face picture score that Energy power gradient obtains, x and y are the pixel coordinate of image respectively.According to calculated result, and will
Obtained gradient energy value normalizes within the scope of 1-100.Energy gradient function is more suitable for Real-Time Evaluation image definition.
Certainly, in other embodiments, can also using entropy function, Brenner gradient function, Tenengrad gradient function,
Laplacian gradient function, SMD (Sum Modul us Difference, gray variance) function, (gray variance multiplies SMD2
Product) modes such as function and variance function calculate image definition score.
In conjunction with the analysis of image definition, the accuracy of picture quality judgement can be improved, greatly improved under true environment
The accuracy rate of recognition of face.
S160, deep learning is carried out to image using deep learning model, to obtain image quality score.
In one embodiment, as shown in figure 5, above-mentioned steps S160 may include S161~S164.
S161, several testing image blocks are extracted to image.
In the present embodiment, testing image block is taken at random having a size of 224mm*224mm's by the image progress inputted
Image block.Image is extracted to the testing image block of several 224mm*224mm sizes at random；Can set herein selection 10 to
Altimetric image block can wait other numbers in the number of other embodiments, the testing image block for 15.
S162, several testing image blocks are inputted into deep learning model, to extract the feature of several testing image blocks
Vector.
10 testing image blocks are sequentially placed into the deep learning model that training obtains, utilize deep learning model extraction
The feature vector of several testing image blocks out.
S163, the mass fraction for predict to the feature vector of several testing image blocks several testing image blocks.
Specifically, feature vector CNN sorter network extracted uses the mass fraction carried out with SVM prediction mode pre-
It surveys, the score that prediction is obtained is as the quality score of face picture.
The full articulamentum of CNN sorter network exports 256 dimensional feature vectors；SVM calls directly the open source library libsvm, uses
RBF core, it is that scikit-learn is arranged automatically that C, which takes 0.9, Gamma,.256 dimensional feature vectors that CNN sorter network is exported as
Obtained mass fraction is normalized to 0-100 by the input feature value of SVM.
The open source library libsvm is called to carry out score calculating, the statement of all functions and structural body definition of libSVM SVM
It is all contained in libSVM.h file, the meaning of some structural bodies is as follows in libSVM.h:
struct svm_node
{
int index；
double value；
}。
The structural body defines one " SVM node ", it may be assumed that index i and its corresponding ith feature value.N in this way
The SVM node of the same category number just constitutes a SVM input vector.That is: a SVM input vector can be expressed as
Form:
Class label index 1: the index of characteristic value 12: the index of characteristic value 23: characteristic value 3...
Input vector as several is input to libSVM to be trained, alternatively, one class label of input is unknown
Vector it is predicted, 0-100 is normalized in prediction result, the mass fraction of testing image block is obtained with this.
S164, obtain several testing image blocks mass fraction average value, to obtain image quality score.
It is averaged by predicting the mass fraction of each testing image block, then by the mass fraction of these testing image blocks
It is worth the mass fraction as testing image picture.
S170, synthetic image articulation score and image quality score form evaluation result.
Specifically, image definition score and image quality score are weighted summation according to specified weight, to be formed
Evaluation result.
In the present embodiment, the image definition score and be based on deep learning that integration capability gradient sharpness computation obtains
The image quality score that model and SVM are predicted provides the evaluation result of image, calculates according to the following equation:
Score=ScoreCNN × 0.6+D (f) × 0.4, wherein score is the comprehensive score of image, i.e. evaluation result,
ScoreCNN is the image quality score predicted based on deep learning model and SVM, and D (f) refers to image definition point
Number.
S180, output evaluation result.
The evaluation result that will acquire feeds back to the picture quality evaluation APP of user terminal 10, so that user knows input
Image score.
Above-mentioned picture quality assessment method, by first being scored from illumination, size and deflection angle training image,
Training image is subjected to quality score, utilizes the training set for the feature vector composition that training image extracts, the quality of training image
Score is trained adjustment to CNN sorter network, to obtain deep learning mode, in use, image is carried out clarity scoring
And deep learning model scores, then is weighted synthesis to two scores, so that evaluation result is obtained, to improve figure
The accuracy that tablet quality determines greatly improves the accuracy rate of true environment human face identification.
Referring to Fig. 6, Fig. 6 is the schematic block diagram for the picture quality evaluation device 200 that the specific embodiment of the invention provides；
As shown in fig. 6, the picture quality evaluation device 200 includes:
Image acquisition unit 204, for obtaining image.
Clarity analytical unit 205, for carrying out clarity analysis to image, to obtain image definition score.
Deep learning unit 206, for using deep learning model to image carry out deep learning, and to learning outcome into
Row Score on Prediction, to obtain image quality score.
Comprehensive unit 207 is used for synthetic image articulation score and image quality score, forms evaluation result.
As a result output unit 208, for exporting evaluation result.
In one embodiment, as shown in fig. 7, above-mentioned device further include:
Training image acquiring unit 201, for obtaining several training images；
Unit 202 is marked, for carrying out quality score to training image and marking；
Model acquiring unit 203, for being trained to the training image after mark, to obtain deep learning model.
In one embodiment, as shown in figure 8, above-mentioned model acquiring unit 203 includes:
Whether size judgment sub-unit 2031, the size for training of judgement image meet preset requirement.
Size handles subelement 2032, is used for if it is not, then carrying out size processing to training image.
Subelement 2033 is screened, for screening several training image blocks from training image.
Training set establishes subelement 2034, for extracting the corresponding feature vector of training image blocks, and establishes training set.
Network training subelement 2035, for training CNN sorter network using training set and training image, to obtain depth
Learning model.
In one embodiment, as shown in figure 9, above-mentioned clarity analytical unit 205 includes:
Sharpness computation subelement 2051, for carrying out energy gradient sharpness computation to image, to obtain calculated result.
Normalizing subelement 2052, for normalizing the calculated result, to obtain image definition score.
In one embodiment, as shown in Figure 10, above-mentioned deep learning unit 206 includes:
Image block extracts subelement 2061, for extracting several testing image blocks to image.
Subelement 2062 is inputted, for several testing image blocks to be inputted deep learning model, is waited for extracting several
The feature vector of altimetric image block.
It predicts subelement 2063, carries out predicting several testing images for the feature vector to several testing image blocks
The mass fraction of block.
Average value obtains subelement 2064, the average value of the mass fraction for obtaining several testing image blocks, to obtain
To image quality score.
It should be noted that it is apparent to those skilled in the art that, above-mentioned picture quality evaluation device
200 and each unit specific implementation process, can with reference to the corresponding description in preceding method embodiment, for convenience of description and
Succinctly, details are not described herein.
Above-mentioned picture quality evaluation device 200 can be implemented as a kind of form of computer program, which can
To be run in computer equipment as shown in figure 11.
Figure 11 is please referred to, Figure 11 is a kind of schematic block diagram of computer equipment provided by the embodiments of the present application.The calculating
Machine equipment 500 can be terminal, be also possible to server, wherein terminal can be smart phone, tablet computer, notebook electricity
Brain, desktop computer, personal digital assistant and wearable device etc. have the electronic equipment of communication function.Server can be independence
Server, be also possible to the server cluster of multiple servers composition.
Refering to fig. 11, which includes processor 502, memory and the net connected by system bus 501
Network interface 505, wherein memory may include non-volatile memory medium 503 and built-in storage 504.
The non-volatile memory medium 503 can storage program area 5031 and computer program 5032.The computer program
5032 include program instruction, which is performed, and processor 502 may make to execute a kind of picture quality assessment method.
The processor 502 is for providing calculating and control ability, to support the operation of entire computer equipment 500.
The built-in storage 504 provides environment for the operation of the computer program 5032 in non-volatile memory medium 503, should
When computer program 5032 is executed by processor 502, processor 502 may make to execute a kind of picture quality assessment method.
The network interface 505 is used to carry out network communication with other equipment.It will be understood by those skilled in the art that in Figure 11
The structure shown, only the block diagram of part-structure relevant to application scheme, does not constitute and is applied to application scheme
The restriction of computer equipment 500 thereon, specific computer equipment 500 may include more more or fewer than as shown in the figure
Component perhaps combines certain components or with different component layouts.
Wherein, the processor 502 is for running computer program 5032 stored in memory, to realize following step
It is rapid:
Obtain image；
Clarity analysis is carried out to image, to obtain image definition score；
Deep learning is carried out to image using deep learning model, and Score on Prediction is carried out to learning outcome, to obtain figure
As mass fraction；
Synthetic image articulation score and image quality score form evaluation result；
Export evaluation result.
In one embodiment, processor 502 carries out deep learning to image in the realization use deep learning model, with
Before obtaining image quality score step, following steps are also realized:
Obtain several training images；
Quality score is carried out to training image and is marked；
Training image after mark is trained, to obtain deep learning model.
In one embodiment, training image of the processor 502 after realizing described pair of mark is trained, to obtain depth
When learning model step, it is implemented as follows step:
Whether the size of training of judgement image meets preset requirement；
If it is not, then carrying out size processing to training image, and enter next step, if so, into next step；
Several training image blocks are screened from training image；
The corresponding feature vector of training image blocks is extracted, and establishes training set；
Using training set and training image training CNN sorter network, to obtain deep learning model.
In one embodiment, processor 502 is described to image progress clarity analysis in realization, to obtain image definition
When score step, it is implemented as follows step:
Energy gradient sharpness computation is carried out to image, to obtain calculated result；
The calculated result is normalized, to obtain image definition score.
In one embodiment, processor 502 carries out deep learning to image in the realization use deep learning model, and
Score on Prediction is carried out to learning outcome and is implemented as follows step when obtaining image quality score step:
Several testing image blocks are extracted to image；
Several testing image blocks are inputted into deep learning model, to extract the feature vector of several testing image blocks；
Predict to the feature vector of several testing image blocks the mass fraction of several testing image blocks；
The average value of the mass fraction of several testing image blocks is obtained, to obtain image quality score.
In one embodiment, processor 502 is realizing the synthetic image articulation score and image quality score, shape
When at evaluation result step, it is implemented as follows step:
Image definition score and image quality score are weighted summation according to specified weight, to form evaluation knot
Fruit.
It should be appreciated that in the embodiment of the present application, processor 502 can be central processing unit (Central
Processing Unit, CPU), which can also be other general processors, digital signal processor (Digital
Signal Processor, DSP), specific integrated circuit (Application Specific Integrated Circuit,
ASIC), ready-made programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic
Device, discrete gate or transistor logic, discrete hardware components etc..Wherein, general processor can be microprocessor or
Person's processor is also possible to any conventional processor etc..
Those of ordinary skill in the art will appreciate that be realize above-described embodiment method in all or part of the process,
It is that relevant hardware can be instructed to complete by computer program.The computer program includes program instruction, computer journey
Sequence can be stored in a storage medium, which is computer readable storage medium.The program instruction is by the department of computer science
At least one processor in system executes, to realize the process step of the embodiment of the above method.
Therefore, the present invention also provides a kind of storage mediums.The storage medium can be computer readable storage medium.This is deposited
Storage media is stored with computer program, and wherein computer program includes program instruction.The program instruction makes when being executed by processor
Processor executes following steps:
Obtain image；
Clarity analysis is carried out to image, to obtain image definition score；
Deep learning is carried out to image using deep learning model, and Score on Prediction is carried out to learning outcome, to obtain figure
As mass fraction；
Synthetic image articulation score and image quality score form evaluation result；
Export evaluation result.
In one embodiment, the processor is realized described using deep learning model pair in the instruction of execution described program
Image carries out deep learning, before obtaining image quality score step, also realizes following steps:
Obtain several training images；
Quality score is carried out to training image and is marked；
Training image after mark is trained, to obtain deep learning model.
In one embodiment, training image of the processor after described pair of mark is realized in the instruction of execution described program
It is trained, when obtaining deep learning model step, is implemented as follows step:
Whether the size of training of judgement image meets preset requirement；
If it is not, then carrying out size processing to training image, and enter next step, if so, into next step；
Several training image blocks are screened from training image；
The corresponding feature vector of training image blocks is extracted, and establishes training set；
Using training set and training image training CNN sorter network, to obtain deep learning model.
In one embodiment, the processor is realized described to image progress clarity point in the instruction of execution described program
Analysis, when obtaining image definition score step, is implemented as follows step:
Energy gradient sharpness computation is carried out to image, to obtain calculated result；
The calculated result is normalized, to obtain image definition score.
In one embodiment, the processor is realized described using deep learning model pair in the instruction of execution described program
Image carries out deep learning, and carries out Score on Prediction to learning outcome, and when obtaining image quality score step, specific implementation is such as
Lower step:
Several testing image blocks are extracted to image；
Several testing image blocks are inputted into deep learning model, to extract the feature vector of several testing image blocks；
Predict to the feature vector of several testing image blocks the mass fraction of several testing image blocks；
The average value of the mass fraction of several testing image blocks is obtained, to obtain image quality score.
In one embodiment, the processor realizes the synthetic image articulation score executing described program instruction
And image quality score is implemented as follows step when forming evaluation result step:
Image definition score and image quality score are weighted summation according to specified weight, to form evaluation knot
Fruit.
The storage medium can be USB flash disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), magnetic disk
Or the various computer readable storage mediums that can store program code such as CD.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware
With the interchangeability of software, each exemplary composition and step are generally described according to function in the above description.This
A little functions are implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Specially
Industry technical staff can use different methods to achieve the described function each specific application, but this realization is not
It is considered as beyond the scope of this invention.
In several embodiments provided by the present invention, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary.For example, the division of each unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation.Such as multiple units or components can be tied
Another system is closed or is desirably integrated into, or some features can be ignored or not executed.
The steps in the embodiment of the present invention can be sequentially adjusted, merged and deleted according to actual needs.This hair
Unit in bright embodiment device can be combined, divided and deleted according to actual needs.In addition, in each implementation of the present invention
Each functional unit in example can integrate in one processing unit, is also possible to each unit and physically exists alone, can also be with
It is that two or more units are integrated in one unit.
If the integrated unit is realized in the form of SFU software functional unit and when sold or used as an independent product,
It can store in one storage medium.Based on this understanding, technical solution of the present invention is substantially in other words to existing skill
The all or part of part or the technical solution that art contributes can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a
People's computer, terminal or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention.
It is above-mentioned that technology contents of the invention are only further illustrated with embodiment, in order to which reader is easier to understand, but not
It represents embodiments of the present invention and is only limitted to this, any technology done according to the present invention extends or recreation, by of the invention
Protection.Protection scope of the present invention is subject to claims.
Claims (10)
1. picture quality assessment method characterized by comprising
Obtain image；
Clarity analysis is carried out to image, to obtain image definition score；
Deep learning is carried out to image using deep learning model, and Score on Prediction is carried out to learning outcome, to obtain image matter
Measure score；
Synthetic image articulation score and image quality score form evaluation result；
Export evaluation result.
2. picture quality assessment method according to claim 1, which is characterized in that described to use deep learning model to figure
As carrying out deep learning, before obtaining image quality score, further includes:
Obtain several training images；
Quality score is carried out to training image and is marked；
Training image after mark is trained, to obtain deep learning model.
3. picture quality assessment method according to claim 2, which is characterized in that described pair mark after training image into
Row training, to obtain deep learning model, comprising:
Whether the size of training of judgement image meets preset requirement；
If it is not, then carrying out size processing to training image, and enter next step, if so, into next step；
Several training image blocks are screened from training image；
The corresponding feature vector of training image blocks is extracted, and establishes training set；
Using training set and training image training CNN sorter network, to obtain deep learning model.
4. picture quality assessment method according to claim 3, which is characterized in that described to carry out clarity point to image
Analysis, to obtain image definition score, comprising:
Energy gradient sharpness computation is carried out to image, to obtain calculated result；
The calculated result is normalized, to obtain image definition score.
5. picture quality assessment method according to claim 4, which is characterized in that described to use deep learning model to figure
Score on Prediction is carried out as carrying out deep learning, and to learning outcome, to obtain image quality score, comprising:
Several testing image blocks are extracted to image；
Several testing image blocks are inputted into deep learning model, to extract the feature vector of several testing image blocks；
Predict to the feature vector of several testing image blocks the mass fraction of several testing image blocks；
The average value of the mass fraction of several testing image blocks is obtained, to obtain image quality score.
6. picture quality assessment method according to any one of claims 1 to 5, which is characterized in that the synthetic image is clear
Clear degree score and image quality score form evaluation result, comprising:
Image definition score and image quality score are weighted summation according to specified weight, to form evaluation result.
7. picture quality evaluation device characterized by comprising
Image acquisition unit, for obtaining image；
Clarity analytical unit, for carrying out clarity analysis to image, to obtain image definition score；
Deep learning unit for carrying out deep learning to image using deep learning model, and carries out score to learning outcome
Prediction, to obtain image quality score；
Comprehensive unit is used for synthetic image articulation score and image quality score, forms evaluation result；
As a result output unit, for exporting evaluation result.
8. picture quality evaluation device according to claim 7, which is characterized in that the clarity analytical unit includes:
Sharpness computation subelement, for carrying out energy gradient sharpness computation to image, to obtain calculated result；
Normalizing subelement, for normalizing the calculated result, to obtain image definition score.
9. a kind of computer equipment, which is characterized in that including memory, processor and be stored on the memory and can be in institute
The computer program run on processor is stated, the processor is realized when executing the computer program as in claim 1 to 6
Picture quality assessment method described in any one.
10. a kind of storage medium, which is characterized in that the storage medium is stored with computer program, the computer program packet
Program instruction is included, described program instruction makes the processor execute such as claim 1 to 6 any one when being executed by a processor
The picture quality assessment method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810869434.3A CN109146856A (en) | 2018-08-02 | 2018-08-02 | Picture quality assessment method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810869434.3A CN109146856A (en) | 2018-08-02 | 2018-08-02 | Picture quality assessment method, device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109146856A true CN109146856A (en) | 2019-01-04 |
Family
ID=64799413
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810869434.3A Pending CN109146856A (en) | 2018-08-02 | 2018-08-02 | Picture quality assessment method, device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109146856A (en) |
Cited By (40)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110321843A (en) * | 2019-07-04 | 2019-10-11 | 杭州视洞科技有限公司 | A kind of face out of kilter method based on deep learning |
| CN110400335A (en) * | 2019-07-25 | 2019-11-01 | 广西科技大学 | Quality Estimation Method of Texture Image Based on Deep Learning |
| CN110458829A (en) * | 2019-08-13 | 2019-11-15 | 腾讯医疗健康(深圳)有限公司 | Image quality control method, device, equipment and storage medium based on artificial intelligence |
| CN110766052A (en) * | 2019-09-20 | 2020-02-07 | 北京三快在线科技有限公司 | Image display method, evaluation model generation device and electronic equipment |
| CN110781633A (en) * | 2019-10-30 | 2020-02-11 | 广东博智林机器人有限公司 | Image-text design quality detection method, device and system based on deep learning model |
| CN110796106A (en) * | 2019-11-04 | 2020-02-14 | 北京迈格威科技有限公司 | Method for establishing portrait quality evaluation model and recognizing portrait from video |
| CN110807491A (en) * | 2019-11-05 | 2020-02-18 | 上海眼控科技股份有限公司 | License plate image definition model training method, definition detection method and device |
| CN110942072A (en) * | 2019-12-31 | 2020-03-31 | 北京迈格威科技有限公司 | Quality evaluation-based quality scoring and detecting model training and detecting method and device |
| CN111091069A (en) * | 2019-11-27 | 2020-05-01 | 云南电网有限责任公司电力科学研究院 | A power grid target detection method and system guided by blind image quality assessment |
| CN111160284A (en) * | 2019-12-31 | 2020-05-15 | 苏州纳智天地智能科技有限公司 | Method, system, equipment and storage medium for evaluating quality of face photo |
| CN111179382A (en) * | 2020-01-02 | 2020-05-19 | 广东博智林机器人有限公司 | Image typesetting method, device, medium and electronic equipment |
| CN111199186A (en) * | 2019-12-03 | 2020-05-26 | 恒大智慧科技有限公司 | Image quality scoring model training method, device, equipment and storage medium |
| CN111210402A (en) * | 2019-12-03 | 2020-05-29 | 恒大智慧科技有限公司 | Face image quality scoring method and device, computer equipment and storage medium |
| CN111291829A (en) * | 2020-03-03 | 2020-06-16 | 支付宝(杭州)信息技术有限公司 | Automatic determination method and system for selected pictures |
| CN111340073A (en) * | 2020-02-12 | 2020-06-26 | 杭州涂鸦信息技术有限公司 | Deep learning training method and system for image grade evaluation |
| CN111506748A (en) * | 2019-01-30 | 2020-08-07 | 斯特拉德视觉公司 | Method and device for managing intelligent database for face recognition |
| CN111584044A (en) * | 2020-05-14 | 2020-08-25 | 上海深至信息科技有限公司 | Method and system for generating standard ultrasonic image training set |
| CN111626248A (en) * | 2020-06-01 | 2020-09-04 | 北京世纪好未来教育科技有限公司 | Face value scoring model training method, face value scoring method and related device |
| CN111753873A (en) * | 2020-05-12 | 2020-10-09 | 北京捷通华声科技股份有限公司 | An image detection method and device |
| CN111815606A (en) * | 2020-07-09 | 2020-10-23 | 浙江大华技术股份有限公司 | Image quality evaluation method, storage medium, and computing device |
| CN112102309A (en) * | 2020-09-27 | 2020-12-18 | 中国建设银行股份有限公司 | Method, device and equipment for determining image quality evaluation result |
| CN112132107A (en) * | 2020-10-09 | 2020-12-25 | 腾讯科技（深圳）有限公司 | Image processing method, apparatus, electronic device, and computer-readable storage medium |
| CN112200804A (en) * | 2020-11-09 | 2021-01-08 | 北京地平线信息技术有限公司 | Image detection method and device, computer readable storage medium and electronic equipment |
| CN112329638A (en) * | 2020-11-06 | 2021-02-05 | 上海优扬新媒信息技术有限公司 | Image scoring method, device and system |
| CN112418009A (en) * | 2020-11-06 | 2021-02-26 | 中保车服科技服务股份有限公司 | Image quality detection method, terminal device and storage medium |
| CN112672143A (en) * | 2020-12-21 | 2021-04-16 | 北京金山云网络技术有限公司 | Video quality evaluation method and device and server |
| CN112712492A (en) * | 2020-02-04 | 2021-04-27 | 首都医科大学附属北京友谊医院 | Method, device, server and storage medium for determining equipment quality |
| CN112785572A (en) * | 2021-01-21 | 2021-05-11 | 上海云从汇临人工智能科技有限公司 | Image quality evaluation method, device and computer readable storage medium |
| CN112825120A (en) * | 2019-11-20 | 2021-05-21 | 北京眼神智能科技有限公司 | Face illumination evaluation method and device, computer readable storage medium and equipment |
| CN113158777A (en) * | 2021-03-08 | 2021-07-23 | 佳都新太科技股份有限公司 | Quality scoring method, quality scoring model training method and related device |
| CN113298139A (en) * | 2021-05-21 | 2021-08-24 | 广州文远知行科技有限公司 | Image data optimization method, device, equipment and medium |
| CN113366420A (en) * | 2019-04-17 | 2021-09-07 | 深圳市欢太科技有限公司 | Image processing method, computer-readable storage medium, and computer device |
| CN113409281A (en) * | 2021-06-24 | 2021-09-17 | 上海云从企业发展有限公司 | Image definition evaluation method and device based on depth model |
| CN113496485A (en) * | 2021-06-24 | 2021-10-12 | 北京市遥感信息研究所 | Satellite remote sensing image quality evaluation method and device |
| CN113642479A (en) * | 2021-08-17 | 2021-11-12 | 北京的卢深视科技有限公司 | Human face image evaluation method and device, electronic equipment and storage medium |
| CN114286081A (en) * | 2021-12-22 | 2022-04-05 | 携程旅游信息技术（上海）有限公司 | Video quality evaluation method, apparatus, and medium |
| CN114520890A (en) * | 2020-11-19 | 2022-05-20 | 华为技术有限公司 | Image processing method and device |
| CN115131218A (en) * | 2021-03-25 | 2022-09-30 | 腾讯科技（深圳）有限公司 | Image processing method, image processing device, computer readable medium and electronic equipment |
| CN116777900A (en) * | 2023-07-31 | 2023-09-19 | 广州医科大学附属第一医院（广州呼吸中心） | A medical image quality scoring method, system and readable storage medium |
| CN117788461A (en) * | 2024-02-23 | 2024-03-29 | 华中科技大学同济医学院附属同济医院 | Magnetic resonance image quality evaluation system based on image analysis |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6023527A (en) * | 1995-06-27 | 2000-02-08 | Ricoh Company, Ltd. | Method and system of selecting a color space mapping technique for an output color space |
| CN103366369A (en) * | 2013-07-01 | 2013-10-23 | 中国矿业大学 | Method and device for evaluating blocking artifacts in image |
| CN105117705A (en) * | 2015-08-26 | 2015-12-02 | 北京无线电计量测试研究所 | Iris image quality cascade type evaluation method |
| CN105160678A (en) * | 2015-09-02 | 2015-12-16 | 山东大学 | Convolutional-neural-network-based reference-free three-dimensional image quality evaluation method |
| CN107633513A (en) * | 2017-09-18 | 2018-01-26 | 天津大学 | The measure of 3D rendering quality based on deep learning |
| CN108257132A (en) * | 2018-03-05 | 2018-07-06 | 南方医科大学 | A kind of method of the CT image quality measures based on machine learning |
-
2018
- 2018-08-02 CN CN201810869434.3A patent/CN109146856A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6023527A (en) * | 1995-06-27 | 2000-02-08 | Ricoh Company, Ltd. | Method and system of selecting a color space mapping technique for an output color space |
| CN103366369A (en) * | 2013-07-01 | 2013-10-23 | 中国矿业大学 | Method and device for evaluating blocking artifacts in image |
| CN105117705A (en) * | 2015-08-26 | 2015-12-02 | 北京无线电计量测试研究所 | Iris image quality cascade type evaluation method |
| CN105160678A (en) * | 2015-09-02 | 2015-12-16 | 山东大学 | Convolutional-neural-network-based reference-free three-dimensional image quality evaluation method |
| CN107633513A (en) * | 2017-09-18 | 2018-01-26 | 天津大学 | The measure of 3D rendering quality based on deep learning |
| CN108257132A (en) * | 2018-03-05 | 2018-07-06 | 南方医科大学 | A kind of method of the CT image quality measures based on machine learning |
Non-Patent Citations (1)
| Title |
|---|
| 王志明: "无参考图像质量评价综述", 《自动化学报》 * |
Cited By (55)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111506748A (en) * | 2019-01-30 | 2020-08-07 | 斯特拉德视觉公司 | Method and device for managing intelligent database for face recognition |
| CN111506748B (en) * | 2019-01-30 | 2024-03-05 | 斯特拉德视觉公司 | Method and device for managing intelligent database for face recognition |
| CN113366420A (en) * | 2019-04-17 | 2021-09-07 | 深圳市欢太科技有限公司 | Image processing method, computer-readable storage medium, and computer device |
| CN113366420B (en) * | 2019-04-17 | 2024-05-03 | 深圳市欢太科技有限公司 | Image processing method, computer-readable storage medium, and computer device |
| CN110321843A (en) * | 2019-07-04 | 2019-10-11 | 杭州视洞科技有限公司 | A kind of face out of kilter method based on deep learning |
| CN110321843B (en) * | 2019-07-04 | 2021-11-09 | 杭州视洞科技有限公司 | Face optimization method based on deep learning |
| CN110400335A (en) * | 2019-07-25 | 2019-11-01 | 广西科技大学 | Quality Estimation Method of Texture Image Based on Deep Learning |
| CN110458829A (en) * | 2019-08-13 | 2019-11-15 | 腾讯医疗健康(深圳)有限公司 | Image quality control method, device, equipment and storage medium based on artificial intelligence |
| CN110458829B (en) * | 2019-08-13 | 2024-01-30 | 腾讯医疗健康(深圳)有限公司 | Image quality control method, device, equipment and storage medium based on artificial intelligence |
| CN110766052A (en) * | 2019-09-20 | 2020-02-07 | 北京三快在线科技有限公司 | Image display method, evaluation model generation device and electronic equipment |
| CN110781633A (en) * | 2019-10-30 | 2020-02-11 | 广东博智林机器人有限公司 | Image-text design quality detection method, device and system based on deep learning model |
| CN110796106A (en) * | 2019-11-04 | 2020-02-14 | 北京迈格威科技有限公司 | Method for establishing portrait quality evaluation model and recognizing portrait from video |
| CN110807491A (en) * | 2019-11-05 | 2020-02-18 | 上海眼控科技股份有限公司 | License plate image definition model training method, definition detection method and device |
| CN112825120A (en) * | 2019-11-20 | 2021-05-21 | 北京眼神智能科技有限公司 | Face illumination evaluation method and device, computer readable storage medium and equipment |
| CN112825120B (en) * | 2019-11-20 | 2024-04-23 | 北京眼神智能科技有限公司 | Face illumination evaluation method, device, computer-readable storage medium and equipment |
| CN111091069A (en) * | 2019-11-27 | 2020-05-01 | 云南电网有限责任公司电力科学研究院 | A power grid target detection method and system guided by blind image quality assessment |
| CN111210402A (en) * | 2019-12-03 | 2020-05-29 | 恒大智慧科技有限公司 | Face image quality scoring method and device, computer equipment and storage medium |
| CN111199186A (en) * | 2019-12-03 | 2020-05-26 | 恒大智慧科技有限公司 | Image quality scoring model training method, device, equipment and storage medium |
| CN110942072A (en) * | 2019-12-31 | 2020-03-31 | 北京迈格威科技有限公司 | Quality evaluation-based quality scoring and detecting model training and detecting method and device |
| CN110942072B (en) * | 2019-12-31 | 2024-02-02 | 北京迈格威科技有限公司 | Quality score based on quality assessment, detection model training and detection method and device |
| CN111160284A (en) * | 2019-12-31 | 2020-05-15 | 苏州纳智天地智能科技有限公司 | Method, system, equipment and storage medium for evaluating quality of face photo |
| CN111179382A (en) * | 2020-01-02 | 2020-05-19 | 广东博智林机器人有限公司 | Image typesetting method, device, medium and electronic equipment |
| CN112712492A (en) * | 2020-02-04 | 2021-04-27 | 首都医科大学附属北京友谊医院 | Method, device, server and storage medium for determining equipment quality |
| CN111340073A (en) * | 2020-02-12 | 2020-06-26 | 杭州涂鸦信息技术有限公司 | Deep learning training method and system for image grade evaluation |
| CN111291829A (en) * | 2020-03-03 | 2020-06-16 | 支付宝(杭州)信息技术有限公司 | Automatic determination method and system for selected pictures |
| CN111753873A (en) * | 2020-05-12 | 2020-10-09 | 北京捷通华声科技股份有限公司 | An image detection method and device |
| CN111584044A (en) * | 2020-05-14 | 2020-08-25 | 上海深至信息科技有限公司 | Method and system for generating standard ultrasonic image training set |
| CN111584044B (en) * | 2020-05-14 | 2023-07-14 | 上海深至信息科技有限公司 | Standard ultrasonic image training set generation method and system |
| CN111626248B (en) * | 2020-06-01 | 2022-05-06 | 北京世纪好未来教育科技有限公司 | Color value scoring model training method, color value scoring method and related device |
| CN111626248A (en) * | 2020-06-01 | 2020-09-04 | 北京世纪好未来教育科技有限公司 | Face value scoring model training method, face value scoring method and related device |
| CN111815606B (en) * | 2020-07-09 | 2023-09-01 | 浙江大华技术股份有限公司 | Image quality evaluation method, storage medium, and computing device |
| CN111815606A (en) * | 2020-07-09 | 2020-10-23 | 浙江大华技术股份有限公司 | Image quality evaluation method, storage medium, and computing device |
| CN112102309A (en) * | 2020-09-27 | 2020-12-18 | 中国建设银行股份有限公司 | Method, device and equipment for determining image quality evaluation result |
| CN112132107A (en) * | 2020-10-09 | 2020-12-25 | 腾讯科技（深圳）有限公司 | Image processing method, apparatus, electronic device, and computer-readable storage medium |
| CN112418009B (en) * | 2020-11-06 | 2024-03-22 | 中保车服科技服务股份有限公司 | Image quality detection method, terminal equipment and storage medium |
| CN112418009A (en) * | 2020-11-06 | 2021-02-26 | 中保车服科技服务股份有限公司 | Image quality detection method, terminal device and storage medium |
| CN112329638A (en) * | 2020-11-06 | 2021-02-05 | 上海优扬新媒信息技术有限公司 | Image scoring method, device and system |
| CN112200804A (en) * | 2020-11-09 | 2021-01-08 | 北京地平线信息技术有限公司 | Image detection method and device, computer readable storage medium and electronic equipment |
| CN114520890A (en) * | 2020-11-19 | 2022-05-20 | 华为技术有限公司 | Image processing method and device |
| CN114520890B (en) * | 2020-11-19 | 2023-07-11 | 华为技术有限公司 | Image processing method and device |
| CN112672143A (en) * | 2020-12-21 | 2021-04-16 | 北京金山云网络技术有限公司 | Video quality evaluation method and device and server |
| CN112785572A (en) * | 2021-01-21 | 2021-05-11 | 上海云从汇临人工智能科技有限公司 | Image quality evaluation method, device and computer readable storage medium |
| CN112785572B (en) * | 2021-01-21 | 2023-10-24 | 上海云从汇临人工智能科技有限公司 | Image quality evaluation method, apparatus and computer readable storage medium |
| CN113158777A (en) * | 2021-03-08 | 2021-07-23 | 佳都新太科技股份有限公司 | Quality scoring method, quality scoring model training method and related device |
| CN115131218A (en) * | 2021-03-25 | 2022-09-30 | 腾讯科技（深圳）有限公司 | Image processing method, image processing device, computer readable medium and electronic equipment |
| CN113298139B (en) * | 2021-05-21 | 2024-02-27 | 广州文远知行科技有限公司 | Image data optimization method, device, equipment and medium |
| CN113298139A (en) * | 2021-05-21 | 2021-08-24 | 广州文远知行科技有限公司 | Image data optimization method, device, equipment and medium |
| CN113496485A (en) * | 2021-06-24 | 2021-10-12 | 北京市遥感信息研究所 | Satellite remote sensing image quality evaluation method and device |
| CN113496485B (en) * | 2021-06-24 | 2022-12-02 | 北京市遥感信息研究所 | Satellite remote sensing image quality evaluation method and device |
| CN113409281A (en) * | 2021-06-24 | 2021-09-17 | 上海云从企业发展有限公司 | Image definition evaluation method and device based on depth model |
| CN113642479A (en) * | 2021-08-17 | 2021-11-12 | 北京的卢深视科技有限公司 | Human face image evaluation method and device, electronic equipment and storage medium |
| CN114286081A (en) * | 2021-12-22 | 2022-04-05 | 携程旅游信息技术（上海）有限公司 | Video quality evaluation method, apparatus, and medium |
| CN116777900A (en) * | 2023-07-31 | 2023-09-19 | 广州医科大学附属第一医院（广州呼吸中心） | A medical image quality scoring method, system and readable storage medium |
| CN117788461A (en) * | 2024-02-23 | 2024-03-29 | 华中科技大学同济医学院附属同济医院 | Magnetic resonance image quality evaluation system based on image analysis |
| CN117788461B (en) * | 2024-02-23 | 2024-05-07 | 华中科技大学同济医学院附属同济医院 | Magnetic resonance image quality evaluation system based on image analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109146856A (en) | Picture quality assessment method, device, computer equipment and storage medium | |
| CN112215180B (en) | Living body detection method and device | |
| CN110399929B (en) | Fundus image classification method, fundus image classification apparatus, and computer-readable storage medium | |
| CN105631439B (en) | Face image processing process and device | |
| CN105138993B (en) | Establish the method and device of human face recognition model | |
| CN102113335B (en) | Image processing apparatus and method | |
| JP6664163B2 (en) | Image identification method, image identification device, and program | |
| CN106056064B (en) | A kind of face identification method and face identification device | |
| CN108898579A (en) | A kind of image definition recognition methods, device and storage medium | |
| CN104866868B (en) | Metal coins recognition methods based on deep neural network and device | |
| CN104123543B (en) | A kind of eye movement recognition methods based on recognition of face | |
| KR20200004841A (en) | System and method for guiding a user to take a selfie | |
| CN108875522A (en) | Face cluster methods, devices and systems and storage medium | |
| CN109753938A (en) | Image-recognizing method and equipment and the training method of application, neural network | |
| CN111935479B (en) | Target image determination method and device, computer equipment and storage medium | |
| CN109117897A (en) | Image processing method, device and readable storage medium storing program for executing based on convolutional neural networks | |
| CN110032925A (en) | A kind of images of gestures segmentation and recognition methods based on improvement capsule network and algorithm | |
| EP1964028A1 (en) | Method for automatic detection and classification of objects and patterns in low resolution environments | |
| CN111160284A (en) | Method, system, equipment and storage medium for evaluating quality of face photo | |
| CN107590473A (en) | A kind of human face in-vivo detection method, medium and relevant apparatus | |
| CN109886153A (en) | A real-time face detection method based on deep convolutional neural network | |
| CN107341505A (en) | A kind of scene classification method based on saliency Yu Object Bank | |
| Sakthimohan et al. | Detection and Recognition of Face Using Deep Learning | |
| CN114066877A (en) | Image screening method and device, electronic equipment and storage medium | |
| Buzuti et al. | Fréchet AutoEncoder distance: a new approach for evaluation of generative adversarial networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190104 |