CN110473181A - Screen content image based on edge feature information without ginseng quality evaluating method - Google Patents
Screen content image based on edge feature information without ginseng quality evaluating method Download PDFInfo
- Publication number
- CN110473181A CN110473181A CN201910704654.5A CN201910704654A CN110473181A CN 110473181 A CN110473181 A CN 110473181A CN 201910704654 A CN201910704654 A CN 201910704654A CN 110473181 A CN110473181 A CN 110473181A
- Authority
- CN
- China
- Prior art keywords
- scis
- image
- quality
- edge
- score
- 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
- 238000000034 method Methods 0.000 title claims abstract description 32
- 241000208340 Araliaceae Species 0.000 title claims abstract description 12
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 title claims abstract description 12
- 235000003140 Panax quinquefolius Nutrition 0.000 title claims abstract description 12
- 235000008434 ginseng Nutrition 0.000 title claims abstract description 12
- 238000012549 training Methods 0.000 claims abstract description 22
- 239000013598 vector Substances 0.000 claims abstract description 20
- 238000010606 normalization Methods 0.000 claims abstract description 14
- 238000013507 mapping Methods 0.000 claims abstract description 11
- 238000000605 extraction Methods 0.000 claims abstract description 6
- 238000001303 quality assessment method Methods 0.000 claims abstract description 3
- 239000000284 extract Substances 0.000 claims abstract 2
- 238000013441 quality evaluation Methods 0.000 claims description 24
- 230000000007 visual effect Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 5
- 230000007613 environmental effect Effects 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000013210 evaluation model Methods 0.000 description 3
- 230000008447 perception Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000011833 dog model Methods 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 238000002834 transmittance Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F18/2148—Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the process organisation or structure, e.g. boosting cascade
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge 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
- G06T2207/20—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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of screen content images based on edge feature information without ginseng quality evaluating method, including the following steps: is calculated by DOG algorithm and obtains the mapping of the image border SCIs;The edge map information that will acquire carries out local normalization, in the case where given SCIs, calculates normalization pixel using local average subtraction and the normalized method of division;The extraction of edge feature vector is carried out, extracts marginal information from part normalization edge graph using the fitting of L- square distribution estimation method;Device is returned using ABPNN, and quality is perceived into maps feature vectors to subjective quality score, L- square is distributed SCIs image n dimension its subjective quality assessment score of edge feature vector sum that estimation is extracted and is put into training in ABPNN network, in the training stage, regression model is established using edge feature vector and its corresponding subjective differences mean value Opinion Score (DMOS) value.
Description
Technical field
The invention belongs to image quality evaluation fields, are related to screen content image without ginseng quality evaluating method.
Background technique
In recent years, the rapid development of communications, promotes the extensive use of telecommunication, such as online news and wide
It accuses, cloud computing etc., to generate the screen content image (SCIs) largely generated by the screen of various terminal equipment.With natural field
Scape image is compared, and SCIs has quite apparent structural information and statistical nature, because SCI includes the content that computer generates,
Such as text, chart, map, mode or illustration (such as logo, bar code, two dimensional code), other icons.
According to the existence of reference image information, traditional Images of Natural Scenery quality evaluating method is generally divided into three types
Type, i.e., entirely with reference to (FR), reduction with reference to (RR) and blind/without with reference to (NR) scheme.It is complete to join natural image quality evaluation algorithm and get over
Subtract parameter natural image evaluation algorithms, using structural similarity index (SSIM), basic theories is human visual system (HVS)
Perceptual image quality effectively estimated to image spatial feature susceptibility.Without ginseng natural image quality evaluation without necessarily referring to letter
Breath is a kind of method for adapting to actual conditions, such as measurement based on natural scene statistics, support vector regression, kernel study, base
In the measurement of deep learning.
Since SCIs and natural image have visibly different structure and statistical property, the particularity of visual quality research
It is still a challenging problem.The image quality evaluation of SCIs is to solve how to assess from various electronic equipments
Into in our eyes deformation SCIs perceived quality the problem of.SCIs is constituted there are two types of method, it is another one is in receiving end
Kind is during image processing, such as to transmit in transmitting terminal and compression, SCIs will appear various distortions.By patent
Applicant retrieve discovery, no ginseng image quality evaluation on screen content image apply it is very few, therefore, this patent proposes that one kind has
Effect based on SCIs visual quality evaluation method, main contents include being able to maintain the stability of digital terminal, and guidance is based on
Coding, enhancing, the image of SCIs processing redirect the improvement and development of scheduling algorithm.
Marginal information is the fundamental of SCIs image-region high-frequency region, and therefore, marginal information can be used to reflect SCI
Distortion level, analyze SCI marginal information be solve SCI perceived quality forecasting problem a kind of feasible method.In addition,
AdaBoosting algorithm gets up multiple relatively weak classifiers combinations, is configured to a stronger regression machine, in algorithm essence
It improves a lot on degree.This patent is based on SCIs marginal information, passes through AdaBoosting algorithm, building prediction SCIs image
Without ginseng Environmental Evaluation Model, SCIs image quality evaluation is predicted.
Summary of the invention
This patent establish it is a kind of without with reference to measure come assess screen content image perceived quality (SCIs) method,
The purpose is to the SCIs edge features by extracting, and are carried out using marginal information of the Gauss model to SCI visual quality effective
The capture of human eye vision difference, and the mapping of two class multi-scale edges is calculated by neural network model, effectively SCI is tested in reflection
Distortion level, using AdaBooting reverse transmittance nerve network, training is carried out without ginseng Environmental Evaluation Model for SCIs image
The prediction of visual quality evaluation.Technical solution is as follows:
A kind of screen content image based on edge feature information without ginseng quality evaluating method, including the following steps:
The first step is calculated by DOG algorithm and obtains the mapping of the image border SCIs;
Second step, the edge map information that will acquire carries out local normalization, in the case where given SCIs, using part
Average subtraction and the normalized method of division calculate normalization pixel;
Third portion carries out the extraction of edge feature vector, normalizes edge from part using the fitting of L- square distribution estimation method
Marginal information is extracted in figure.
4th step returns device for quality perception maps feature vectors to subjective quality score using ABPNN, L- square is distributed
SCIs image n dimension its subjective quality assessment score of edge feature vector sum that estimation is extracted is put into training in ABPNN network, is instructing
Practice the stage, establishes regression model using edge feature vector and its corresponding subjective differences mean value Opinion Score (DMOS) value;Prediction
Stage will test the feature vector that SCIs is calculated and be input in the model of study, predicts SCIs image quality evaluation score.
Training stage in step 4, using following method:
Guarantee the quantity Q of Weak Classifier, and by the subjective differences mean value Opinion Score (DMOS) of the SCIs in training set
It is mapped to [0,1];Using AdaBoosting algorithm, pass through the original quality evaluation score and prediction matter to i-th of Weak Classifier
Difference between amount evaluation score is integrated, and estimates its error of quality appraisement Γ and corresponding distribution;It will be single weak using convex function
The weight of classifier is converted into corresponding weight, advantageously ensures that the weight of lower Weak Classifier is bigger, lower weak typing
The weight of device is smaller;Final SCIs image quality evaluation score output is calculated in conjunction with prediction output phase by weight.
It is proposed by the present invention effectively without ginseng screen content image quality evaluating method.This method is by human eye vision edge mould
Type is combined with AdaBoosing BP neural network.Two kinds of multi-scale edges of SCIs image are calculated first with DOG operator
Mapping.Then, after locally normalization edge mapping, estimation is distributed using L- square and is joined to be fitted its coefficient of the image border SCIs figure
Number is used as edge feature.Finally, using AdaBoosting BP neural network, by the n dimension edge feature vector of said extracted with
Its corresponding image quality evaluation score is put into training in model, and the objective quality of quality perception Feature Mapping to SCIs are commented
Valence score carries out SCIs image to predict its quality evaluation score.This method competitiveness with higher in SCI database
Can, it is stronger with the consistency of human visual system.
Detailed description of the invention
Fig. 1 SCIs example images figure
The flow chart that designs a model of Fig. 2 SCI image quality evaluation
Fig. 3 ABPNN algorithm flow chart
Fig. 4 ABPNN structure chart
Specific embodiment
SCIs image is as shown in Figure 1, detailed process trains as shown in Fig. 2, being divided into and predicts two stages, in training rank
Section, if the brightness of image of SCIs is I (x, y).
Evaluation procedure is as follows:
Step 1: edge graph calculates
This step calculates edge mapping using the difference algorithm (DOG algorithm) of Gaussian function.During image perception,
Human visual system is highly sensitive to marginal information, and details is the abundant component part of SCIs.Therefore, human vision DOG model
The difference of edge mapping can effectively be handled.In formula, I indicates the brightness mapping of test,Indicate convolution algorithm, image
Edge graph E can be indicated are as follows:
Wherein GσIt is isotropism Gaussian function and scale parameter σ
DOG model is expressed as
DOG=Gσ1(x,y)-Gσ2(x,y)
DOG algorithm can be used to improve visibility and the profile information of image at edge etc. as a kind of enhancing algorithm
Details.By using different parameter (σ1σ2) and the two kinds of edges mappings of 3 × 3 Gauss square window can effectively capture
To edge degradation.
Step 2: part normalization
Normalization is the integral process based on statistical property, as contrast normalizes.The Shandong of model can be improved in normalization
Stick calculates normalization pixel using local average subtraction and the normalized method of division in the case where given SCI image,
Its expression formula is
P (x, y) andRespectively indicate the original value and normal value at position (x, y);X ∈ 1,2,3 ..., X } and y
∈ 1,2,3 ..., and Y } representation space index;usAnd σsShow normalized mean value and standard deviation square value;C is a small constant,
The case where to avoid denominator being zero.μpAnd σpIt may be calculated:
W (r, z) | r=-R ..., R;Z=-Z ..., Z } unit volume Gauss window can be defined as;R and Z are set as
3, to ensure that window size is fairly small, smaller window size provides higher performance.This step passes through part normalization rule
Model SCIs image edge information.
Step 3: the distribution estimation of L square
This step is fitted the probability distribution of the image border SCIs figure by L square, and coefficient characterizes edge feature.L- square is one
Group is used to summarize the statistical series of probability distribution shape, including generalized Gaussian distribution, laplacian distribution, and L- moments estimation is a kind of
The statistical parameter estimation method of function admirable has stronger robustness to little deviation model, and the formula of square is defined as follows:
ρ1=η0
ρ2=2 η1-η0
ρ3=6 η2-6η1+η0
ρ4=20 η3-30η2+12η1-η0
L=(η1,η2,...,ηi), i=4
In addition, the distortion of image will affect the Multi-scale model of image, multi-scale information is introduced and is commented without ginseng picture quality
Preferable performance can be obtained in valence.To distortion SCI parameter attribute vector by
Feature=[LD1,LD2,...,LDi], i=7
The extraction that this step carries out edge graph using down-sampled method, and feature extraction is carried out on three scales.Its
Middle symbol D indicates the response of DOG, and L is L- square fitting coefficient.
Step 4: image quality evaluation model training
The one group of n dimension edge feature vector and its corresponding visual quality evaluation mark that this step is generated based on above-mentioned steps
Label, the calculating Environmental Evaluation Model of SCIs is established using homing method.In the training stage, edge feature and its corresponding subjectivity are utilized
Difference mean value Opinion Score (DMOS) value establishes regression model.Forecast period calculates test SCIs image through the above steps
Obtained edge feature vector, is input in training pattern, predicts SCIs in image quality evaluation score.
Wherein model chooses ABPNN frame, and particular content is, firstly, we will guarantee the quantity Q of Weak Classifier, and
By the O of the SCIs in training settrainjSubjective differences mean value Opinion Score (DMOS) is mapped to [0,1].AdaBoosting algorithm
Multiple relatively weak classifiers combinations are got up, a stronger regression machine is configured to, there is very big mention in arithmetic accuracy
Height, i-th of Weak Classifier is in training set training ΩtrainAnd Otrainj.Therefore, it can estimate O' in each training settrainiIt is pre-
Survey output.Its formula is as follows:
The calculating training error of each training set, this is considered as Γ distribution.Γ is in training set ΩtrainI-th of expression
Weak Classifier:
Using AdaBoosting algorithm, commented by the original quality evaluation score to i-th of Weak Classifier with forecast quality
Difference between valence score is integrated, and estimates its error of quality appraisement Γ and corresponding distribution.
Corresponding weight is converted by the weight of single Weak Classifier using convex function, advantageously ensures that lower weak typing
The weight of device is bigger, and the weight of lower Weak Classifier is smaller.The i component value λ of Weak ClassifieriIt can be expressed as
Wherein b and c is small normal number.
Therefore, final SCIs image quality evaluation score output O' can be calculated in conjunction with prediction output phase by weight, be determined
Justice are as follows:
Forecast period calls the model trained, and the n that SCIs image is obtained through the above steps ties up edge feature vector
Input model obtains its corresponding image quality evaluation score.
Claims (2)
1. a kind of screen content image based on edge feature information without ginseng quality evaluating method, including the following steps:
The first step is calculated by DOG algorithm and obtains the mapping of the image border SCIs;
Second step, the edge map information that will acquire carries out local normalization, in the case where given SCIs, using local average
Subtraction and the normalized method of division calculate normalization pixel;
Third step carries out the extraction of edge feature vector, using the fitting of L- square distribution estimation method from part normalization edge graph
Extract marginal information.
4th step returns device using ABPNN and quality is perceived maps feature vectors to subjective quality score, L- square is distributed and is estimated
SCIs image n dimension its subjective quality assessment score of edge feature vector sum of extraction is put into training in ABPNN network, in training rank
Section establishes regression model using edge feature vector and its corresponding subjective differences mean value Opinion Score (DMOS) value;Forecast period,
The feature vector that SCIs is calculated will be tested to be input in the model of study, predict SCIs image quality evaluation score.
2. the method according to claim 1, wherein the training stage in step 4, using following method:
Guarantee the quantity Q of Weak Classifier, and the subjective differences mean value Opinion Score (DMOS) of the SCIs in training set is mapped to [0,
1];Using AdaBoosting algorithm, pass through the original quality evaluation score and forecast quality evaluation point to i-th of Weak Classifier
Difference between number is integrated, and estimates its error of quality appraisement Γ and corresponding distribution;Using convex function by single Weak Classifier
Weight is converted into corresponding weight, advantageously ensures that the weight of lower Weak Classifier is bigger, the weight of lower Weak Classifier
It is smaller;Final SCIs image quality evaluation score output is calculated in conjunction with prediction output phase by weight.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910704654.5A CN110473181A (en) | 2019-07-31 | 2019-07-31 | Screen content image based on edge feature information without ginseng quality evaluating method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910704654.5A CN110473181A (en) | 2019-07-31 | 2019-07-31 | Screen content image based on edge feature information without ginseng quality evaluating method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110473181A true CN110473181A (en) | 2019-11-19 |
Family
ID=68509473
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910704654.5A Pending CN110473181A (en) | 2019-07-31 | 2019-07-31 | Screen content image based on edge feature information without ginseng quality evaluating method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110473181A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112348809A (en) * | 2020-11-30 | 2021-02-09 | 天津大学 | No-reference screen content image quality evaluation method based on multitask deep learning |
CN112950567A (en) * | 2021-02-25 | 2021-06-11 | 北京金山云网络技术有限公司 | Quality evaluation method, quality evaluation device, electronic device, and storage medium |
CN113222902A (en) * | 2021-04-16 | 2021-08-06 | 北京科技大学 | No-reference image quality evaluation method and system |
CN117275130A (en) * | 2023-11-17 | 2023-12-22 | 长春金融高等专科学校 | Intelligent access control verification system based on face recognition |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2015037973A1 (en) * | 2013-09-12 | 2015-03-19 | Data Calibre Sdn Bhd | A face identification method |
CN104658002A (en) * | 2015-03-10 | 2015-05-27 | 浙江科技学院 | Non-reference image objective quality evaluation method |
CN107680077A (en) * | 2017-08-29 | 2018-02-09 | 南京航空航天大学 | A kind of non-reference picture quality appraisement method based on multistage Gradient Features |
CN108664840A (en) * | 2017-03-27 | 2018-10-16 | 北京三星通信技术研究有限公司 | Image-recognizing method and device |
CN108830829A (en) * | 2018-05-08 | 2018-11-16 | 天津大学 | Combine the reference-free quality evaluation algorithm of a variety of edge detection operators |
CN109544504A (en) * | 2018-10-16 | 2019-03-29 | 天津大学 | Screen picture quality evaluating method based on rarefaction representation |
CN110046673A (en) * | 2019-04-25 | 2019-07-23 | 上海大学 | No reference tone mapping graph image quality evaluation method based on multi-feature fusion |
-
2019
- 2019-07-31 CN CN201910704654.5A patent/CN110473181A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2015037973A1 (en) * | 2013-09-12 | 2015-03-19 | Data Calibre Sdn Bhd | A face identification method |
CN104658002A (en) * | 2015-03-10 | 2015-05-27 | 浙江科技学院 | Non-reference image objective quality evaluation method |
CN108664840A (en) * | 2017-03-27 | 2018-10-16 | 北京三星通信技术研究有限公司 | Image-recognizing method and device |
CN107680077A (en) * | 2017-08-29 | 2018-02-09 | 南京航空航天大学 | A kind of non-reference picture quality appraisement method based on multistage Gradient Features |
CN108830829A (en) * | 2018-05-08 | 2018-11-16 | 天津大学 | Combine the reference-free quality evaluation algorithm of a variety of edge detection operators |
CN109544504A (en) * | 2018-10-16 | 2019-03-29 | 天津大学 | Screen picture quality evaluating method based on rarefaction representation |
CN110046673A (en) * | 2019-04-25 | 2019-07-23 | 上海大学 | No reference tone mapping graph image quality evaluation method based on multi-feature fusion |
Non-Patent Citations (2)
Title |
---|
LIXIONGLIU等: "Blind image quality assessment by relative gradient statistics and adaboosting neural network", 《SIGNAL PROCESSING: IMAGE COMMUNICATION》 * |
YAQI LV等: "Difference of Gaussian statistical features based blind image quality assessment: A deep learning approach", 《2015 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP)》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112348809A (en) * | 2020-11-30 | 2021-02-09 | 天津大学 | No-reference screen content image quality evaluation method based on multitask deep learning |
CN112348809B (en) * | 2020-11-30 | 2023-05-23 | 天津大学 | No-reference screen content image quality evaluation method based on multitask deep learning |
CN112950567A (en) * | 2021-02-25 | 2021-06-11 | 北京金山云网络技术有限公司 | Quality evaluation method, quality evaluation device, electronic device, and storage medium |
CN113222902A (en) * | 2021-04-16 | 2021-08-06 | 北京科技大学 | No-reference image quality evaluation method and system |
CN113222902B (en) * | 2021-04-16 | 2024-02-02 | 北京科技大学 | No-reference image quality evaluation method and system |
CN117275130A (en) * | 2023-11-17 | 2023-12-22 | 长春金融高等专科学校 | Intelligent access control verification system based on face recognition |
CN117275130B (en) * | 2023-11-17 | 2024-01-26 | 长春金融高等专科学校 | Intelligent access control verification system based on face recognition |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110473181A (en) | Screen content image based on edge feature information without ginseng quality evaluating method | |
CN110662484B (en) | System and method for whole body measurement extraction | |
CN108229509B (en) | Method and device for identifying object class and electronic equipment | |
US20200042770A1 (en) | Facial recognition method and apparatus, electronic device, and storage medium | |
CN108765423B (en) | Convolutional neural network training method and device | |
CN109389135B (en) | Image screening method and device | |
CN104658002B (en) | Non-reference image objective quality evaluation method | |
CN113674421B (en) | 3D target detection method, model training method, related device and electronic equipment | |
CN104036493B (en) | No-reference image quality evaluation method based on multifractal spectrum | |
CN101567044A (en) | Method for detecting quality of human face image | |
CN103942525A (en) | Real-time face optimal selection method based on video sequence | |
CN103714349A (en) | Image recognition method based on color and texture features | |
CN106600597A (en) | Non-reference color image quality evaluation method based on local binary pattern | |
CN113449704B (en) | Face recognition model training method and device, electronic equipment and storage medium | |
CN110570435A (en) | method and device for carrying out damage segmentation on vehicle damage image | |
CN104063686A (en) | System and method for performing interactive diagnosis on crop leaf segment disease images | |
CN111860091A (en) | Face image evaluation method and system, server and computer readable storage medium | |
CN112633221A (en) | Face direction detection method and related device | |
CN110991434B (en) | Self-service terminal certificate identification method and device | |
CN115375663A (en) | Image quality evaluation method based on depth mutual learning and double-scale feature fusion | |
CN108537990B (en) | All-in-one machine fraud judgment method, device, equipment and computer readable storage medium | |
Yang et al. | EHNQ: Subjective and objective quality evaluation of enhanced night-time images | |
CN117078670A (en) | Production control system of cloud photo frame | |
CN112581001B (en) | Evaluation method and device of equipment, electronic equipment and readable storage medium | |
CN111582278A (en) | Portrait segmentation method and device and electronic equipment |
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: 20191119 |