CN109410203A - A kind of picture picture quality detection method based on machine learning - Google Patents
A kind of picture picture quality detection method based on machine learning Download PDFInfo
- Publication number
- CN109410203A CN109410203A CN201811287856.6A CN201811287856A CN109410203A CN 109410203 A CN109410203 A CN 109410203A CN 201811287856 A CN201811287856 A CN 201811287856A CN 109410203 A CN109410203 A CN 109410203A
- Authority
- CN
- China
- Prior art keywords
- picture
- image quality
- feature
- library
- machine learning
- 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.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of picture picture quality detection method based on machine learning, belongs to picture recognition field, described method includes following steps: collecting the data of problematic picture;It is analyzed and processed according to the image data of collection and classification processing;Picture after analysis processing and classification processing is established picture preliminary image quality contrasting detection library;Extract the image quality feature of each picture in preliminary image quality library;Image quality feature database is established according to the image quality feature of picture;Picture is detected according to image quality feature database and preliminary image quality library;Detection output result.Pass through the method for machine learning, feature learning is extracted from previous image quality function problem picture, quickly which class problem positioning belongs to, and carries out feature learning to image quality function problem picture using depth learning technology, and classify to the feature learnt, obtain deep learning model;Camera image quality function picture is judged using deep learning model and exports result.
Description
[technical field]
The present invention relates to picture recognition fields, and in particular to a kind of picture picture quality detection method based on machine learning.
[background technique]
With the development of society, people increasingly like taking pictures, but the difference of the shooting technology due to everyone, thus
So that the problem of this or that occurs in the photo taken, is not usually able to satisfy the demand of people.With the development of internet technology,
Picture expresses the advantages such as intuitive, abundant in content because it has with respect to text, is answered extensively in more and more webpages and application
With.But existing camera can only take pictures to picture, could not automatically be classified to picture or exposure image divides
Class or handsome choosing etc. checking and selecting so that people carry out each again after having clapped photo, to greatly increase
The workload of people.
With the quickening of social step, the time of people is more and more valuable, and people is in out on tours or outdoor activities, often
More photo can be often arranged, but since the factors such as technology and environment can make some photos expose, exposure
Picture can usually occupy the memory space of equipment, while selecting picture to user and making troubles, and therefore, it is necessary to design one kind certainly
The method of dynamic identification picture image quality, automatic point so as to complete picture is carried out handsome choosing.
[summary of the invention]
The present invention is directed to disclose a kind of picture picture quality detection method based on machine learning, existing exposure image image quality is solved
The technical issues of being unable to automatic screening and classification.
The technical scheme adopted by the invention is as follows:
A kind of picture picture quality detection method based on machine learning, described method includes following steps:
Collect the data of problematic picture;
It is analyzed and processed according to the image data of collection and classification processing;
Picture after analysis processing and classification processing is established picture preliminary image quality contrasting detection library;
Extract the image quality feature of each picture in preliminary image quality library;
Image quality feature database is established according to the image quality feature of picture;
Picture is detected according to image quality feature database and preliminary image quality library;
Whether artificial detection output picture upon selection manually classifies to the image data after detection for artificial selection
Processing, output point come as a result, when not selecting, direct output test result.
Further, the data for collecting problematic picture include collecting the picture of previous appearance image quality exposure problems
The picture to go wrong with the daily image quality fed back of user.
Further, when the data for collecting problematic picture, it is automatic below each picture or in catalogue or
Manually using the content of the specific image quality problem of explanatory note picture.
Further, the image data according to collection is analyzed and processed the detailed process with classification processing are as follows: inspection
Characteristic color region is surveyed, picture feature is described and is analyzed by color histogram and histogram of gradients, then according to face
Color Histogram and histogram of gradients again classify to picture.
Further, the detailed process for establishing picture preliminary image quality contrasting detection library is that elder generation is according to image quality color diagram
Classify with the explanatory note of picture, then carried out establishing picture function model according to different types of picture, then building
The picture function model stood is stored;The classification is to be carried out being divided into major class, major class the inside according to the description of text
There are several groups, forms association class between group and group.
Further, the picture function model is after in addition the later period is collected into image quality problem image, image quality problem image
It is automatically added to picture function library, automatic training study increases picture function library performance.
Further, the image quality characteristic extraction procedure of each picture is in the preliminary image quality library of extraction, to each picture
Image quality feature extraction is carried out, the image quality feature of extraction and the image quality comment content of picture are compared, when endless whole is consistent
When, secondary image quality feature extraction is carried out to the picture.
Further, the process for establishing image quality feature database are as follows: classification processing first is carried out to image quality feature, then basis
Classification situation establishes image quality characteristic model, and then image quality characteristic model is stored, and it is special the image quality newly extracted occur in the later period
When sign, new image quality feature is automatically added to image quality characteristic model, improves image quality characteristic model data accuracy.
Further, the process that picture is detected according to image quality feature database and preliminary image quality library are as follows:
Picture is carried out preliminary analysis processing and classification processing, treated picture and the preliminary image quality contrasting detection of picture
Library is detected, when detection picture go wrong, exported, when there is no problem, to picture carry out image quality extraction, then with picture
Matter feature is established image quality feature database and is compared, and comparison structure data are then exported.
Further, when the artificial detection output picture, the result data of artificial detection is different from the result of original detection
When, artificial selection is trained again using the different picture of result as the data of problem picture.
It is had the advantage that using technical solution of the present invention
The present invention extracts feature learning from previous image quality function problem picture by the method for machine learning, quickly fixed
Which class problem position belongs to, and carries out feature learning to image quality function problem picture using depth learning technology, and to the spy learnt
Sign is classified, and deep learning model is obtained;Using deep learning model to camera image quality function picture carry out judgement and it is defeated
Result out.By inputting different types of image quality problem picture early period and learning the different types of feature of mass data automatically, obtain
To learning model, subsequent need to provide picture and input whether i.e. exportable image quality result passes through, and eliminate manpower, substantially increase
Efficiency realizes automatic classification so as to sort out the picture of some image quality differences such as exposure from the data of magnanimity.
[Detailed description of the invention]
Fig. 1 is a kind of picture picture quality detection method flow chart based on machine learning of the present invention.
The present invention that the following detailed description will be further explained with reference to the above drawings.
[specific embodiment]
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
It should be noted that the term used in embodiments of the present invention is only merely for the mesh of description specific embodiment
, it is not intended to limit the invention." the one of the embodiment of the present invention and singular used in the attached claims
Kind ", " described " and "the" are also intended to including most forms, unless the context clearly indicates other meaning.It is also understood that this
Term "and/or" used herein refers to and includes one or more associated any or all possible group for listing project
It closes.
Term " includes " in description and claims of this specification and above-mentioned attached drawing and " having " and they appoint
What is deformed, it is intended that is covered and non-exclusive is included.Such as contain the process, method, system, production of a series of steps or units
Product or equipment are not limited to listed step or unit, but optionally further comprising the step of not listing or unit, or
Optionally further comprising other step or units intrinsic for these process, methods, product or equipment.
It should be noted that the embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings,
In which the same or similar labels are throughly indicated same or similar element or elements with the same or similar functions.Under
Face is exemplary by reference to the embodiment that attached drawing describes, and for explaining only the invention, and cannot be construed to of the invention
Limitation.
Following disclosure provides many different embodiments or example is used to realize different structure of the invention.For letter
Change disclosure of the invention, hereinafter the component of specific examples and setting are described.Certainly, they are merely examples, and
It is not intended to limit the present invention.In addition, the present invention can in different examples repeat reference numerals and/or letter.It is this heavy
It is for purposes of simplicity and clarity, itself not indicate the relationship between discussed various embodiments and/or setting again.
Deep learning realizes that complicated function approaches and input data table by learning the nonlinear network structure of deep layer a kind of
Sign, shows the learning ability of powerful data set substantive characteristics.Convolutional neural networks, as a kind of typical deep learning side
Method is a specially designed multilayer perceptron for two dimensional image processing.Convolutional neural networks do not need artificially to participate in feature
Selection process, can automatically learn mass data concentration target signature.The detection method of deep learning is defeated by early period
Enter different types of image quality problem picture and learn the different types of feature of mass data automatically, obtains learning model, subsequent
Picture need to be provided and input whether i.e. exportable image quality result passes through, equipment and manpower is eliminated, greatly improve the efficiency.Pass through benefit
With the method for deep learning, feature learning is extracted from previous image quality function problem picture, quickly which class problem positioning belongs to.Benefit
Feature learning is carried out to image quality function problem picture with depth learning technology, and is classified to the feature learnt, depth is obtained
Spend learning model;Camera image quality function picture is judged using deep learning model and exports result.With terminal visitor
Requirement of the family to image effect is higher and higher, and using currently burning hoter deep learning identification technology, deep learning is in recent years
Great development is achieved in computer vision field, it is right by constructing a kind of deep-neural-network similar to human brain structure
The data such as the image of input are analyzed deeply, by profound neural network and a large amount of training data, improve spy
The extraction of sign as a result, also improve the accuracy of target classification simultaneously.
Neural network is made of a lot of layers, and with the raising of training scale, the number of plies also can be more and more, and precision can also be got over
Come higher.First layer neuron can pass to processed data second layer neuron, and second layer neuron goes to complete oneself
Processing task, in this way processing be continued until the last layer, final output.A perceptron is had in input layer
Thing is working, and perceptron can receive binary input, can according to need selection and more or less inputs.Finally produce
Raw binary system output.The simple rule of a weight can be introduced in the process of output simultaneously to calculate output.Such as this patent
In to carry out image quality detection, by the input of picture to be detected, picture is split into multiple fritters, is gone point by neural network
Analysis, such as stain, extract in shape, shade, the size of size and the position in picture etc. inputted, then
Weight size final output is provided by correspondence according to system trained before.
Referring to Fig. 1, being a kind of picture picture quality detection method process based on machine learning provided in an embodiment of the present invention
Figure, described method includes following steps:
Step S10: collecting the data of problematic picture, and the data for collecting problematic picture include collecting previous appearance picture
The picture that the picture of matter exposure problems and the daily image quality fed back of user go wrong.Collect the data of problematic picture
When, either automatically or manually using the content of the specific image quality problem of explanatory note picture below each picture or in catalogue.
For example people are when taking pictures, bad due to dimming, taking the photo come is exposure, therefore this photo taken should not belong to
Photo is needed in people, so that an other file can be deleted or is referred to for it automatically.
Step S20: being analyzed and processed according to the image data of collection and classification processing.According to the image data of collection into
The detailed process of row analysis processing and classification processing are as follows: detection characteristic color region passes through color histogram and histogram of gradients
Picture feature is described and is analyzed, is then classified again to picture according to color histogram and histogram of gradients.It establishes
The detailed process in picture preliminary image quality contrasting detection library is first to be classified according to the explanatory note of image quality color diagram and picture,
Then it is carried out establishing picture function model according to different types of picture, then established picture function model is deposited
Storage;The classification is to carry out being divided into major class according to the description of text, there is several groups inside major class, group and group it
Between form association class.Picture function model carries out certainly picture function model after in addition the later period is collected into image quality problem image
Dynamic addition data, improve accuracy.
Picture function library is mainly the library that classification foundation is carried out according to the function of picture, and it is existing for specifically establishing the model in library
Deep learning model.By input early period different types of image quality problem picture and automatically study mass data it is different types of
Feature obtains learning model, and subsequent need to provide picture and input whether i.e. exportable image quality result passes through, and eliminates equipment and people
Power greatly improves the efficiency.
Step S30: the picture after analysis processing and classification processing is established picture preliminary image quality contrasting detection library.Picture function
For energy model after in addition the later period is collected into image quality problem image, image quality problem image is automatically added to picture function library, automatic to instruct
Practice study, increases picture function library performance.First from the quality of the preliminary interpretation picture of angle of color, so as to quick well
Classification and screening picture, speed is faster.
Step S40: extracting the image quality feature of each picture in preliminary image quality library, extracts each picture in preliminary image quality library
Image quality characteristic extraction procedure is to carry out image quality feature extraction to each picture, the image quality feature of extraction and the image quality of picture are said
Bright word content comparison carries out secondary image quality feature extraction to the picture when endless whole is consistent.According to the different of picture
Problem can directly extract corresponding feature, and corresponding feature represents the image quality problem of the picture.Characteristic procedure by locating in advance
Reason, characteristic point are chosen, the processes such as feature point extraction.
Step S50: image quality feature database is established according to the image quality feature of picture, establishes the process of image quality feature database are as follows: first right
Image quality feature carries out classification processing, then establishes image quality characteristic model according to classification situation, then image quality characteristic model is carried out
Storage, when there is the image quality feature newly extracted in the later period, new image quality feature is automatically added to image quality characteristic model, improves image quality
Characteristic model data accuracy.Classified to image quality feature and mainly classified to the image quality problem, for example is picture
Matter has stain or image quality to classify than darker etc..Establishing image quality characteristic model is the characteristic model for establishing deep learning,
Basis is provided for the deep learning comparison in later period.
Step S60: detecting picture according to image quality feature database and preliminary image quality library, according to image quality feature database and tentatively
The process that picture is detected in image quality library are as follows:
Picture is carried out preliminary analysis processing and classification processing, treated picture and the preliminary image quality contrasting detection of picture
Library is detected, when detection picture go wrong, exported, when there is no problem, to picture carry out image quality extraction, then with picture
Matter feature is established image quality feature database and is compared, and comparison structure data are then exported.
Step S70: whether artificial detection exports picture for artificial selection, upon selection, manually to the image data after detection
Classification processing is carried out, output point comes as a result, when not selecting, direct output test result.When artificial detection exports picture, manually
When the result data of detection and the result difference of former detection, artificial selection is using the different picture of result as the data of problem picture
It is trained again.
Claims (10)
1. a kind of picture picture quality detection method based on machine learning, which is characterized in that described method includes following steps:
Collect the data of problematic picture;
It is analyzed and processed according to the image data of collection and classification processing;
Picture after analysis processing and classification processing is established picture preliminary image quality contrasting detection library;
Extract the image quality feature of each picture in preliminary image quality library;
Image quality feature database is established according to the image quality feature of picture;
Picture is detected according to image quality feature database and preliminary image quality library;
Whether artificial detection exports picture for artificial selection, upon selection, manually carries out classification processing to the image data after detection,
Output point comes as a result, when not selecting, direct output test result.
2. a kind of picture picture quality detection method based on machine learning according to claim 1, which is characterized in that the receipts
The data for collecting problematic picture include collecting the picture and the daily picture fed back of user of previous appearance image quality exposure problems
The picture that matter goes wrong.
3. a kind of picture picture quality detection method based on machine learning according to claim 2, it is characterised in that: the receipts
When collecting the data of problematic picture, the tool of explanatory note picture is either automatically or manually used below each picture or in catalogue
The content of body image quality problem.
4. a kind of picture picture quality detection method based on machine learning according to claim 1, it is characterised in that: described
The detailed process with classification processing is analyzed and processed according to the image data of collection are as follows: detection characteristic color region passes through color
Histogram and histogram of gradients are described and analyze to picture feature, then right again according to color histogram and histogram of gradients
Picture is classified.
5. a kind of picture picture quality detection method based on machine learning according to claim 4, it is characterised in that: described to build
The detailed process in vertical picture preliminary image quality contrasting detection library is first to be divided according to the explanatory note of image quality color diagram and picture
Then class carries out establishing picture function model according to different types of picture, then established picture function model is carried out
Storage;The classification is to be carried out being divided into major class according to the description of text, has several groups, group and group inside major class
Between form association class.
6. a kind of picture picture quality detection method based on machine learning according to claim 5, it is characterised in that: the figure
For piece functional mode after in addition the later period is collected into image quality problem image, image quality problem image is automatically added to picture function library, from
Dynamic training study, increases picture function library performance.
7. a kind of picture picture quality detection method based on machine learning according to claim 1, which is characterized in that described to mention
The image quality characteristic extraction procedure for taking each picture in preliminary image quality library is image quality feature extraction to be carried out to each picture, extraction
Image quality feature and the image quality comment content of picture compare, when endless whole is consistent, secondary image quality is carried out to the picture
Feature extraction.
8. a kind of picture picture quality detection method based on machine learning according to claim 1, it is characterised in that: described to build
The process of vertical image quality feature database are as follows: classification processing first is carried out to image quality feature, image quality character modules are then established according to classification situation
Then type stores image quality characteristic model, when there is the image quality feature newly extracted in the later period, new image quality feature adds automatically
It is added to image quality characteristic model, improves image quality characteristic model data accuracy.
9. a kind of picture picture quality detection method based on machine learning according to claim 1, which is characterized in that described
The process that picture is detected according to image quality feature database and preliminary image quality library are as follows:
Picture is carried out preliminary analysis processing and classification processing, treated picture and picture preliminary image quality contrasting detection library into
Row detection goes wrong when detecting picture, is exported, and when there is no problem, carries out image quality extraction to picture, then special with image quality
Sign is established image quality feature database and is compared, and comparison structure data are then exported.
10. a kind of picture picture quality detection method based on machine learning according to claim 1, which is characterized in that described
When artificial detection exports picture, when the result data of artificial detection is with the former result difference detected, artificial selection is result difference
Picture trained again as the data of problem picture.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811287856.6A CN109410203A (en) | 2018-10-31 | 2018-10-31 | A kind of picture picture quality detection method based on machine learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811287856.6A CN109410203A (en) | 2018-10-31 | 2018-10-31 | A kind of picture picture quality detection method based on machine learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109410203A true CN109410203A (en) | 2019-03-01 |
Family
ID=65470692
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811287856.6A Withdrawn CN109410203A (en) | 2018-10-31 | 2018-10-31 | A kind of picture picture quality detection method based on machine learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109410203A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111062957A (en) * | 2019-10-28 | 2020-04-24 | 广西科技大学鹿山学院 | Non-classical receptive field based contour detection method |
CN111077326A (en) * | 2019-12-20 | 2020-04-28 | 中译语通文娱科技(青岛)有限公司 | Detection technology based on AI technology pre-perception capability |
WO2020259209A1 (en) * | 2019-06-26 | 2020-12-30 | 平安科技(深圳)有限公司 | Fundus image recognition method, apparatus and device, and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20110056999A (en) * | 2009-11-23 | 2011-05-31 | 한국전자통신연구원 | Method and apparatus for malicious photo filtering using semantic features of digital photo |
CN106897748A (en) * | 2017-03-02 | 2017-06-27 | 上海极链网络科技有限公司 | Face method for evaluating quality and system based on deep layer convolutional neural networks |
CN108257117A (en) * | 2018-01-02 | 2018-07-06 | 中兴通讯股份有限公司 | The evaluating method and device of image exposure degree |
-
2018
- 2018-10-31 CN CN201811287856.6A patent/CN109410203A/en not_active Withdrawn
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20110056999A (en) * | 2009-11-23 | 2011-05-31 | 한국전자통신연구원 | Method and apparatus for malicious photo filtering using semantic features of digital photo |
CN106897748A (en) * | 2017-03-02 | 2017-06-27 | 上海极链网络科技有限公司 | Face method for evaluating quality and system based on deep layer convolutional neural networks |
CN108257117A (en) * | 2018-01-02 | 2018-07-06 | 中兴通讯股份有限公司 | The evaluating method and device of image exposure degree |
Non-Patent Citations (3)
Title |
---|
党亚民: "《海岛礁测绘技术与方法》", 30 November 2012, 测绘出版社 * |
徐敏霞: "手机照片质量评价算法研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
樊田田: "证件照的质量预检测***", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020259209A1 (en) * | 2019-06-26 | 2020-12-30 | 平安科技(深圳)有限公司 | Fundus image recognition method, apparatus and device, and storage medium |
CN111062957A (en) * | 2019-10-28 | 2020-04-24 | 广西科技大学鹿山学院 | Non-classical receptive field based contour detection method |
CN111062957B (en) * | 2019-10-28 | 2024-02-09 | 广西科技大学鹿山学院 | Non-classical receptive field contour detection method |
CN111077326A (en) * | 2019-12-20 | 2020-04-28 | 中译语通文娱科技(青岛)有限公司 | Detection technology based on AI technology pre-perception capability |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Wiley et al. | Computer vision and image processing: a paper review | |
CN112926405B (en) | Method, system, equipment and storage medium for detecting wearing of safety helmet | |
CN108537191B (en) | Three-dimensional face recognition method based on structured light camera | |
CN109447977B (en) | Visual defect detection method based on multispectral deep convolutional neural network | |
CN104182765B (en) | Internet image driven automatic selection method of optimal view of three-dimensional model | |
CN110598800A (en) | Garbage classification and identification method based on artificial intelligence | |
CN109410203A (en) | A kind of picture picture quality detection method based on machine learning | |
CN106127780A (en) | A kind of curved surface defect automatic testing method and device thereof | |
CN106407369A (en) | Photo management method and system based on deep learning face recognition | |
CN109684913A (en) | A kind of video human face mask method and system based on community discovery cluster | |
CN105678344A (en) | Intelligent classification method for power instrument equipment | |
CN112837282A (en) | Small sample image defect detection method based on cloud edge cooperation and deep learning | |
CN110348505B (en) | Vehicle color classification model training method and device and vehicle color identification method | |
CN109165564B (en) | Electronic photo album, generation method, system, storage medium and computer equipment | |
CN111401149A (en) | Lightweight video behavior identification method based on long-short-term time domain modeling algorithm | |
CN111914902A (en) | Traditional Chinese medicine identification and surface defect detection method based on deep neural network | |
CN109472790A (en) | A kind of machine components defect inspection method and system | |
CN113158969A (en) | Apple appearance defect identification system and method | |
CN107578067A (en) | A kind of recognition methods of pipe tobacco component and device based on convolutional neural networks | |
CN107944373A (en) | A kind of video anomaly detection method based on deep learning | |
Tyukin et al. | Exploring automated pottery identification [Arch-I-Scan] | |
CN109741351A (en) | A kind of classification responsive type edge detection method based on deep learning | |
CN112949369A (en) | Mass face gallery retrieval method based on man-machine cooperation | |
CN111652242A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
CN107133586A (en) | A kind of authentication method of calligraphy class historical relic |
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 | ||
WW01 | Invention patent application withdrawn after publication | ||
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190301 |