CN113159150A - Branch intervention pearl sorting method based on multi-algorithm integration - Google Patents
Branch intervention pearl sorting method based on multi-algorithm integration Download PDFInfo
- Publication number
- CN113159150A CN113159150A CN202110386522.XA CN202110386522A CN113159150A CN 113159150 A CN113159150 A CN 113159150A CN 202110386522 A CN202110386522 A CN 202110386522A CN 113159150 A CN113159150 A CN 113159150A
- Authority
- CN
- China
- Prior art keywords
- pearl
- classification
- divergence
- models
- sorting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 230000010354 integration Effects 0.000 title claims abstract description 10
- 238000012549 training Methods 0.000 claims abstract description 16
- 238000011156 evaluation Methods 0.000 claims abstract description 8
- 238000012795 verification Methods 0.000 claims abstract description 8
- 238000012360 testing method Methods 0.000 claims abstract description 6
- 238000002372 labelling Methods 0.000 claims abstract description 5
- 239000011049 pearl Substances 0.000 claims description 72
- 230000007547 defect Effects 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 230000000717 retained effect Effects 0.000 claims description 2
- 238000007636 ensemble learning method Methods 0.000 claims 1
- 238000013527 convolutional neural network Methods 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- -1 color Substances 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000002932 luster Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/34—Sorting according to other particular properties
- B07C5/3412—Sorting according to other particular properties according to a code applied to the object which indicates a property of the object, e.g. quality class, contents or incorrect indication
-
- 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
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/06—Recognition of objects for industrial automation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Evolutionary Biology (AREA)
- Educational Administration (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Game Theory and Decision Science (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
Abstract
A bifurcation intervention pearl sorting method based on multi-algorithm integration comprises the following steps: step 1: respectively recording pearl data sets as Class1-7 according to categories, labeling each picture, and recording the data sets according to the ratio of 6: 2: 2, dividing the ratio into a training set, a verification set and a test set; step 2: selecting three models ResNet50, SE-ResNet50 and Vgg16 which are mainstream and advanced at present for independent training, wherein the models ResNet50, SE-ResNet50 and Vgg16 can finish pearl sorting tasks and respectively store optimal models; and step 3: the obtained three optimal models are used for an arbitration system, namely divergence arbitration is carried out on prediction results given by the constructed primary system and the constructed secondary system; and 4, step 4: in an experimental verification stage, a divergence accuracy index alpha and an additional cost index beta are used as evaluation indexes of system divergence to evaluate the overall performance of a system; and 5: selecting a system combination which achieves higher sorting precision and requires less labor cost as a final pearl sorting system according to the evaluation indexes in the step four; step 6: and outputting a final classification result. The invention finds out the error prediction possibly existing in machine classification through divergence based on a plurality of redundant algorithms, and corrects the prediction output with the intervention of least labor cost, thereby improving the classification precision of the whole pearl sorting.
Description
Technical Field
The invention relates to a pearl detection and classification method which is suitable for solving the problem that pure deep learning is limited in improvement of pearl sorting precision by utilizing a human intervention scheme.
Background
Traditional pearl sorting is mainly performed manually, and practitioners with abundant experience observe various appearance characteristics of pearls such as color, luster, shape, texture and the like through naked eyes, and then classify the pearls according to certain predetermined grade rules. It is easy to recognize that the observation and sorting of a large number of pearls whose appearance differences may not be too great is a repeated and tedious work; and with the increase of time, the fatigue of people can greatly influence the sorting accuracy. However, misclassifying a high-grade pearl as a low grade inherently reduces the value of the sale, and conversely, misclassifying a low-grade pearl as a high grade also affects the quality and the reputation of the product. Therefore, it is necessary to introduce an automated sorting method for efficient and stable sorting.
Since pearl sorting is mainly based on pearl appearance, automated sorting methods are therefore mostly based on methods and techniques of machine vision. The manual marking of pearl features in the traditional computer vision method depends on expert knowledge, and the cost is relatively high. In recent years, deep learning methods, particularly convolutional neural networks CNN, have been highly successful in image classification, which can automatically generate useful features, thereby saving a lot of manual labeling work. The current automatic classification system trained by the deep learning method achieves the sorting precision of 92.57%, but the requirement of enterprises is still insufficient. For pearl production enterprises, the improvement of the sorting accuracy by 1 percent also has great economic value. The price of the pearls is high, the price of the pearls in different grades is even different by hundreds of times, the quality of the pearls with flaws is seriously lost when the pearls with superior grades are divided into inferior grades, and the quality of the pearls with flaws is seriously lost when the pearls with flaws are divided into superior grades. Therefore, the method provides a practical demand for improving the pearl sorting accuracy.
It will be appreciated that further improvements in accuracy on this basis present significant difficulties. On one hand, the reasons are technical, and the current artificial intelligence technology based on deep learning still has the defects of difficult interpretation, poor robustness and the like, so that an algorithm with higher precision is difficult to train, and the existing error output cannot be predicted. On the other hand, the reason is that the cost is high, and a large number of accurately labeled pearl data sets are difficult to obtain. The analysis means that if the pearl sorting accuracy needs to be further improved, the existing pure deep learning technical framework is likely to need to be jumped out.
Disclosure of Invention
The invention provides a bifurcation intervention pearl sorting method based on multi-algorithm integration, which aims to overcome the defects of the prior art, namely the practical requirement on pearl sorting accuracy and the limitation of improving the accuracy under the existing framework.
The pearl sorting method based on the multiple redundant algorithms can use a method based on the multiple redundant algorithms without changing the algorithms, pearls which are judged to be branched by the multiple algorithms are identified as sorting errors by introducing the redundant automatic sorting algorithms, and the parts with the sorting errors are judged again with the least labor cost, so that an idea is provided for improving the precision of the pearl sorting method.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a bifurcation intervention pearl sorting method based on multi-algorithm integration comprises the following steps:
the method comprises the following steps: respectively recording pearl data sets as Class1-7 according to categories, labeling each picture, and recording the data sets according to the ratio of 6: 2: 2, dividing the ratio into a training set, a verification set and a test set;
step two: selecting three models ResNet50, SE-ResNet50 and Vgg16 which are mainstream and advanced at present for independent training, wherein the models ResNet50, SE-ResNet50 and Vgg16 can finish pearl sorting tasks and respectively store optimal models;
step three: the obtained three optimal models are used for an arbitration system, namely divergence arbitration is carried out on prediction results given by the constructed primary system and the constructed secondary system;
step four: in an experimental verification stage, a divergence accuracy index alpha and an additional cost index beta are used as evaluation indexes of system divergence to evaluate the overall performance of a system;
step five: selecting a system combination which achieves higher sorting precision and requires less labor cost as a final pearl sorting system according to the evaluation indexes in the step four;
step six: and outputting a final classification result.
The pearl sorting method based on the multiple redundant algorithms is provided under the condition that the algorithms are not changed, the pearl which is judged to be branched by the multiple algorithms is identified as a sorting fault by introducing the redundant automatic sorting algorithm, and the part with the sorting fault is judged again with the least labor cost, so that an idea is provided for improving the precision of the pearl sorting method. The convolutional neural network CNN can extract the features in the pictures and identify the information in the new pictures, so that the pictures with unknown labels can be classified after a large number of pictures with known labels are trained; the effective intervention of the human can correct the prediction deviation problem existing in the system.
Compared with the prior art, the technical scheme of the invention has the advantages that:
(1) the situation that only one type of algorithm is used for predicting classification errors can be detected more simply and effectively by utilizing the divergence of a plurality of redundant algorithms;
(2) a small amount of labor cost is introduced to correct the error classification when necessary, the advantages of people are effectively utilized to make up for the machine errors, and higher sorting precision is easy to obtain.
Drawings
FIG. 1: a flow chart of the method of the invention;
FIG. 2: the branch arbitration block diagram of the method of the invention;
FIG. 3: the method is based on a human bifurcation intervention block diagram of a multi-algorithm;
FIG. 4: the method is based on a divergence intervention block diagram of algorithm integration;
FIG. 5: a data set for use in the method of the invention.
Detailed Description
In order to facilitate the understanding and implementation of the present invention for those of ordinary skill in the art, the present invention is further described in detail below with reference to the accompanying drawings and examples.
A bifurcation intervention pearl sorting method based on multi-algorithm integration comprises the following steps:
the method comprises the following steps: respectively recording pearl data sets as Class1-7 according to categories, labeling each picture, and recording the data sets according to the ratio of 6: 2: 2, dividing the ratio into a training set, a verification set and a test set;
step two: three models ResNet50, SE-ResNet50 and Vgg16 which are mainstream and advanced at present are selected for independent training, pearl sorting tasks can be completed, and the optimal models are respectively stored: setting the maximum training times to be 10 epochs when training the model, setting the single processing sample number Batch during training and testing to be 32, setting the learning rate to be 0.0001, using an Adam optimizer, setting the step-size to be 5, setting the gamma to be 0.1, gradually changing the learning rate to be 0.1 after 5 changes in the training process, and finally obtaining the accuracy of the three optimal models to be 91.84%, 92.25% and 91.29% respectively;
step three: the obtained three optimal models are used for an arbitration system, namely divergence arbitration is carried out on prediction results given by the constructed primary system and the constructed secondary system: if the divergence occurs, the system classification is considered to be possibly wrong, and an interventionalist is needed to make a final decision; if there is no divergence, the prediction result of the main algorithm is retained. Here, the three models are grouped in pairs, the remaining model serves as a main system, the collocated two models form a subsystem by an integrated learning method, and the final prediction output of the subsystem obtained by integration is defined as:
wherein h isi(x) Is the probability output result of the ith model, T is the number of models, wiIs a single model hiRight of (1)
Step four: in the experimental verification stage, the divergence accuracy index alpha and the additional cost index beta are used as evaluation indexes of system divergence to evaluate the overall performance of the system: the divergence accuracy index alpha refers to the ratio of the number of the corrected pictures after the classification error of the main algorithm to the total number of the pictures of the classification error of the model, and is used for evaluating the capability of the method for finding out the classification error; the additional cost index beta refers to the ratio of the number of pictures which are correctly classified by the main algorithm and are submitted to a person for judgment to the total number of pictures which are correctly classified by the optimal model, and is used for representing the invalid human cost consumed by the method;
step five: selecting a system combination which achieves higher sorting precision and requires less labor cost as a final pearl sorting system according to the evaluation indexes in the step four;
step six: outputting a final classification result: if the arbiter selects the classification result given by the main system as the pearl type, otherwise, the arbiter gives the prediction result again after the classification is judged by the person, and the result is considered as the final type.
The pearl data set classification step specifically comprises the following steps: first, classification according to the rough rule can be roughly divided into two categories: 1) one type is the low value flat or deep defect pearl; 2) the other is pearl with smaller flaw degree; secondly, according to a more refined classification standard, the pearl classification with poor first major classification can be divided into three minor classifications: A1) a pearl having a plurality of flat surfaces; A2) pearls of symmetrical shape; A3) other pearls; the second major category of better pearl classification can be divided into four subclasses: B1) pearls having a ratio of minor to major radii greater than 0.7; B2) a light colored pearl; B3) pearls with hidden speckles; B4) the rest of pearl.
The Convolutional Neural Network and Pearl data set referred to herein are disclosed in "Automatic Pearl Classification Machine base ON a Multistream conditional Neural Network" of IEEE TRANSACTIONS ON INDUSTRIAL ELECTRICAL 2018 by Qi Xuan, Binwei Fang, et al.
The invention finds out the error prediction possibly existing in machine classification through divergence based on a plurality of redundant algorithms, and corrects the prediction output with the intervention of least labor cost, thereby improving the classification precision of the whole pearl sorting.
The embodiments described in this specification are merely illustrative of implementations of the inventive concept and the scope of the present invention should not be considered limited to the specific forms set forth in the embodiments but rather by the equivalents thereof as may occur to those skilled in the art upon consideration of the present inventive concept.
Claims (2)
1. A bifurcation intervention pearl sorting method based on multi-algorithm integration comprises the following steps:
the method comprises the following steps: respectively recording pearl data sets as Class1-7 according to categories, labeling each picture, and recording the data sets according to the ratio of 6: 2: 2, dividing the ratio into a training set, a verification set and a test set;
step two: three models ResNet50, SE-ResNet50 and Vgg16 which are mainstream and advanced at present are selected for independent training, pearl sorting tasks can be completed, and the optimal models are respectively stored: setting the maximum training times to be 10 epochs when training the model, setting the single processing sample number Batch during training and testing to be 32, setting the learning rate to be 0.0001, using an Adam optimizer, setting the step-size to be 5, setting the gamma to be 0.1, gradually changing the learning rate to be 0.1 after 5 changes in the training process, and finally obtaining the accuracies of the three optimal models to be 91.84%, 92.25% and 91.29% respectively;
step three: the obtained three optimal models are used for an arbitration system, namely divergence arbitration is carried out on prediction results given by the constructed primary system and the constructed secondary system: if the divergence occurs, the system classification is considered to be possibly wrong, and an interventionalist is needed to make a final decision; if there is no divergence, the prediction result of the main algorithm is retained. Here, the three models are grouped in pairs, the remaining model is used as a main system, the two collocated models form a subsystem by an ensemble learning method, and the final prediction output of the subsystem obtained by integration is defined as:
wherein h isi(x) Is the probability output result of the ith model, T is the number of models, wiIs a single model hiIs usually given a weight of
Step four: in the experimental verification stage, the divergence accuracy index alpha and the additional cost index beta are used as evaluation indexes of system divergence to evaluate the overall performance of the system: the divergence accuracy index alpha refers to the ratio of the number of the corrected pictures after the classification error of the main algorithm to the total number of the pictures of the classification error of the model, and is used for evaluating the capability of the method for finding out the classification error; the additional cost index beta refers to the ratio of the number of pictures which are correctly classified by the main algorithm and are submitted to a person for judgment to the total number of pictures which are correctly classified by the optimal model, and is used for representing the invalid human cost consumed by the method;
step five: selecting a system combination which achieves higher sorting precision and requires less labor cost as a final pearl sorting system according to the evaluation indexes in the step four;
step six: outputting a final classification result: if the arbiter selects the classification result given by the main system as the pearl type, otherwise, the arbiter gives the prediction result again after the classification is judged by the person, and the result is considered as the final type.
2. The method for detecting and classifying pearls according to claim 1, wherein: step one, the pearl data set classification step specifically comprises: first, classification according to the rough rule can be roughly divided into two categories: 1) one type is the low value flat or deep defect pearl; 2) the other is pearl with smaller flaw degree; secondly, according to a more refined classification standard, the pearl classification with poor first major classification can be divided into three minor classifications: A1) a pearl having a plurality of flat surfaces; A2) pearls of symmetrical shape; A3) other pearls; the second major category of better pearl classification can be divided into four subclasses: B1) pearls having a ratio of minor to major radii greater than 0.7; B2) a light colored pearl; B3) pearls with hidden speckles; B4) the rest of pearl.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110386522.XA CN113159150B (en) | 2021-04-12 | Multi-algorithm integration-based branch intervention pearl sorting method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110386522.XA CN113159150B (en) | 2021-04-12 | Multi-algorithm integration-based branch intervention pearl sorting method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113159150A true CN113159150A (en) | 2021-07-23 |
CN113159150B CN113159150B (en) | 2024-07-19 |
Family
ID=
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106529568A (en) * | 2016-10-11 | 2017-03-22 | 浙江工业大学 | Pearl multi-classification method based on BP neural network |
CN106650721A (en) * | 2016-12-28 | 2017-05-10 | 吴晓军 | Industrial character identification method based on convolution neural network |
CN109684922A (en) * | 2018-11-20 | 2019-04-26 | 浙江大学山东工业技术研究院 | A kind of recognition methods based on the multi-model of convolutional neural networks to finished product dish |
CN110263774A (en) * | 2019-08-19 | 2019-09-20 | 珠海亿智电子科技有限公司 | A kind of method for detecting human face |
CN110689081A (en) * | 2019-09-30 | 2020-01-14 | 中国科学院大学 | Weak supervision target classification and positioning method based on bifurcation learning |
CN110932274A (en) * | 2019-12-18 | 2020-03-27 | 辽宁工业大学 | Power system measurement and load parameter analysis and identification method |
CN112381127A (en) * | 2020-11-03 | 2021-02-19 | 浙江工业大学 | Pearl sorting method based on human bifurcation intervention |
CN112559791A (en) * | 2020-11-30 | 2021-03-26 | 广东工业大学 | Cloth classification retrieval method based on deep learning |
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106529568A (en) * | 2016-10-11 | 2017-03-22 | 浙江工业大学 | Pearl multi-classification method based on BP neural network |
CN106650721A (en) * | 2016-12-28 | 2017-05-10 | 吴晓军 | Industrial character identification method based on convolution neural network |
CN109684922A (en) * | 2018-11-20 | 2019-04-26 | 浙江大学山东工业技术研究院 | A kind of recognition methods based on the multi-model of convolutional neural networks to finished product dish |
CN110263774A (en) * | 2019-08-19 | 2019-09-20 | 珠海亿智电子科技有限公司 | A kind of method for detecting human face |
CN110689081A (en) * | 2019-09-30 | 2020-01-14 | 中国科学院大学 | Weak supervision target classification and positioning method based on bifurcation learning |
CN110932274A (en) * | 2019-12-18 | 2020-03-27 | 辽宁工业大学 | Power system measurement and load parameter analysis and identification method |
CN112381127A (en) * | 2020-11-03 | 2021-02-19 | 浙江工业大学 | Pearl sorting method based on human bifurcation intervention |
CN112559791A (en) * | 2020-11-30 | 2021-03-26 | 广东工业大学 | Cloth classification retrieval method based on deep learning |
Non-Patent Citations (3)
Title |
---|
QI XUAN 等: "Automatic Pearl Classification Machine Based on a Multistream Convolutional Neural Network", IEEE TRANSACTIONS ON INDUSTRIAL ELECTEONICS, vol. 65, no. 8, 13 April 2018 (2018-04-13), pages 6538 - 6547, XP011682147, DOI: 10.1109/TIE.2017.2784394 * |
QI XUAN 等: "Multiview Generative Adversarial Network and Its Application in Pearl Classification", IEEE TRANSACTIONS ON INDUSTRISAL ELECTEONICS, vol. 66, no. 10, 1 October 2019 (2019-10-01), pages 8244 - 8252, XP011727390, DOI: 10.1109/TIE.2018.2885684 * |
秦海;张东波;王俊超;颜霜;: "流数据环境下基于分歧策略的高效能集成学习", 计算机工程与应用, no. 13, 12 May 2015 (2015-05-12), pages 89 - 94 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108256482B (en) | Face age estimation method for distributed learning based on convolutional neural network | |
CN110689089A (en) | Active incremental training method for deep learning of multi-class medical image classification | |
CN108805196B (en) | Automatic incremental learning method for image recognition | |
CN106096661B (en) | The zero sample image classification method based on relative priority random forest | |
CN110688888B (en) | Pedestrian attribute identification method and system based on deep learning | |
US20220222807A1 (en) | Ai-based new learning model generation system for vision inspection on product production line | |
CN110610482A (en) | Workpiece flaw detection method based on resnet | |
CN109214444B (en) | Game anti-addiction determination system and method based on twin neural network and GMM | |
CN113887908A (en) | Bridge risk assessment method considering subjective and objective cross fusion weight | |
Tian et al. | CASDD: Automatic surface defect detection using a complementary adversarial network | |
CN109598681A (en) | The reference-free quality evaluation method of image after a kind of symmetrical Tangka repairs | |
CN116468935A (en) | Multi-core convolutional network-based stepwise classification and identification method for traffic signs | |
CN114580517A (en) | Method and device for determining image recognition model | |
CN111144462A (en) | Unknown individual identification method and device for radar signals | |
Babnik et al. | Optimization-based improvement of face image quality assessment techniques | |
CN114372926A (en) | Traditional Chinese medicine tongue tenderness identification method based on image restoration and convolutional neural network | |
CN105373568B (en) | Problem answers Auto-learning Method and device | |
CN113159150A (en) | Branch intervention pearl sorting method based on multi-algorithm integration | |
CN113159150B (en) | Multi-algorithm integration-based branch intervention pearl sorting method | |
CN115759205A (en) | Negative sample sampling method based on multi-model cooperation contrast learning | |
CN116137012A (en) | Online teaching quality supervision and management system based on Internet education | |
Sun et al. | A classification and location of surface defects method in hot rolled steel strips based on YOLOV7 | |
CN110992333B (en) | COF Film defect AITraining database and application | |
CN107016131A (en) | Machine learning algorithm based on enhanced clustering and application of algorithm | |
CN114519092A (en) | Large-scale complex relation data set construction framework oriented to Chinese field |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant |