CN112541407A - Visual service recommendation method based on user service operation flow - Google Patents
Visual service recommendation method based on user service operation flow Download PDFInfo
- Publication number
- CN112541407A CN112541407A CN202011365755.3A CN202011365755A CN112541407A CN 112541407 A CN112541407 A CN 112541407A CN 202011365755 A CN202011365755 A CN 202011365755A CN 112541407 A CN112541407 A CN 112541407A
- Authority
- CN
- China
- Prior art keywords
- service
- user
- algorithm
- recommendation
- machine vision
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
- G06F16/535—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Library & Information Science (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a visual service recommendation method based on user service operation flow, which is a method for recommending machine vision algorithm service to a user by a visual service recommendation technology based on user service operation flow, wherein FCM clustering is performed on bottom-layer service flows of a machine vision algorithm by the aid of the user service operation flow, and user preference relations among the bottom-layer machine vision algorithm service are classified and screened; and then, the LFM algorithm is used for carrying out matrix modeling analysis on the operation flow preference of the user, and the dependence degree and preference degree of different users on the bottom layer machine vision algorithm service are calculated, so that the repeated recommendation of the traditional algorithm recommendation service left in a surface layer algorithm module is broken, and the algorithm service module recommendation can not be deeply carried out from the operation of the bottom layer algorithm. And the visual service used by the user bottom layer of the user is known by analyzing the visual algorithm used by the platform user, so that accurate visual service recommendation is carried out.
Description
Technical Field
The invention relates to the technical field of machine vision information processing and data service analysis, in particular to a machine vision service platform which realizes accurate recommendation and personalized service of machine vision science and technology service resources by researching and utilizing technologies such as machine learning, data analysis and recommendation algorithm under the service operation background.
Background
From 2012, the artificial intelligence technology developed by ImageNet competition explosion to the present, the application of the machine vision technology in industry and life is more and more extensive, and a large number of machine vision technologies are packaged into service and applied to a machine vision platform. In developed countries in the western world, a cloud service platform is deployed in large numbers in recent years to provide artificial intelligence service technology, and machine vision is a comprehensive technology including image processing, mechanical engineering technology, control, electric light source illumination, optical imaging, sensors, analog and digital video technology, and computer software and hardware technology (image enhancement and analysis algorithms, image cards, I/O cards, and the like). A typical machine vision application system comprises an image capture module, a light source system, an image digitization module, a digital image processing module, an intelligent judgment decision module and a mechanical control execution module. The most basic feature of machine vision systems is to increase the flexibility and automation of production. In some dangerous working environments which are not suitable for manual operation or occasions where manual vision is difficult to meet the requirements, machine vision is often used to replace the manual vision. Meanwhile, in the process of mass repetitive industrial production, the machine vision detection method can greatly improve the production efficiency and the automation degree.
Cloud computing is the development of parallel computing, distributed computing, and grid computing, or rather the commercial implementation of these computational science concepts. Cloud computing is the result of a mixed evolution and leap of concepts such as virtualization, utility computing, infrastructure as a service, platform as a service, and software as a service. For a machine vision service platform, algorithm recommendation has the characteristics of low efficiency and waste caused by repeated calling of service resources, and most recommendation algorithms do not consider the operation flow of a user and the service use preference of an invisible underlying algorithm.
Disclosure of Invention
The invention aims to overcome the defects that the existing machine vision service recommendation technology is left in a service algorithm recommendation surface layer and does not consider the characteristics of an operation process of user operation on a vision platform, and provides a service recommendation method based on a user service operation flow to provide machine vision service for a user. Based on LFM (tension Factor model) algorithm, the proposal for generating accurate recommendation is generated.
A personalized service recommendation method based on a big data platform is characterized in that a method for recommending machine vision algorithm services to a user by a vision service recommendation technology based on user service operation flows is used for generating FCM clustering on bottom layer service flows of a machine vision algorithm through the user service operation flows, and classifying and screening user preference relations among the bottom layer machine vision algorithm services; and then, the LFM algorithm is used for carrying out matrix modeling analysis on the operation flow preference of the user, and the dependence degree and preference degree of different users on the bottom layer machine vision algorithm service are calculated, so that the repeated recommendation of the traditional algorithm recommendation service left in a surface layer algorithm module is broken, and the algorithm service module recommendation can not be deeply carried out from the operation of the bottom layer algorithm. And the visual service used by the user bottom layer of the user is known by analyzing the visual algorithm used by the platform user, so that accurate visual service recommendation is carried out.
The method specifically comprises the following steps:
1) fine modularization of algorithm bottom layer service: decoupling bottom codes of a large machine vision service algorithm, repackaging the bottom codes into a small service algorithm module, and exposing the small service algorithm module to a user through a service interface; module processing is carried out on various bottom layer repeated service modules, and multithreading concurrent encapsulation is carried out on service modules which cannot be multithreaded;
2) analyzing the user operation flow: according to the sequence of algorithms used by users, the dynamic implementation condition of algorithm calling and the algorithm data flow direction, user operation flow state data are constructed, matrix modeling is carried out on the data flow state of the users, rows of the matrix are different users, and the rows are the operation flow sequence of the algorithm calling carried out by the users; after the modeling is successful, storing the model into a user database dictionary; when a user calls a service algorithm module, tracking and analyzing the calling sequence of the service algorithm module, analyzing the type of data transmitted by the user, and simultaneously constructing an operation flow matrix and storing the operation flow matrix in a user database dictionary;
3) FCM analysis step: extracting machine vision service module information and a user dictionary from a database to obtain a user operation flow and a user data flow, classifying the machine vision service module from functions and calling characteristics through FCM fuzzy clustering analysis, and carrying out fuzzy clustering on operation flow matrixes and user data flow matrixes of different users, so as to classify the users and store classification results into a user service matching database;
4) and a user service matching step: service data users and service use frequency in a User service matching database are counted, coarse matching is carried out on the bottom layer service class data and the User service module after FCM analysis, and a User-Item matrix is roughly constructed;
5) LFM recommendation optimization algorithm steps: the method comprises the steps of using a User-Item matrix constructed by User service matching to carry out parameter iteration, calculating included angles between Item vectors of users to obtain similar User groups, carrying out machine vision service supplementary recommendation among the same User groups, recording service recommendation data streams, recording services which are presented to the users but not used by the users, updating the User-Item matrix again by using the data, and carrying out online learning, thereby realizing accurate User service recommendation.
Furthermore, the machine vision algorithm service is ultra-fine modularized, the machine vision algorithm service is finely layered from the algorithm module layer to the bottom service module calling layer, the large machine vision algorithm is disassembled and re-packaged, and a hard interface is provided for the analysis of the user service operation flow.
Further, the statistical calculation of the user service operation flow uses the FCM fuzzy clustering algorithm to perform rough machining on the user operation flow, and the algorithm used by the user and the operation flow are analyzed, so that the hierarchical relation of the user operation flow and the correlation coefficient between the bottom service modules of the machine vision algorithm are analyzed.
Further, the LFM algorithm is used for processing the hierarchical relation of the user operation flow and the correlation coefficient of the algorithm module, the screened algorithm service module is recommended to the user, and the user can freely select the needed service algorithm module.
Drawings
FIG. 1 is a diagram of a visual service recommendation technique architecture based on user service operation flows.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
Examples
As shown in fig. 1, the visual service recommendation method based on user service operation flow includes the following specific steps:
1. fine modularization of algorithm bottom layer service: decoupling bottom codes of a large machine vision service algorithm, repackaging the bottom codes into a small service algorithm module, and exposing the small service algorithm module to a user through a service interface; and module processing is carried out on various bottom layer repeated service modules, and multithreading concurrent encapsulation is carried out on service modules which cannot be multithreaded, so that the independence and high concurrency of calling of the bottom layer service modules are ensured.
2. Analyzing the user operation flow: according to the sequence of the algorithms used by the users, the dynamic implementation condition of algorithm calling and the algorithm data flow direction, user operation flow state data are constructed, matrix modeling is carried out on the data flow states of the users, rows of the matrix are different users, and the rows are listed as the operation flow sequence of the algorithm calling carried out by the users. And storing the model into a user database dictionary after the model is successfully built. When a user calls the service algorithm module, tracking and analyzing the calling sequence of the service algorithm module, analyzing the type of data transmitted by the user, and simultaneously constructing an operation flow matrix and storing the operation flow matrix in a user database dictionary.
In the storing process, repeated operation of the user needs to be performed with duplicate removal processing, so that data redundancy of the user dictionary is reduced.
FCM analysis step: and extracting machine vision service module information and a user dictionary from a database to obtain a user operation flow and a user data flow, classifying the functions and calling characteristics of the machine vision service module through FCM fuzzy clustering analysis, and carrying out fuzzy clustering on operation flow matrixes and user data flow matrixes of different users, so that the users are classified, and classification results are stored in a user service matching database.
4. And a user service matching step: and counting service data users and service use frequency in the database, roughly matching the bottom service class data and the User service module after the FCM analysis, and roughly constructing a User-Item matrix.
Step 5, LFM recommendation optimization algorithm: the method comprises the steps of using a User-Item matrix constructed by User service matching to carry out parameter iteration, calculating included angles between Item vectors of users to obtain similar User groups, carrying out machine vision service supplementary recommendation among the same User groups, recording service recommendation data streams, recording services which are presented to the users but not used by the users, updating the User-Item matrix again by using the data, and carrying out online learning, thereby realizing accurate User service recommendation.
Based on the technical scheme, when the system is used:
decoupling a rear-end machine vision algorithm, carrying out refined module packaging on different algorithm services, putting service algorithm codes into a database for storage, butting a front-end user click time and an algorithm operation time recommendation system of the machine vision service, transmitting a user operation flow to a rear end, calling FCM (fuzzy C-means) analysis by the rear end, matching the user service with the database for LFM service recommendation, recommending the algorithm service to a user, and allowing the user to select the algorithm service at the front end to carry out machine vision algorithm operation.
Claims (2)
1. A personalized service recommendation method based on a big data platform is characterized in that a method for recommending machine vision algorithm services to a user by a vision service recommendation technology based on user service operation flows is used for generating FCM clustering on bottom layer service flows of a machine vision algorithm through the user service operation flows, and classifying and screening user preference relations among the bottom layer machine vision algorithm services; and then, the LFM algorithm is used for carrying out matrix modeling analysis on the operation flow preference of the user, and the dependence degree and preference degree of different users on the bottom layer machine vision algorithm service are calculated, so that the repeated recommendation of the traditional algorithm recommendation service left in a surface layer algorithm module is broken, and the algorithm service module recommendation can not be deeply carried out from the operation of the bottom layer algorithm. And the visual service used by the user bottom layer of the user is known by analyzing the visual algorithm used by the platform user, so that accurate visual service recommendation is carried out.
2. The method of claim 1, comprising the steps of:
1) fine modularization of algorithm bottom layer service: decoupling bottom codes of a large machine vision service algorithm, repackaging the bottom codes into a small service algorithm module, and exposing the small service algorithm module to a user through a service interface; module processing is carried out on various bottom layer repeated service modules, and multithreading concurrent encapsulation is carried out on service modules which cannot be multithreaded;
2) analyzing the user operation flow: according to the sequence of algorithms used by users, the dynamic implementation condition of algorithm calling and the algorithm data flow direction, user operation flow state data are constructed, matrix modeling is carried out on the data flow state of the users, rows of the matrix are different users, and the rows are the operation flow sequence of the algorithm calling carried out by the users; after the modeling is successful, storing the model into a user database dictionary; when a user calls a service algorithm module, tracking and analyzing the calling sequence of the service algorithm module, analyzing the type of data transmitted by the user, and simultaneously constructing an operation flow matrix and storing the operation flow matrix in a user database dictionary;
3) FCM analysis step: extracting machine vision service module information and a user dictionary from a database to obtain a user operation flow and a user data flow, classifying the machine vision service module from functions and calling characteristics through FCM fuzzy clustering analysis, and carrying out fuzzy clustering on operation flow matrixes and user data flow matrixes of different users, so as to classify the users and store classification results into a user service matching database;
4) and a user service matching step: service data users and service use frequency in a User service matching database are counted, coarse matching is carried out on the bottom layer service class data and the User service module after FCM analysis, and a User-Item matrix is roughly constructed;
5) LFM recommendation optimization algorithm steps: the method comprises the steps of using a User-Item matrix constructed by User service matching to carry out parameter iteration, calculating included angles between Item vectors of users to obtain similar User groups, carrying out machine vision service supplementary recommendation among the same User groups, recording service recommendation data streams, recording services which are presented to the users but not used by the users, updating the User-Item matrix again by using the data, and carrying out online learning, thereby realizing accurate User service recommendation.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2020108418529 | 2020-08-20 | ||
CN202010841852 | 2020-08-20 |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112541407A true CN112541407A (en) | 2021-03-23 |
CN112541407B CN112541407B (en) | 2022-05-13 |
Family
ID=75017065
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011365755.3A Active CN112541407B (en) | 2020-08-20 | 2020-11-29 | Visual service recommendation method based on user service operation flow |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112541407B (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2008077022A2 (en) * | 2006-12-18 | 2008-06-26 | The Johns Hopkins University | Himf and btk in pulmonary, cardiac, and inflammation disorders |
CN104166668A (en) * | 2014-06-09 | 2014-11-26 | 南京邮电大学 | News recommendation system and method based on FOLFM model |
CN105913296A (en) * | 2016-04-01 | 2016-08-31 | 北京理工大学 | Customized recommendation method based on graphs |
CN108198045A (en) * | 2018-01-30 | 2018-06-22 | 东华大学 | The design method of mixing commending system based on e-commerce website data mining |
CN108460489A (en) * | 2018-03-15 | 2018-08-28 | 重庆邮电大学 | A kind of user behavior analysis based on big data technology and service recommendation frame |
CN108763590A (en) * | 2018-06-20 | 2018-11-06 | 合肥工业大学 | A kind of data clustering methods based on double change Weighted Kernel FCM algorithms |
CN108763478A (en) * | 2018-05-29 | 2018-11-06 | 努比亚技术有限公司 | User's recessive character computational methods, server and computer readable storage medium |
CN111199422A (en) * | 2019-12-25 | 2020-05-26 | 浙江工商大学 | Improved LFM (Linear frequency modulation) collaborative filtering method fusing time attenuation factors |
-
2020
- 2020-11-29 CN CN202011365755.3A patent/CN112541407B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2008077022A2 (en) * | 2006-12-18 | 2008-06-26 | The Johns Hopkins University | Himf and btk in pulmonary, cardiac, and inflammation disorders |
CN104166668A (en) * | 2014-06-09 | 2014-11-26 | 南京邮电大学 | News recommendation system and method based on FOLFM model |
CN105913296A (en) * | 2016-04-01 | 2016-08-31 | 北京理工大学 | Customized recommendation method based on graphs |
CN108198045A (en) * | 2018-01-30 | 2018-06-22 | 东华大学 | The design method of mixing commending system based on e-commerce website data mining |
CN108460489A (en) * | 2018-03-15 | 2018-08-28 | 重庆邮电大学 | A kind of user behavior analysis based on big data technology and service recommendation frame |
CN108763478A (en) * | 2018-05-29 | 2018-11-06 | 努比亚技术有限公司 | User's recessive character computational methods, server and computer readable storage medium |
CN108763590A (en) * | 2018-06-20 | 2018-11-06 | 合肥工业大学 | A kind of data clustering methods based on double change Weighted Kernel FCM algorithms |
CN111199422A (en) * | 2019-12-25 | 2020-05-26 | 浙江工商大学 | Improved LFM (Linear frequency modulation) collaborative filtering method fusing time attenuation factors |
Also Published As
Publication number | Publication date |
---|---|
CN112541407B (en) | 2022-05-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Oh et al. | Hybrid genetic algorithms for feature selection | |
CN116152267B (en) | Point cloud instance segmentation method based on contrast language image pre-training technology | |
CN109271521A (en) | A kind of file classification method and device | |
Kang | Rotation-invariant wafer map pattern classification with convolutional neural networks | |
Chen et al. | Cf-vit: A general coarse-to-fine method for vision transformer | |
Bespalov et al. | Scale-space representation of 3d models and topological matching | |
CN110263151B (en) | Latent semantic learning method for multi-view multi-label data | |
CN110188816B (en) | Image fine granularity identification method based on multi-stream multi-scale cross bilinear features | |
CN107918761A (en) | A kind of single sample face recognition method based on multiple manifold kernel discriminant analysis | |
Kamran et al. | Efficient yet deep convolutional neural networks for semantic segmentation | |
CN115294563A (en) | 3D point cloud analysis method and device based on Transformer and capable of enhancing local semantic learning ability | |
Han et al. | 3CROSSNet: Cross-level cross-scale cross-attention network for point cloud representation | |
CN110991247B (en) | Electronic component identification method based on deep learning and NCA fusion | |
Depuru et al. | Convolutional neural network based human emotion recognition system: A deep learning approach | |
CN111881906A (en) | LOGO identification method based on attention mechanism image retrieval | |
Terziyan et al. | Causality-aware convolutional neural networks for advanced image classification and generation | |
Pandey et al. | Do-it-yourself recommender system: Reusing and recycling with blockchain and deep learning | |
Sarpietro et al. | Explainable deep learning system for advanced silicon and silicon carbide electrical wafer defect map assessment | |
CN112541407B (en) | Visual service recommendation method based on user service operation flow | |
CN113569955A (en) | Model training method, user portrait generation method, device and equipment | |
Schönhof et al. | Towards automated capability assessment leveraging deep learning | |
CN115293639A (en) | Battlefield situation studying and judging method based on hidden Markov model | |
CN109472319B (en) | Three-dimensional model classification method and retrieval method | |
CN113837271A (en) | Classification improvement algorithm based on feature selection | |
Chen et al. | Review of Lightweight Deep Convolutional Neural Networks |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |