Unlock instant, AI-driven research and patent intelligence for your innovation.

SVM image recognition system and method based on cloud platform

An image recognition and cloud platform technology, applied in the field of image recognition, can solve the problem of low retrieval efficiency and achieve the effect of reducing classification time

Inactive Publication Date: 2019-08-27
HARBIN UNIV OF SCI & TECH
View PDF3 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to solve the problem of low retrieval efficiency of the existing SVM image recognition method, and propose a SVM image recognition system and method based on the cloud platform

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • SVM image recognition system and method based on cloud platform

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0019] This embodiment is based on an SVM image recognition system under a cloud platform. The cloud platform is mainly composed of three parts: a file system, a database, and distributed parallel computing; the most important calculation and processing parts in the cloud platform are mainly through distributed file management The realization of two key technologies of system and parallel processing;

[0020] As a storage and computing processing platform, the core part of cloud platform processing is still a distributed file system and parallel processing. The superior hardware system also enables the platform to reflect the characteristics of scalability, low cost, high fault tolerance, high efficiency and stability; the cloud platform has a complete structure, and storage and computing can be directly expanded without changes. The scalability is the cloud platform Key attributes.

specific Embodiment approach 2

[0022] In this embodiment, a SVM image recognition method based on a cloud platform, the amount of training sample data of the SVM method gradually increases, and the time of training samples also shows an exponential increase trend, which is still very difficult to perform in a stand-alone mode. This is also a problem caused by the increase in the training sample size. In order to solve this problem and accelerate the training speed of the SVM algorithm, the present invention studies the parallel computing SVM method based on the cloud platform, so that the computing time is further shortened. The main idea of ​​the SVM algorithm is to find the classification corresponding to the decision function in the training data set for analysis, and find the support vector of the data set; all support vectors have the characteristics of sparseness, and they occupy a small proportion in the data vector set. This feature realizes the parallel SVM algorithm for the data; in the operation p...

specific Embodiment approach 3

[0024] The difference from the second embodiment is that the SVM image recognition method based on the cloud platform in this embodiment is mainly implemented through the following steps:

[0025] Step 1: Upload data information to the cloud platform. Upload data information and submit jobs to the cloud platform, mainly obtain data sources from HDFS, divide the data according to the data cluster configuration, and classify and process the image samples of the job, and enter the nodes required in the process information.

[0026] Step two, realize the operation process of image sample reading. Read the image samples stored in HDFS into the system, and convert the parameter types of the data samples in the block. After the conversion, genetic algorithm is used to optimize the combined parameters of the conversion. After all the preparatory work, the svm_train function is called in, and the sample training process is performed to obtain the support vector of the data, namely Form ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an SVM image recognition system and method based on a cloud platform, belonging to the field of image processing. An existing SVM image recognition system and an existing recognition method have the problem of low efficiency. The invention discloses an SVM image recognition system based on a cloud platform. The method comprises the following steps of finding out the classification corresponding to the decision function in the training data set for analysis, and finding out the support vector of the data set. All support vectors have the characteristic of sparsity and occupy a very small specific gravity in a data vector set, and the parallel SVM algorithm of the data is realized by utilizing the characteristic. In the operation process, firstly, training data is segmented, and blocking processing is carried out; and then SVM algorithm solution is carried out on each segmented data block to achieve the purpose of shortening the solution time.

Description

Technical field [0001] The invention relates to an image recognition method, in particular to an SVM image recognition system and method based on a cloud platform. Background technique [0002] At this stage, personal computers and mobile Internet are widely used, and various pictures, sounds, videos and other digital information are widely interacted on the Internet, and the amount of data is already immeasurable. As the most vivid and direct picture and image information, it shines because of the use of the Internet, and it has increasingly become an important part of people's daily communication and learning. Massive image data is widely spread on the Internet, and hundreds of millions of pictures are uploaded or downloaded on the Internet every day. In order to facilitate people to retrieve the images they need from a large number of images, experts have conducted extensive research and learning on image retrieval methods. Among them, the support vector machine (SVM) model ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06F16/583
CPCG06F16/583G06F18/2411
Inventor 房国志李玉龙
Owner HARBIN UNIV OF SCI & TECH