Blind domain image sample classification method based on over-limit hidden feature model

A technology of image samples and classification methods, which is applied in biological neural network models, character and pattern recognition, instruments, etc., and can solve problems such as performance discounts of adaptive models

Active Publication Date: 2019-06-25
CHONGQING UNIV OF POSTS & TELECOMM
View PDF11 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

When the amount of data in the target domain is seriously insufficient, the

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
  • Blind domain image sample classification method based on over-limit hidden feature model
  • Blind domain image sample classification method based on over-limit hidden feature model
  • Blind domain image sample classification method based on over-limit hidden feature model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, and are not intended to limit the invention.

[0033] This embodiment provides a blind domain image sample classification method based on an over-limit latent feature model, please refer to figure 1 shown, including the following steps:

[0034] S1: In the training stage of the source domain ultra-limited hidden feature model, obtain the source domain image data matrix for model training, the corresponding label matrix, and the hidden layer output matrix output by the hidden nodes of the ultra-limited learning machine according to the image data matrix .

[0035] Specifically, the image data set used for source domain mo...

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 a blind domain image sample classification method based on an over-limit hidden feature model, and the method employs a cross-domain data generation and augmentation strategy,and efficiently reduces the deviation between different data domains. Specifically, the over-limit hidden feature model incorporates an ELM and ELM-AE (an automatic encoder based on ELM) into a unified optimization model, and advantages of the ELM and ELM-AE are inherited, and good image classification and image reconstruction effects can be obtained under the condition that original data information is protected, the image classification capability and the image reconstruction capability are achieved, the potential relation between the original image data and high-level semantics can be better disclosed, information loss is reduced, and the image classification precision is improved. The blind domain adaptation problem is solved by using the over-limit hidden feature model, cross-domain knowledge is fused, domain deviation between different domain data is effectively reduced, and the blind domain data classification accuracy is improved.

Description

technical field [0001] The present invention relates to the technical field of image classification and intelligent optimization, and more specifically, relates to a blind field image sample classification method based on an over-limit hidden feature model. Background technique [0002] In the field of computer vision, people often collect data from multiple fields, and there are differences in distribution. Traditional machine learning methods assume that the training set and the test set obey the same distribution. This assumption is often difficult to be satisfied in reality. Cross-domain image recognition is currently a research hotspot in the field of computer vision. Domain adaptation is to solve the problem of cross-domain image recognition. It is One of the research contents of transfer learning focuses on solving the problem that the feature space is consistent, the category space is consistent, and only the feature distribution is inconsistent. [0003] The basic ...

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
IPC IPC(8): G06K9/62G06N3/04
Inventor 郭坦张磊谭晓衡杨柳鲁银芝梁志芳胡昊
Owner CHONGQING UNIV OF POSTS & TELECOMM
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap