Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A Method of Object Recognition Based on Autoencoder

An object recognition and self-encoding technology, applied in the field of object recognition based on self-encoding, can solve problems such as unfavorable classifier classification applications, and achieve the effect of improving the classification effect and distinguishing features.

Active Publication Date: 2018-11-16
CHINA UNIV OF PETROLEUM (EAST CHINA)
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the feature extraction process of existing autoencoders is an unsupervised process, that is, there is no constraint between similar and heterogeneous sample points in the hidden layer mapping space, which is not conducive to the classification application of classifiers

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
  • A Method of Object Recognition Based on Autoencoder
  • A Method of Object Recognition Based on Autoencoder
  • A Method of Object Recognition Based on Autoencoder

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0037] The present invention proposes an object recognition method based on self-encoding. Firstly, the data of the existing labeled image database is used to train the self-encoder and the softmax classifier according to the steps, and the optimal classification function parameters are obtained as fixed parameters for identification; Then, during classification and recognition, the image data to be recognized is input into the trained self-encoder and classifier for classification and recognition. Considering that the existing autoencoder feature extraction process is an unsupervised process, that is, there is no constraint between similar and heterogeneous sample points in the hidden layer mapping space, so the present invention designs a supervised feature extraction process with Large-Margin regularization , so that the sample points in the hidden layer mapping space are close to each other and far away from each other, which can better facilitate the classification applica...

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 present invention relates to a kind of object recognition method based on self-encoder, at first it is training process, namely self-encoder and classifier are trained, in the training process to self-encoder, added Large-Margin regularization term; Then for recognition In the process, the image data of the object to be recognized is converted into the corresponding format, and input into the trained autoencoder and classifier for classification and recognition. The present invention adds the supervised regularization of Large-Margin in the process of training the autoencoder, so that the similar sample points in the mapping space are gathered together, and the different types are far away from each other, so that the feature distinction between different categories is more obvious. Therefore, after the feature data is input into the classifier, its classification recognition effect is improved.

Description

technical field [0001] The invention relates to the technical field of object recognition, in particular to an object recognition method based on self-encoding. Background technique [0002] Object recognition is one of the basic functions of machine intelligence, and it is the core problem and key technology in any practical application system that takes images or videos as input. Object recognition technology has a wide range of needs and applications in both military and civilian applications. [0003] In the prior art, the deep neural network has been widely used in the field of object recognition, and the autoencoder as its basic architecture is also being continuously improved and perfected. However, the feature extraction process of existing autoencoders is an unsupervised process, that is, there is no constraint between similar and heterogeneous sample points in the hidden layer mapping space, which is not conducive to the classification application of classifiers. ...

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 Patents(China)
IPC IPC(8): G06K9/62
CPCG06F18/214G06F18/24
Inventor 刘伟锋马腾洲
Owner CHINA UNIV OF PETROLEUM (EAST CHINA)
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com