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

Unsupervised Hash Fast Image Retrieval System and Method Based on Convolutional Neural Network

A convolutional neural network and neural network technology, applied in the field of unsupervised hashing fast image retrieval system, can solve the problems of unbalanced training data and uncoordinated training, so as to ensure mutual adaptability, improve accuracy, and ensure balance. Effect

Active Publication Date: 2020-01-21
上海媒智科技有限公司
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a convolutional neural network-based unsupervised hash fast image retrieval system and method to solve the problem of unbalanced training data and uncoordinated training in the existing methods. question

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
  • Unsupervised Hash Fast Image Retrieval System and Method Based on Convolutional Neural Network
  • Unsupervised Hash Fast Image Retrieval System and Method Based on Convolutional Neural Network
  • Unsupervised Hash Fast Image Retrieval System and Method Based on Convolutional Neural Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0061] This embodiment provides an unsupervised hash fast image retrieval system and method based on a convolutional neural network. The system and method utilize data enhancement technology to implement an unsupervised hashing algorithm based on a convolutional neural network. Through this model, The input image is mapped to a shorter binary hash code, and similar images can be screened out by comparing the Hamming distance in the image retrieval process. The present invention is a network structure capable of using unlabeled training data to train a more discriminative ability in the field of fast image retrieval. By using the added ternary loss function, minimum quantization error loss function and maximum entropy loss function, Significantly improved the accuracy of fast image retrieval.

[0062] The unsupervised hash fast image retrieval system based on the convolutional neural network provided in this embodiment includes the following modules:

[0063] The unsupervised ...

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 proposes an unsupervised hash fast image retrieval system and method based on a convolutional neural network. The system and method utilize the existing hash algorithm structure and propose an efficient unsupervised hash model based on data enhancement technology for the field of fast image retrieval. Through the data enhancement method, triplet training samples are constructed for unlabeled data, and the network is driven to make full use of the information of each picture through triplet loss function, minimum quantization error loss function and maximum entropy loss function, and learn a series of more expressive Capability parameters to improve the accuracy of fast image retrieval. The present invention is a hash fast image retrieval method capable of using unlabeled data to learn a network, uses data enhancement to construct a triplet training sample training network with stronger expressive ability, and significantly improves the accuracy of fast image retrieval.

Description

technical field [0001] The invention relates to the fields of computer vision and image processing, in particular to an unsupervised hash fast image retrieval system and method based on a convolutional neural network. Background technique [0002] With the explosive growth of existing multimedia content, how to speed up image retrieval has received extensive attention. As an algorithm that can convert a high-dimensional feature vector into a compact and expressive binary code through multiple mapping equations, hashing has achieved considerable success in the field of fast image retrieval. In recent years, with the rapid development of deep convolutional neural networks, many hashing algorithms based on convolutional neural networks have been proposed and have shown great prospects. In particular, unsupervised hashing algorithms based on unlabeled data have received extensive attention due to the lack of existing labeled data and the labor and resources required to label im...

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): G06F16/583G06N3/04G06N3/08
CPCG06F16/5838G06N3/088G06N3/045
Inventor 王延峰张娅黄杉杉熊意超
Owner 上海媒智科技有限公司
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