Check patentability & draft patents in minutes with Patsnap Eureka AI!

Scene investigation image retrieval method based on low-level image features and CNN features

An image feature and image retrieval technology, applied in the fields of instruments, character and pattern recognition, computer components, etc., can solve the problem of time-consuming training of CNN models, and achieve the effect of improving retrieval accuracy

Active Publication Date: 2018-12-11
厦门中盾安图威科技有限公司
View PDF5 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Although CNN-based image features can effectively express the content of surveyed images, it takes a long time to train the CNN model, and there is no standard surveyed image library to fully train the CNN model

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
  • Scene investigation image retrieval method based on low-level image features and CNN features
  • Scene investigation image retrieval method based on low-level image features and CNN features
  • Scene investigation image retrieval method based on low-level image features and CNN features

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0045] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

[0046] In order to solve the problem of low retrieval performance accuracy of low-level image features in the prior art, the present invention provides an image retrieval method for integrating low-level image features and CNN (Convolutional Neural Network, Convolutional Neural Network) features, such as figure 1 , figure 2 As shown, the steps are as follows:

[0047] 1) Extract CNN intermediate layer features, extract low-level image features;

[0048] 2) Fusion of low-level image features and CNN middle-level features to obtain fusion features;

[0049] 3) Calculate the similarity of the fusion features by using the block distance to obtain the retrieval results. Among them, the calculation results of the similarity are sorted, so that the search results can be viewed more quickly.

[0050] In step 1), the extraction method of CNN inter...

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 scene investigation image retrieval method based on low-level image features and CNN features, the method mainly utilizing image features extracted from a CNN modeland enabling fusion of the image features based on a CNN and traditional image features to solve the problem of describing the effective features of database images with complex contents and eliminating the semantic gap. In order to further improve the retrieval efficiency, the similarity of images is calculated by using the middle-level feature of a convolutional neural network and the low-levelfeature vector of traditional images, and the final image similarity sim is calculated. The similarity between images is measured by the distance between blocks of image feature vectors or feature matrices. Compared with a conventional low-level feature retrieval method, the method improves the retrieval accuracy to a considerable extent through experimental verification.

Description

technical field [0001] The invention relates to digital image processing technology, and more specifically, relates to an on-site survey image retrieval method that integrates low-level image features and CNN features. Background technique [0002] Deep learning technology is outstanding in image retrieval. Using deep convolutional neural network (Convolutional Neural Network, CNN) can adaptively learn semantic features from big data instead of artificially designed features, which is the biggest difference between it and traditional pattern recognition methods. Studies have shown that using the image features extracted by the CNN network as the input of the fully connected layer (FCN) can effectively improve the accuracy of image classification and recognition. However, the selection of the upper semantic layer of the convolutional neural network (CNN) is not conducive to target retrieval, because the upper semantic layer loses the spatial information of the target, and se...

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): G06K9/62
CPCG06F18/253
Inventor 刘颖胡丹王富平
Owner 厦门中盾安图威科技有限公司
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com