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

A Fine-Grained Image Retrieval Method for Simultaneous Localization and Hashing

An image retrieval, fine-grained technology, applied in still image data retrieval, neural learning methods, still image data query, etc., can solve the problems of not very good results, little research in the field of image retrieval, etc., to improve the accuracy and retrieval accuracy. boosted effect

Active Publication Date: 2022-04-22
SUN YAT SEN UNIV
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The related tasks of fine-grained images have attracted attention in the past two years. However, most of the current work focuses on image classification tasks, and there are few studies in the field of image retrieval.
Although some fine-grained image retrieval work that can be found currently try to fuse features of different scales or find key areas, they only consider the output of different layers of the fusion network or learn a weight to multiply with the original image. Direct and simple, the effect is not very good

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 Fine-Grained Image Retrieval Method for Simultaneous Localization and Hashing
  • A Fine-Grained Image Retrieval Method for Simultaneous Localization and Hashing
  • A Fine-Grained Image Retrieval Method for Simultaneous Localization and Hashing

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0052] The accompanying drawings are for illustrative purposes only and cannot be construed as limiting the patent;

[0053] In order to better illustrate this embodiment, some parts in the drawings will be omitted, enlarged or reduced, and do not represent the size of the actual product;

[0054] For those skilled in the art, it is understandable that some well-known structures and descriptions thereof may be omitted in the drawings.

[0055] The technical solutions of the present invention will be further described below in conjunction with the drawings and embodiments.

[0056] Such as Figure 1-2 As shown, a fine-grained image retrieval method that realizes simultaneous localization and hashing includes the following steps:

[0057] S1: For an input fine-grained picture, first extract the image features through the shared feature extractor ResNet-18, and the 512x7x7 feature output by the last convolutional layer of ResNet-18 is used as the input of the positioning module...

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 provides a fine-grained image retrieval method that realizes simultaneous positioning and hashing. Through simultaneous hashing and positioning, the method can better learn the discriminative region of the input image and integrate input information of different scales, thereby improving the efficiency of retrieval. In addition, the hash and localization modules in the method can be jointly trained in a mutually promoting manner, and can avoid manual extraction of image features in traditional methods, realize end-to-end training, and greatly improve the accuracy of retrieval.

Description

technical field [0001] The invention relates to the field of image processing algorithms, and more specifically, relates to a fine-grained image retrieval method that realizes simultaneous positioning and hashing. Background technique [0002] Browse pictures. This makes how to retrieve images efficiently has become a growing concern. The goal of content-based image retrieval (CRIR) is to efficiently and accurately retrieve the most relevant visual content to a query (image or text) from a large-scale database. [0003] Since the 1990s, image retrieval has attracted widespread attention from academia and industry, and among various retrieval methods, hashing methods have become an important technology because of their high efficiency. From the depth of the model, image hashing techniques can be roughly divided into traditional methods and deep network-based methods. Traditional methods generally need to manually extract the features of the image (such as SIFT features), a...

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): G06F16/53G06F16/583G06N3/04G06N3/08
CPCG06F16/53G06F16/583G06N3/08G06N3/045
Inventor 曾海恩赖韩江印鉴
Owner SUN YAT SEN UNIV
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