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

Reranking Method for Long Query Image Retrieval Combining Semantic and Visual Information

A technology for querying images and visual information, applied in the field of information retrieval, can solve problems such as low correlation with query words, unsatisfactory results, unreliable initial retrieval results, etc.

Active Publication Date: 2018-05-01
HEFEI UNIV OF TECH
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, many studies have pointed out that reranking using only image visual information cannot achieve satisfactory results.
At the same time, when using long queries for retrieval, the initial retrieval results are usually unreliable, that is, the images ranked in front of the initial retrieval results have low correlation with the query words

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
  • Reranking Method for Long Query Image Retrieval Combining Semantic and Visual Information
  • Reranking Method for Long Query Image Retrieval Combining Semantic and Visual Information
  • Reranking Method for Long Query Image Retrieval Combining Semantic and Visual Information

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0067] In this embodiment, a long query image retrieval and reordering method combining semantic and visual information is to reorder the search results returned by the image search engine, and proceed as follows:

[0068] Step 1. On the search engine, enter a long query statement Q for image retrieval. A long query statement is a natural language query statement composed of several closely related concepts. For example, the long query "people dancing on the wedding" contains three The concepts are "people", "dancing" and "wedding", respectively, and these three concepts are closely related. From the returned several long query images, select the long query images sorted as the top N in the long query images, and form the initial return list X={x by the first N long query images 1 ,x 2 ,...,x u ,...,x N},x u Indicates the uth long query image in the initial return list, u represents the long query image x u The position in the initial return list is the uth, u=1,...,N;

...

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 long inquiring image searching reordering algorithm based on semantic and visual information. The long inquiring image searching reordering algorithm is characterized by comprising the following steps: 1, inputting a long inquiring statement to acquire an initial returning list; 2, building a visual dictionary; 3, segmenting the long inquiring statement and extracting a visual concept; 4, acquiring initial returning lists respectively based on the visual concept; 5, extracting textual features and visual features; 6, establishing a probabilistic model; 7, estimating semantic correlations; 8, estimating visual correlations; 9, combing the semantic correlations and the visual correlations; 10, reordering to acquire a reordering result. According to the invention, the long inquiring image searching reordering algorithm can make full use of image feature information, so that the accuracy of image searching and reordering can be effectively improved.

Description

technical field [0001] The invention belongs to the technical field of information retrieval, in particular to a long query image retrieval and reordering method combining semantic and visual information. Background technique [0002] The 21st century is the information age. With the development of Internet technology and network sharing services, the image data on the network is increasing exponentially. Image retrieval has become an indispensable activity in people's daily life. As the search behavior of network users becomes more and more precise, query words become more and more complex, and complex long queries can express more specific and precise information than simple queries. However, existing web search engines usually have wrong sorting of the retrieval results returned by long queries. The reasons are mainly because: First, long queries are composed of multiple concepts, which further widens the semantic gap between textual query words and visual content. Seco...

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): G06F17/30
CPCG06F16/36G06F16/58G06F16/583
Inventor 洪日昌高鹏飞汪萌刘奕群孙茂松刘学亮郝世杰
Owner HEFEI UNIV OF TECH
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