Unlock instant, AI-driven research and patent intelligence for your innovation.

Image block and label matching method and system based on tripartite graph model

A matching method and image block technology, applied in the field of image block and label matching, can solve problems such as narrowing the semantic gap

Pending Publication Date: 2022-06-07
SHANDONG UNIV OF FINANCE & ECONOMICS
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Machines face two major challenges in image annotation tasks: the first is to narrow the semantic gap between low-level visual features and high-level semantic labels; the second is to learn the correspondence between labels and image regions in the training data, however, traditional Method assigns multiple labels to entire images without analyzing correspondences in fine-grained data

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
  • Image block and label matching method and system based on tripartite graph model
  • Image block and label matching method and system based on tripartite graph model
  • Image block and label matching method and system based on tripartite graph model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0043] This embodiment provides a method for matching image blocks and labels based on a tripartite graph model, constructs a tripartite graph model, and fully discovers the internal relationship between image blocks and labels through information diffusion, specifically including the following steps:

[0044] Step 1. Obtain the image set and its corresponding label set, divide each image in the image set into several blocks, cluster all the image blocks, and use the centroid of each cluster as a visual word to obtain the visual word set; based on the image Set and visual word set, obtain visual word-image bipartite graph; Based on image set and label set, obtain image-label bipartite graph; Described visual word-image bipartite graph and described image-label bipartite graph are combined into tripartite graph Model; Calculate the correlation matrix between visual words and tags based on the tripartite graph model.

[0045] Specifically, each image in the training data set is ...

Embodiment 2

[0068] The present embodiment provides an image block and tag matching system based on a tripartite graph model, which specifically includes the following modules:

[0069] A data acquisition module configured to: acquire an image set and its corresponding label set;

[0070] A visual word set building module configured to: divide each image in the image set into several image blocks, cluster all the image blocks, and use the centroid of each cluster as a visual word to obtain a visual word set;

[0071] Tripartite graph model construction module, it is configured to: Based on image set and visual word set, obtain visual word-image bipartite graph; Based on image set and label set, obtain image-label bipartite graph; Described visual word-image bipartite The graph and the image-label bipartite graph are combined into a tripartite graph model;

[0072] A correlation matrix building block configured to: calculate a correlation matrix between visual words and labels based on a t...

Embodiment 3

[0076] This embodiment provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, a method for matching image blocks and labels based on a tripartite graph model as described in the first embodiment above is realized in the steps.

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 provides an image block and label matching method and system based on a tripartite graph model. The method comprises the following steps: acquiring an image set and a corresponding label set; dividing each image in the image set into a plurality of image blocks, clustering all the image blocks, and taking the centroid of each cluster as a visual word to obtain a visual word set; obtaining a visual word-image bipartite graph based on the image set and the visual word set; obtaining an image-label bipartite graph based on the image set and the label set; combining the visual word-image bipartite graph and the image-label bipartite graph into a tripartite graph model; calculating a correlation matrix of visual words and labels based on a tripartite graph model; and for each image, constructing a sorting matrix based on the correlation matrix of the visual words and the labels, and matching the labels of all the image blocks in each image based on the sorting matrix. And the internal relation between the image block and the label is fully found.

Description

technical field [0001] The invention belongs to the technical field of image block and label matching, and in particular relates to an image block and label matching method and system based on a tripartite graph model. Background technique [0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art. [0003] In recent years, many users share images on social networking sites every day. Therefore, there is a growing need for an effective understanding of the semantics of images, which is of great benefit in improving the performance of image retrieval, recommendation, and management. Due to its critical importance to image semantic understanding, image annotation has attracted increasing research interest in the field of computer vision. [0004] Image annotation can be done by machines or humans, however, performing such tasks manually is time-consuming, costly, and ambiguous. F...

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): G06V10/25G06V10/75G06K9/62G06N20/00
CPCG06N20/00G06F18/22
Inventor 刘峥高珊珊迟静袁韶璟苏宜俊裴新蕾
Owner SHANDONG UNIV OF FINANCE & ECONOMICS