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

A Co-occurrence Image Pattern Mining Method

An image mode, a technology in images, applied in character and pattern recognition, instruments, calculations, etc., can solve problems such as high computational complexity and inability to pick out

Active Publication Date: 2021-09-28
SUZHOU UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, there are serious problems in these two methods: Even if the optimization method [11] has been adopted, the search and matching process between feature points in the pairwise matching algorithm will lead to high computational complexity
For context-aware clustering algorithms, although searching and matching can be avoided when looking for co-occurrence patterns, the problem is that the most meaningful visual patterns cannot be picked out from all the patterns in the clustering results

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 Co-occurrence Image Pattern Mining Method
  • A Co-occurrence Image Pattern Mining Method
  • A Co-occurrence Image Pattern Mining Method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0032] Embodiment one: see figure 1 As shown, a co-occurrence image pattern mining method includes the following steps:

[0033] 1) Use the SIFT (Scale Invariant Feature Transform) algorithm to extract the visual primitives {v i};

[0034] 2) Cluster these visual primitives using a context-aware clustering algorithm: Context-aware clustering aims to classify all visual primitives into higher-level candidate patterns to discover meaningful co-occurrence patterns. Co-occurring visual patterns often have similar spatial structure as well as similar feature descriptors. Therefore, in order to cluster the visual primitives in the feature domain, context-aware clustering is adopted as the first step of the algorithm. In context-aware clustering, visual primitives are classified into M distinct visual vocabularies by using K-means clustering of raw features. Then, within a predetermined spatial neighborhood of each primitive, an M-dimensional aggregate (visual phrase) vector can ...

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 co-occurrence image pattern mining method, comprising the following steps: (1) extracting the visual primitives in the image by using the SIFT algorithm; (2) clustering the visual primitives into context similar groups by using the context-aware clustering ; (3) Use spatial clustering to divide contextually similar groups into object groups; (4) Merge matching patterns to delineate object groups; (5) SVD-SIFT detection for each object group; (6) Double-layer filtering rules Screen out meaningful co-occurrence patterns; (7) Confinement frame refinement. The invention can quickly and accurately discover the symbiotic visual pattern in the picture, so as to facilitate subsequent visual tasks.

Description

technical field [0001] The invention relates to a method for mining co-occurrence image patterns, belonging to the technical field of image retrieval. Background technique [0002] Meaningful co-occurring visual patterns are defined as patterns that occur multiple times in images with similar spatial structure. Meaningful patterns have high visual salience compared with the background, and meaningful visual patterns are important features of images, so they usually make images more vivid. Discovering meaningful recurrent visual patterns has many applications in computer vision, such as image recognition and segmentation, image coding, compression and summarization, image classification and annotation, and object retrieval. There are basically two approaches to mining and discovering thematic patterns in images: (1) pairwise matching, which compares the number between feature points; (2) context-aware clustering, which takes visual primitives into account when clustering sp...

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): G06K9/46G06K9/62
CPCG06V10/462G06F18/23213
Inventor 杨剑宇黄瑶邓宇阳朱晨
Owner SUZHOU UNIV