Feature clustering method based on self-adaptive chart filter

A clustering method and filter technology, applied in the field of signal processing, can solve the problems of reducing the optimization matching speed, the classification problem is complex, and the image signal classification effect is not very good, so as to reduce the impact of noise, high recognition rate and robustness. , the effect of known noise effects

Inactive Publication Date: 2018-11-23
SOUTHEAST UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

And because the number of filter taps is limited at the beginning, it also reduces the optimization matching speed
Classification under graph signals is a complex problem
Under the GF (Graph Filter) algorithm, the classification effect of this kind of graph signal with noise 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
  • Feature clustering method based on self-adaptive chart filter
  • Feature clustering method based on self-adaptive chart filter
  • Feature clustering method based on self-adaptive chart filter

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0058] The technical solutions and beneficial effects of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0059] Such as figure 1 As shown, the present invention provides a feature clustering method based on an adaptive graph filter, comprising the following steps:

[0060] 1) The feature data is standardized, and the features of the same dimension require unitization;

[0061] 2) Feature data label records provide a certain amount of prior knowledge;

[0062] The specific process is: randomly extract a certain amount of samples from the feature data, and assign a certain label value (+1 for positive examples, -1 for negative examples), and assign 0 to the selected feature data for the remaining bits;

[0063] 3) Build a graph signal based on the data source, map the data to the data structure graph, and obtain the initial graph signal s (known) ;

[0064] 4) Full variational smoothing filtering to obtain the image sign...

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 feature clustering method based on a self-adaptive chart filter. The method comprises the following steps: step one, feature data normalization, unitizing the features at thesame dimension; step two, feature data label recording, providing a certain amount of prior knowledge; step three, constructing a chart signal based on a data source, mapping the data to a data structure chart, and acquiring an initial chart signal s(known); step four, total-variation smooth filtering, acquiring the noise-reduced chart signal s (true); step five, chart filter self-matching, acquiring a tap array h; and step six, global filtering, acquiring a classification result s(pred). Through the method disclosed by the invention, the utilization on the prior knowledge can be enhanced, and a new chart signal weight side is constructed.

Description

technical field [0001] The invention belongs to the technical field of signal processing, relates to a graph signal processing theory, in particular to a feature clustering method based on an adaptive graph filter. Background technique [0002] Graph Signal Processing (Graph Signal Processing, GSP) is an extremely effective way to study unstructured data, and has a wide range of applications in data classification, compressed sensing, and linear prediction. Graph signal theory has established a relatively complete set of graph signal theory, including graph signal definition, graph filter, graph frequency definition, etc. In graph theory, it was first proposed by Sandryhaila et al. to convert the data classification problem under traditional machine learning into the filter matching problem under digital signal processing (DSP) by means of filters. When solving the classification problem under the GF (Graph Filter) algorithm, it is necessary to establish an index matrix, an...

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
IPC IPC(8): G06K9/62
CPCG06F18/2321G06F18/24147
Inventor 舒华忠陈晓鹏孔佑勇伍家松杨淳沨
Owner SOUTHEAST UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap