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

Block-based robust tensor principal component analysis method

A principle component analysis and robust technology, applied in the field of image processing, can solve problems such as large noise points or abnormal points, rough processing of sparse components, etc., and achieve the effect of easy parameter setting, accurate and clear details

Active Publication Date: 2019-03-08
UNIV OF ELECTRONIC SCI & TECH OF CHINA
View PDF4 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0017] The IBTSVT method is too rough for sparse components, and the result may have large noise points or outliers

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
  • Block-based robust tensor principal component analysis method
  • Block-based robust tensor principal component analysis method
  • Block-based robust tensor principal component analysis method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0104] In order to verify the effectiveness of the present invention, a color image denoising experiment was carried out in the present invention. Its running platform is MatLab R2016a, the processor is Intel 2.60GB i5-3230M, RMB is 8GB, and a notebook with Windows 10 system.

[0105] For space size N 1 ×N 2 RGB color image, the RGB color image is essentially a 3-dimensional tensor Each channel of a color image can be seen as Of a frontal slice. The image can be approximately reconstructed by a low-order matrix. Considering that t-SVD is a multi-linear extension of SVD, in this embodiment, a low tube rank tensor is used to approximate a color image.

[0106] Apply the RBTPCA method of the present invention to the color image denoising processing, and compare the performance with the existing IBTSVT method and RTPCA method:

[0107] Randomly select 50 color images for testing. For each color image, 10% of the pixels are randomly selected, which is randomly selected in the interva...

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 block-based robust tensor principal component analysis method. By introducing a cascade operation, the whole tensor is divided into cascades of several block tensors of the same size, and an image denoising experiment is carried out in a tensor of a more suitable size. Alternating direction multiplier method divides the optimization model into two sub-problems of low rankcomponent approximation and sparse component approximation. Iterative tensor singular value soft threshold operator and iterative soft threshold operator are used to solve these two sub-problems. Thelow rank component is denoised image, and the sparse component is noise. The invention is used for extracting the low-rank component and the sparse component of the multi-path data. By introducing the block idea and adding the sparse constraint, the low-rank component is extracted from the smaller block tensor, and more accurate and clear details can be obtained.

Description

Technical field [0001] The present invention relates to the field of image processing, in particular to a tensor low-rank decomposition method based on block Background technique [0002] Tensor is multi-dimensional data, it is a high-order generalization of vector and matrix data. Signal processing based on tensor data has played an important role in a wide range of applications, such as recommendation systems, data mining, image / video denoising and restoration, etc. However, many data processing methods are only developed for two-dimensional data. It has become increasingly important to extend these effective methods to the tensor domain. [0003] Principal component analysis (PCA), as one of the most commonly used statistical tools for two-dimensional data analysis, can extract potential low-rank structures in data. However, PCA is very sensitive to large noise points or outliers, and its estimated value can be arbitrarily far away from the true value. Therefore, Robust Prin...

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): G06F17/14G06F17/16G06F17/17G06T5/00
CPCG06F17/14G06F17/16G06F17/175G06T5/70
Inventor 刘翼鹏冯兰兰陈龙喜曾思行朱策
Owner UNIV OF ELECTRONIC SCI & TECH OF CHINA
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