A representation and classification method of high-order and high-dimensional image data

An image data and classification method technology, applied in the field of pattern recognition, can solve the problems of data pollution and distortion, damage and loss, loss of correlation information between pixels, etc., to avoid dimensional disaster, improve classification accuracy, and reduce the number of features.

Active Publication Date: 2019-01-11
中国科学院电子学研究所苏州研究院
View PDF3 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] 1) Errors often occur in the process of image acquisition, transmission and storage, resulting in data being polluted and distorted by noise, or even damaged and missing, which is especially common in high-order and high-dimensional data
[0007] 2) Traditional methods generally use vectors to represent images, and need to expand the collected images
This process not only destroys the original spatial structure of the image, loses the correlation information between pixels, but also increases the data dimension, which is easy to cause dimension disaster and small sample problem.
[0008] The representation algorithm and the classification algorithm are often designed independently, and there are often situations where the representation algorithm and the classification algorithm do not match

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 representation and classification method of high-order and high-dimensional image data
  • A representation and classification method of high-order and high-dimensional image data
  • A representation and classification method of high-order and high-dimensional image data

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment

[0054] Step 1: Represent image data in tensor form:

[0055] For example: We have 1000 grayscale images, divided into 5 categories with 200 images each, and the image resolution is 960x1024. For each image, it is expressed as a 2-order matrix Y∈R 960×1024 , then the entire data set can be expressed as a third-order tensor Y∈R 960×1024×1000 , where the first order is the row space of the image, the second order is the column space of the image, and the third order is the sample space of the image.

[0056] Step 2: Use the original data Y to calculate the low-rank projection matrix set {U 1 ∈R 960×960 ,U 2 ∈R 1024×1024}, and project the original data into the subspace according to the mode to obtain low-rank data

[0057] The essence of step 2 is to calculate the following optimization problem:

[0058]

[0059] where λ is a trade-off parameter for balancing low rank and projection error, D ∈ R 983040×1000 is a sparse representation dictionary, A∈R 1000×1000 is the...

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 representation and classification method of high-order and high-dimensional image data, belonging to the field of pattern recognition, which solves the problems of noise influence in the process of image recognition, destruction of original image structure, high data feature dimension, incompatibility between expression algorithm and classification algorithm and the like.The technical scheme adopted by the invention is that the image data is projected from the original space into the low-rank subspace according to a pattern by using a projection matrix set to obtaina low-rank representation of the data; a sparse representation dictionary and a linear classifier are trained in the low rank subspace to classify the image data. The image data is projected from theoriginal space to the low rank subspace by using the projection matrix set, and the low rank representation of the image data is obtained. Sparse representation dictionaries and linear classifiers aretrained in low rank subspaces to classify image data.

Description

technical field [0001] The invention belongs to the field of pattern recognition, in particular to a representation and classification method of high-order high-dimensional image data. Background technique [0002] With the development of human society, a large amount of image data is collected every day, and intelligent processing and analysis of image data has become a research hotspot. The purpose of image recognition is to replace human beings with computers to complete the recognition task of image data. Image representation and classification are the focus and difficulty of image recognition research. [0003] Image representation is further divided into image preprocessing and image feature extraction. Image preprocessing includes eliminating noise and distortion, repairing missing information, and image segmentation; image feature extraction is based on task requirements, using dimensionality reduction algorithms and image processing techniques to extract more stre...

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): G06K9/62
CPCG06V10/36G06F18/241
Inventor 胡岩峰陆成韬周鹏杭谊青陆茜茜廉海明彭晨
Owner 中国科学院电子学研究所苏州研究院
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