DocumentCode
2294280
Title
Feature Extraction Using Polynomial and Sigmoidal Kernels for Classification of Radar SAR Images
Author
Enderli, Cyrille Jean
Author_Institution
Radar Dept., Thales Airborne Syst., Elancourt
fYear
2006
fDate
16-19 Oct. 2006
Firstpage
1
Lastpage
5
Abstract
This paper investigates the interest of nonlinear feature extraction for classification of radar SAR images. It is shown that polynomial and sigmoidal filter models allow to significantly improve performances of standard classifiers. In this paper, an original application of N-LDA to radar data identification, and the interest of nonlinear filter models for SAR image identification are described
Keywords
feature extraction; image classification; nonlinear filters; polynomials; radar imaging; synthetic aperture radar; SAR image classification; image identification; nonlinear feature extraction; nonlinear filter models; polynomial kernels; radar data identification; sigmoidal kernels; synthetic aperture radar; Airborne radar; Covariance matrix; Feature extraction; Kernel; Linear discriminant analysis; Polynomials; Principal component analysis; Radar applications; Radar imaging; Synthetic aperture radar; Classification; Feature extraction; Nonlinear Filtering; SAR images;
fLanguage
English
Publisher
ieee
Conference_Titel
Radar, 2006. CIE '06. International Conference on
Conference_Location
Shanghai
Print_ISBN
0-7803-9582-4
Electronic_ISBN
0-7803-9583-2
Type
conf
DOI
10.1109/ICR.2006.343505
Filename
4148482
Link To Document