Paper Info
Title
Semi-Supervised Kernel Orthogonal Subspace Projection
Abstract
The Orthogonal Subspace Projection (OSP) algorithm is sub- stantially a kind of matched filter that requires the evaluat ion of a prototype for each class to be detected. The kernel OSP (KOSP) has recently demonstrated improved results for target detection in hyperspectral images. The use of kernel helps to combat the high dimensionality problem and makes the method robust to noise. This paper presents a semi-supervised graph-based approach to improve KOSP. The proposed algo- rithm deforms the kernel by approximating the marginal dis- tribution using the unlabeled samples. The good performance of the proposed method is illustrated in a toy dataset and an hyperspectral image target detection problem.
Year
DOI
Venue
2008
10.1109/IGARSS.2008.4779696
IGARSS (4)
Keywords
Field
DocType
index terms— orthogonal subspace projection osp,kosp,kernel method,semi-supervised learning,graph.,geometry,hyperspectral imaging,matched filters,detectors,kernel,image classification,pixel,graph theory,matched filter,semi supervised learning,indexing terms,hyperspectral sensors,remote sensing,graph,probability density function
Kernel (linear algebra),Computer vision,Object detection,Semi-supervised learning,Pattern recognition,Subspace topology,Computer science,Kernel embedding of distributions,Kernel principal component analysis,Artificial intelligence,Kernel method,Variable kernel density estimation
Conference
Citations 
PageRank 
References 
2
0.39
8
Authors
3
Name
Order
Citations
PageRank
Luca Capobianco1906.20
Andrea Garzelli257441.36
Gustavo Camps-Valls32011114.02