7 November 2008 New methods of H-SVMs for the classification of multi-spectral remote sensing imagery
Author Affiliations +
Proceedings Volume 7147, Geoinformatics 2008 and Joint Conference on GIS and Built Environment: Classification of Remote Sensing Images; 714706 (2008) https://doi.org/10.1117/12.813206
Event: Geoinformatics 2008 and Joint Conference on GIS and Built Environment: Geo-Simulation and Virtual GIS Environments, 2008, Guangzhou, China
Abstract
Through systematically analysises of existing multi-class SVMs (M-SVMs) methods, it is shown that hierarchy multi-class SVMs (H-SVMs) can be relatively effective. Further analysis shown that existing methods that measure separability between different classes are not suitable for kernel feature space. A new method is presented for separability measure in feature space based on the characters of RBF kernel function and SVMs. Based on the new separability measure, two kinds of H-SVMs, Binary Tree SVMs (BT SVMs) and Single Layer Clustering SVMs (SLC SVMs) are presented. They are both implements of following ideal: the higher a pair of two sub-classes is in the hierarchy, the easier to separate them. In this way, we can not only achieve classification accuracy by alleviate error accumulation from top to bottom, but also rise classification speed by reduce support vectors in classifier. Experimental results justify the rationality of the new separability measure and effectiveness of BT SVMs and SLC SVMs.
© (2008) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Hou Bin, Hou Bin, } "New methods of H-SVMs for the classification of multi-spectral remote sensing imagery", Proc. SPIE 7147, Geoinformatics 2008 and Joint Conference on GIS and Built Environment: Classification of Remote Sensing Images, 714706 (7 November 2008); doi: 10.1117/12.813206; https://doi.org/10.1117/12.813206
PROCEEDINGS
13 PAGES


SHARE
Back to Top