An Efficient Gabor Feature-Based Multi-task Joint Support Vector Machines Framework for Hyperspectral Image Classification

Abstract

In this paper, a novel multi-task learning (MTL) framework for a series of Gabor features via joint probabilistic outputs of support vector machines (SVM), abbreviated as GF-MTJSVM, has been proposed for Hyperspectral image (HSI) classification. Specifically, we firstly use a series of Gabor wavelet filters with different scales and frequencies to extract spectral-spatial-combined features from the HSI data. Then, we apply these Gabor features into the multi-task learning framework via joint probabilistic outputs of SVM. Experimental results on two widely used real HSI data indicate that the proposed GF-MTJSVM approach outperforms several well-known classification methods.

Bin Deng

Lecturer

My research interests include hyperspectral image processing, pattern recognition and machine learning.