Paper Title
Dimensionality Reduction Of Hyperspectral Images Using Folded Pca Approach And Classification Using Svm

Abstract
Hyperspectral imaging provides a wealth of information for each pixel in the image that is captured in a wide range of electromagnetic spectrum. The analysis of hyperspectral image is complex due to its high dimensional nature. In this paper, a dimension reduction (DR) method is presented called Folded PCA (FPCA), which can be applied to the hyperspectral image dimension reduction. The main shortcoming of the common Principal Component Analysis (PCA) method is that it does not consider spatial relation among image points. FPCA considers the spatial relation among the neighboring image pixels and also takes into account both global and local structures while preserving all useful properties of PCA. The FPCA method outperforms PCA in terms of efficiency and memory requirement which is shown through simulation results. The FPCA extracted features are given as input for Support Vector Machine (SVM) Classification. The experimental results also show the identification of four different classes from the hyperspectral images. IndexTerms- Dimension Reduction (Dr), Hyperspectral Imaging (Hsi), Principal Component Analysis (Pca), Support Vector Machine (Svm).