Volume XXXIX-B3
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XXXIX-B3, 503-508, 2012
https://doi.org/10.5194/isprsarchives-XXXIX-B3-503-2012
© Author(s) 2012. This work is distributed under
the Creative Commons Attribution 3.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XXXIX-B3, 503-508, 2012
https://doi.org/10.5194/isprsarchives-XXXIX-B3-503-2012
© Author(s) 2012. This work is distributed under
the Creative Commons Attribution 3.0 License.

  01 Aug 2012

01 Aug 2012

SPECTRAL REGRESSION DISCRIMINANT ANALYSIS FOR HYPERSPECTRAL IMAGE CLASSIFICATION

Y. Pan, J. Wu, H. Huang, and J. Liu Y. Pan et al.
  • Key Lab. on Opto-electronic Technique and systems, Ministry of Education, Chongqing University, Chongqing, P.R. China, 400044

Keywords: Dimensionality reduction, Hyperspectral image classification, manifold learning, Eigen-decomposition, Spectral regression discriminant analysis, Embedding function

Abstract. Dimensionality reduction algorithms, which aim to select a small set of efficient and discriminant features, have attracted great attention for Hyperspectral Image Classification. The manifold learning methods are popular for dimensionality reduction, such as Locally Linear Embedding, Isomap, and Laplacian Eigenmap. However, a disadvantage of many manifold learning methods is that their computations usually involve eigen-decomposition of dense matrices which is expensive in both time and memory. In this paper, we introduce a new dimensionality reduction method, called Spectral Regression Discriminant Analysis (SRDA). SRDA casts the problem of learning an embedding function into a regression framework, which avoids eigen-decomposition of dense matrices. Also, with the regression based framework, different kinds of regularizes can be naturally incorporated into our algorithm which makes it more flexible. It can make efficient use of data points to discover the intrinsic discriminant structure in the data. Experimental results on Washington DC Mall and AVIRIS Indian Pines hyperspectral data sets demonstrate the effectiveness of the proposed method.