Exploiting Class Hierarchies for Knowledge Transfer in Hyperspectral Data

Obtaining ground truth for hyperspectral data is an expen- sive task. In addition, a number of factors cause the spectral signatures of the same class to vary with location and/or time. Therefore, adapting a classier designed from available labeled data to classify new hyperspec- tral images is dicult, but invaluable to the remote sensing community. In this paper, we use the Binary Hierarchical Classier to propose a knowledge transfer framework that leverages the information gathered from existing labeled data to classify the data obtained from a spatially separate test area. Experimental results show that in the absence of any labeled data in the new area, our approach is better than a direct applica- tion of the old classier on the new data. Moreover, when small amounts of labeled data are available from the new area, our framework oers further improvements through semi-supervised learning mechanisms.