Data fusion and feature extraction using tree structured filter banks

Three feature extraction methods are considered for neural network classifiers. The first two feature extraction methods are based on the wavelet and the translation-invariant wavelet transformations. The feature extraction is in these cases based on the fact that the wavelet transformation transforms a signal from the time domain to the scale-frequency domain and is computed at levels with different time/scale-frequency resolution. The third feature extraction method is based on tree structured multirated filter banks but the tree structured filter banks can be tailored for multisource remote sensing and geographic data. In experiments, the proposed feature extraction methods performed well in neural networks classifications of multisource remote sensing and geographic data.