On the use of Gaussian synapse ANNs in multi and hyperspectral image data analysis

A new type of artificial neural network is used to identify different types of crops and ground elements from hyperspectral remote sensing data sets. These networks have Gaussian synapses and were trained using the GSBP algorithm. The intrinsic filtering ability of the Gaussian synapses permit concentrating on what is relevant in the spectra and automatically discarding what is not. In addition, the networks are structurally adapted to the problem complexity as superfluous synapses and/or nodes are implicitly eliminated by the training procedures, thus pruning the network to the required size straight from the training set.