作者: SVEN BUCHHOLZ , NICOLAS LE BIHAN
DOI: 10.1142/S0129065708001403
关键词:
摘要: For polarized signals, which arise in many application fields, a statistical framework terms of quaternionic random processes is proposed. Based on it, the ability real-, complex- and quaternionic-valued multi-layer perceptrons (MLPs) performing classification tasks for such signals evaluated. multi-dimensional neural networks relevance class label representations discussed. signal to noise separation it shown that MLP yields an optimal solution. Results two different are also reported.