AN ADAPTIVE FOURIER ESTIMATOR FOR 2D-SIGNAL PROCESSING

An adaptive least mean square estimator for modelling data through the discrete bi-dimensional Fourier Transform is introduced. Such an esti mator presents a relevant potential value for applications concerning spatial data, and especially image analysis, extraction of 2D power sp ectra, tracking of bidimensional signal embedded in noise. Besides the formal description, considerations about the theoretical analysis of the device, ifs biological plausibility, ifs applicability to the trea tment of diverse kinds of signals, are also provided. The performance of the estimator is evaluated by means of several evaluation tasks; th e results provide evidence in favour of the feasibility of dynamic 2D signal modelling by the device.