Abstract: The recognition of the actual ship noises based on the kernel Fisher discriminant analysis is investigated. The main idea of this method is to find a nonlinear direction by first mapping the data nonlinearly into some feature space and compute Fisher's linear discriminant there, thus implicitly yielding a nonlinear discriminant in input space. The mechanism of the kernel Fihser discriminant analysis is particularly analyzed and the classification algorithm is developed. The recognition results are encouraging. In addition, this method is compared with other state-of-the-art classification techniques. Experiment shows that the kernel Fisher discriminant(plus a linear support vector machine to estimate the threshold) is superior to other algorithms.

Key words: target recognition, Fisher discriminant, kernel function, feature space