Parameter Optimization of Median epsilon-Filter Based on Correlation Maximization

Median epsiv-filter is a nonlinear filter, which can reduce the impulsive noise with preserving the edge information. It does not require the learning process of signal and noise models. The algorithm is simple and the calculation cost is small because it requires only switching operation. However, it requires the empirical operation to set the appropriate parameter of median epsiv-filter. There is also no criterion to decide the optimal parameters. This paper therefore presents an optimal parameter estimation of median e-filter based on correlation maximization. Experimental results are also given to show the effectiveness of the proposed method.