A new approach to select the optimum neighborhood for edge preserving filter

In this paper, we propose a new approach to select the optimum neighborhood for edge preserving filter, which is based on the minimum distance of gray value between the current pixel and its neighborhood. Compared with the earlier algorithm searching the least mean variance of neighborhood, our approach can consider not only the neighborhood character itself but also the relationship with the current pixel This advantage has been directly shown to us from the equation deducing of the new algorithm, and the experiment has good results also.