Return to top English
A. Asano, K. Itoh, and Y. Ichioka,
Analysis of Nonlinear Filters Using Inductive Inference
Tech. Rep. Osaka Univ. 41, 2056, 221-233 (1991).


It is generally difficult to analyze nonlinear filters. We propose a novel approach for this problem. This approach utilizes inductive inference, which is an algorithmic technique that extracts general properties from a set of concrete examples. We apply this technique and derive properties of some rank-order type filters. The properties are described as a theorem that states modification of a certain local feature of images by filtering.