Adaptive Max-Margin One-Class Classifier for SAR Target Discrimination in Complex Scenes

Synthetic aperture radar (SAR) target discrimination is an important stage that distinguishes targets from clutters in the automatic recognition field. However, complex SAR scenes, performance of some traditional discriminators will degrade. As effective tool for one-class classification (OCC), max-margin classifier has attracted much attention discrimination, as it can effectively reduce impact multiple clutters. very sensitive to values kernel parameters. To solve problem, this paper proposes adaptive scenes. In a with suitable parameter, distance between sample and boundary satisfies certain geometric relationship, i.e., edge samples input space are transformed region close boundary, while interior far away boundary. Therefore, we define information entropy measure automatically obtain optimal parameter classifier, first selected, then optimization performed by minimizing simultaneously maximizing samples. Experimental results synthetic datasets measured validate effectiveness our method.