Ricci-Flow Method for Ship Detection in SAR Image

This letter proposes a novel method for ship detection in synthetic aperture radar (SAR) images by investigating Ricci-flow locally. The proposed method can be divided into three steps. The first step in the process is to present the definitions of the statistical manifold and Ricci curvature, which are used for the description and extraction of SAR image statistical feature. In the second step, the definition of Finsler metric in Ricci curvature space is given, and a novel feature representation, which can depict the statistical information of sea clutter, is proposed to train one-class support vector machine (OC-SVM) model. Finally, the designed anomaly detection system based on OC-SVM is used to extract outlier values (corresponding to the target) from the Ricci-flow features of SAR image. Experimental results show that the feature extraction and representation method is very effective, and the designed detector has high detection performance.