基于修正核函数SVM的一维距离像识别

支持向量机的识别性能很大程度上依赖于核函数的使用。根据黎曼几何理论，提出了一种新的保角变换，对核函数进行数据依赖性改进。该方法通过扩大分类边界处的黎曼张量，使得分类间隔扩大，从而提高支持向量机的分类能力。针对多类舰船目标的识别，利用聚类分析中的均值距离来生成二叉树，将分类器分布在各个节点上，构成多分类支持向量机。对四类舰船目标仿真实验的结果表明，该分类方法无论识别率还是识别速度都具有优势。

Ship Target Recognition Using Range Profiles Based on SVM with Modified Kernel Function

The performance of a support vector machine（SVM） largely depends on the kernel function used. Based on the Riemannian geometry, a new conformal transformation is proposed to modify the kernel function in a data dependent way in this paper. The method is to magnify the Riemannian metric around the separating boundary, so that the separation between classes is increased and the performance of a support vector machine classifier is improved. To solve multi-class problems, the average class distance of clustering is used to construct binary tree and classifiers are distributed to each node, which constitutes multi-class SVM. Experimental results on range profiles of four ship targets demonstrate that the method is feasible.