Image analysis using radial bi-Krawtchouk moments
Image analysis using radial bi-Krawtchouk moments
复制标题
DOI：
10.1117/12.2010862
复制
发表时间：
2013-03
期刊：
影响因子：
--
通讯作者：
Bo Li;B. Fu
中科院分区：
文献类型：
--
作者：
Bo Li;B. Fu
径向离散正交矩因其良好的特性而经常用于表征图像形状，以应用于模式识别和图像分析。在本文中，提出了一组新的径向 Krawtchouk 矩。与经典的径向离散矩相比，所提出的矩采用了一种新的采样方法，即针对不同半径使用不同数量的采样点。另外，通过使径向正交的离散Krawtchouk多项式和周向正交的离散Krawtchouk多项式构造新的离散正交向量集。实验结果表明，所提出的矩在图像重建和计算效率方面具有良好的性能。
Radial discrete orthogonal moments are often used in characterizing image shapes for applications in pattern recognition and image analysis for their good properties. In this paper, a new set of radial Krawtchouk moments is presented. Compared with classical radial discrete moments, the proposed moments take a new sampling method, which uses variational number of sampling point for different radius. In addition, a new discrete orthogonal vector set is constructed by making the discrete Krawtchouk polynomials orthogonal in radial direction to the discrete Krawtchouk polynomials orthogonal in circumferential direction. Experimental results show that the proposed moments have good performance in image reconstruction and computing efficiency.