文章在分析了多分辨率塔形结构算法和基于遗传算法的图像匹配方法的基础上，有机地结合了这两种方法的优点提出了一种新的快速匹配算法，并与传统的基于遗传算法的图像匹配方法做了比较，试验表明算法大幅度降低了匹配时间。-In this paper, an analysis of the structure of multi-resolution pyramid algorithm and genetic algorithm-based image matching method based on the c

:研究了一种多目标识别算法,该算法用SUSAN角点形成SIFT特征点,采用阶梯图像金字塔结构实现尺度不变,为所有匹配点建立统一的超定线性方程组并对该方程组系数矩阵进行简 化使其维数降低一半,得到增广矩阵.对增广矩阵进行列变换,依据坐标转换的特性可从中提取多目标的稳定正常点,实现了快速分离多目标的匹配点. -: Study of a multi-target recognition algorithm using SUSAN corner formed SIFT feature point

Template matching is concerned with measuring the similarity between patterns of two objects. This paper proposes a massively parallel approach to pattern recognition with a large template set. A class of image recognition problems inherently