摘要
Zernike矩作为一种基于区域的形状描述子,具有良好的旋转不变性,能够很好地提取图像内部的形状信息,广泛应用于基于内容的图像检索(CBIR)、图像描述、物体定向及模式识别等各个领域。介绍了Zernike矩的定义和基本特性,并对Zernike矩算法提出了改进,使其具有较好的比例不变性,最后讨论了Zernike矩在基于内容的文物图像检索系统中的应用,同时给出部分实验数据和实验结果。
As a region-based shape descriptor, Zernike moments have good rotation invariance, and can capture the shape information of the object image efficiently. Recently Zernike moments are widely used in the research fields of content-based image retrieval and image description, object directional and pattern recognition. The paper introduces the definition and basic features of Zernike moments. And the improved Zernike moments also have better scale invariance. Finally, an application of improved Zernike moments＇ rotation invariance in content-based cultural relics image retrieval and some experimental data and results are given.
出处
《计算机应用与软件》
CSCD
北大核心
2007年第3期135-136,186,共3页
Computer Applications and Software
