摘要
在图像处理与模式识别领域经常根据物体的边界来识别物体。对物体边界特征描述的三种主要方法,即基于极半径函数的傅里叶描述子、边界序列矩和改进的不变矩进行了比较研究,给出了实验结果和结论。
description or discrimination of boundary curves (shapes) is an important problem in image processing and pattern recognition. Three main boundary feature description methods for object recognition are the Fourier descriptors, contour sequence moments, improved moment invariants. The performance of each of the methods was compared, and the experimental results were given. Brief conclusions of the work were presented.
出处
《光学技术》
CAS
CSCD
2003年第3期284-287,共4页
Optical Technique
基金
国家自然基金项目(60272094)资助
山东省自然科学基金项目(Y2001G10)
济南市青年科技明星计划资助