摘要
物体的对称性检测是图像分析、计算机视觉中一个重要研究课题 .在形状匹配、基于模型的目标匹配、三维物体的重建、图像压缩和图像数据库检索等许多领域都有重要应用 .但是 ,对称性检测一直是计算机视觉的难题 ,至今缺少通用性好 ,能同时检测各种对称性并且效率较高的算法 .利用隐含多项式曲线对复杂物体的良好描述能力 ,提出了基于隐含多项式曲线的物体对称性检测方法 .它将平面图形的对称性检测问题转化为对隐含多项式曲线的研究 ,理论与实验表明该算法能检测物体的各种对称性 ,算法简单 。
Symmetry detection of the object is an important research area in image analysis and computer vision, and it is usually applied in shape matching, model-based object matching, reconstruction of 3D objects, image compression, image database retrieval and so on. Fast and effective symmetry detection is still a difficult problem in computer vision, and none of the existing algorithms can detect all the types of symmetry of a 2D image. In this paper an algorithm to detect symmetry is proposed based on the implicit polynomial curves, which can efficiently describe the complex object. The algorithm converts the symmetry detection of planar image to the analysis of implicit polynomial curves. The theory and experiment show that all types of symmetry of object including noise and loss of information can be detected with the algorithm.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2002年第10期1337-1342,共6页
Journal of Computer Research and Development
