摘要
针对SIFT算法在极值点搜索和特征计算方面的低效,提出一种基于分维搜索和环状描述符的SIFT匹配算法(SC-SIFT)。该算法将SIFT算法中的三维极值点搜索策略分解(separate)为两个维度上的逐维搜索,同时引入了一种新的环状(cricoid)特征描述算子来代替原来高维低效的特征。实验证明,该方法不仅能够提高SIFT算法的执行效率,而且提高匹配正确率,实现了对SIFT算法的优化。
A improved SIFT（SC-SIFT）,based on separated dimension searching and cricoid feature descriptor,was proposed to deal with the poor efficiency in the Maxima and minima point searching and feature compute.The searching method in new algorithm works in two dimension first,and then in the third dimension,which in the original algorithm is working in three dimension once.A new cricoid feature descriptor used in the new algorithm instead of the original descriptor which is high dimension but low efficency.Experimental results show that：the new algorithm can not only improve the execution efficiency of SIFT,but also raised the matching accuracy,optimize the SIFT algorithm.
出处
《电脑开发与应用》
2010年第5期1-4,共4页
Computer Development & Applications
基金
山西省科技攻关基金资助项目(20080322008)
关键词
SIFT
分维搜索
环状特征描述符
匹配正确率
SIFT method
search of separate dimension
cricoid feature descriptor
matching accuracy
