KAZE Feature Point with Modified-SIFT Descriptor
international conference on model transformation(2013)
摘要
In this paper, we present a novel descriptor, called Modified-SIFT (M-SIFT), which is more suitablefor KAZE feature point than other descriptors such as: SURF and SIFT. Their combining even outperformsSIFT matching algorithm and standard KAZE algorithm. We present extensive experimental image matching results on the Mikolajczyk and Schmid datasetwhich show clear advantages of our descriptor against others.
更多查看译文
关键词
feature descriptor,local feature,image match
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要