Similarity Computing of Landmarks at World Scale
It is a challenging task that recognizing landmarks at world-scale. It is very important that computing similarity between landmarks with high efficiency and great exactitude. Inthis paper, we present a novel algorithm to compute the similarity among landmarks based on vast mount of multimedia data on the web with combination text annotation and image features. And we choose to use spatial PACT to extract image features. Experimental results, which demonstrate the algorithms performance, are also shown.
landmark similarity geography cluster census transform spatial PACT
Guo Jia Zhang Honggang
School of Information and Telecommunications Engineering Beijing University of Posts and Telecommunications Beijing, China
国际会议
秦皇岛
英文
282-285
2010-11-05(万方平台首次上网日期,不代表论文的发表时间)