Abstract

There are a wide variety of street maps available on the Web, but many of these maps cannot be easily located. In addition, many of these maps lack detailed metadata that describes important information, such as the geocoordinates and scale of the maps. Image search engines, such as Google Images can be used to find maps, but accurately and efficiently identifying street maps among different images remains a challenging task. This is because image sources might include images that are not street maps, such as photographs, icons, scanned documents, political maps and climate maps. Those images are not relevant for applications which seek only street maps. These sources provide data from their own spatial databases and stored information services. They are searched on the basis of context, for example, on the basis of filenames included in a query. It is important to find a method that does content …

Date

November 4, 2005

Authors

Sneha Desai, Craig A Knoblock, Yao-Yi Chiang, Kandarp Desai, Ching-Chien Chen

Book

Proceedings of the 2005 workshop on Geographic information retrieval

Pages

35-38