some stuff on image-based location

[Clayton H Lewis]

kind of... this is work by qin lv of colorado and colleagues on:

*matching in large collections of big, high dimensional things, like  
images http://www.cs.princeton.edu/cass/papers/mplsh_vldb07.pdf

*doing image searching on a mobile device (though not in the same way  
as above): http://www.cs.colorado.edu/~lv/docs/mobisys09.pdf

qin has some students who will do an experiment using their  
techniques on a mobile device to determine the location within their  
lab... I'll report results when available

one of qin's collaborators, li shang, points out that there is  
context information that can be used to improve image matching, such  
as that a picture taken not long after an earlier picture can't be  
very far from it... using this idea one might have quite a large  
collection of images, giving fine grained spatial coverage, but not  
have to search very much of it, on the average

Clayton Lewis
Professor of Computer Science
Scientist in Residence, Coleman Institute for Cognitive Disabilities
University of Colorado
http://www.cs.colorado.edu/~clayton



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://fluidproject.org/pipermail/fluid-work/attachments/20090503/1f2c9abd/attachment.html>