Comments on: Eisenstein at a GIS conference? http://www.edparsons.com/2006/09/eisenstein-at-a-gis-conference/ The blog of Ed Parsons, Geographer. Mon, 06 Feb 2012 22:11:39 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Tim Woolford http://www.edparsons.com/2006/09/eisenstein-at-a-gis-conference/#comment-542 Tim Woolford Mon, 25 Sep 2006 16:50:32 +0000 http://www.edparsons.com/?p=334#comment-542 Hi Ed I agree, the issue with image tagging is it can be so subjective. Another approach could be to tag the image with OS co-ords or even TOID's so that there is some degree of commonality between our images and other databases, but again this looks like a big manual task! To start with we planning to do some basic content tagging such as Pub-Coach and Horses, Station-Warren Street, Shop-Fred etc and then develop from there. In reality I think our images sit somewhere in the middle between massively structured datasets such as OS mastermap and 'wisdom of crowds' type tagging at the other extreme. Hi Ed
I agree, the issue with image tagging is it can be so subjective. Another approach could be to tag the image with OS co-ords or even TOID’s so that there is some degree of commonality between our images and other databases, but again this looks like a big manual task! To start with we planning to do some basic content tagging such as Pub-Coach and Horses, Station-Warren Street, Shop-Fred etc and then develop from there. In reality I think our images sit somewhere in the middle between massively structured datasets such as OS mastermap and ‘wisdom of crowds’ type tagging at the other extreme.

]]> By: Ed http://www.edparsons.com/2006/09/eisenstein-at-a-gis-conference/#comment-541 Ed Mon, 25 Sep 2006 15:48:27 +0000 http://www.edparsons.com/?p=334#comment-541 Hi Tim, Interesting point, I'm not sure that a completely people driven tagging system actually works as your example image illustrates, it could be, a shopping street, a shopping street in london, an up market shopping street in london, new bond street or just Fred the jewellers.. Hi Tim,

Interesting point, I’m not sure that a completely people driven tagging system actually works as your example image illustrates, it could be, a shopping street, a shopping street in london, an up market shopping street in london, new bond street or just Fred the jewellers..

]]> By: Tim Woolford http://www.edparsons.com/2006/09/eisenstein-at-a-gis-conference/#comment-540 Tim Woolford Mon, 25 Sep 2006 10:57:44 +0000 http://www.edparsons.com/?p=334#comment-540 Hi Ed Just wanted to thank Ordnance Survey for an excellent event, certainly a lot of 'food for thought'. In the context of the 'semantic web' we should also consider how digital images are handled. While tagging images is popular (flickr etc) there is relatively little structure. Do you take the 'wisdom of crowds' approach (similar to flickr) or attempt to impose a struture for tagging images. Having location (along with date/time) is a start, but tagging the content of an image is a real challenge for example in what way could I tag this image: http://www.trekwireless.co.uk/posipix/B20050820_031_0223.jpg so that relevant OS data could be integrated with it? Hi Ed

Just wanted to thank Ordnance Survey for an excellent event, certainly a lot of ‘food for thought’.

In the context of the ‘semantic web’ we should also consider how digital images are handled. While tagging images is popular (flickr etc) there is relatively little structure. Do you take the ‘wisdom of crowds’ approach (similar to flickr) or attempt to impose a struture for tagging images. Having location (along with date/time) is a start, but tagging the content of an image is a real challenge for example in what way could I tag this image: http://www.trekwireless.co.uk/posipix/B20050820_031_0223.jpg so that relevant OS data could be integrated with it?

]]>