Purely speculative: several research institute in Switzerland have develop apps to recognize mountain peaks, I wonder if, using the pictures of monuments in Commons, such institute couldn’t develop a database of « profiles » used to recognize monuments in Switzerland. Based on the the geolocalization of the camera you should easily determine which profile to use for each monument?
charles ___________________________________________________________ I use this email for mailing list only. Charles ANDRES, Chief Science Officer "Wikimedia CH" – Association for the advancement of free knowledge – www.wikimedia.ch Skype: charles.andres.wmch IRC://irc.freenode.net/wikimedia-ch Le 6 janv. 2014 à 21:02, rupert THURNER <rupert.thur...@gmail.com> a écrit : > Fyi > > ---------- Weitergeleitete Nachricht ---------- > Von: "john cummings" <mrjohncummi...@gmail.com> > Datum: 05.01.2014 01:00 > Betreff: Re: [cultural-partners] Successor to QRpedia? > An: <kerry.raym...@gmail.com>, "Wikimedia Chapters cultural partners > coordination - closed list" <cultural-partn...@wikimedia.ch> > > Hi All > > I have been working on something using the redirecting service that QRpedia > provides without using the actual QR code for the Natural History Museum in > London (final write up coming soon). > > We get around 50% foreign language visitors and offer them very little (apart > from a newly released audio guide). > > What I've made is basically a photographic representation of the museum that > links to the Wikipedia articles for the species, I've included a link to the > mobile site version as it looks nicer > > https://en.m.wikipedia.org/wiki/Wikipedia:GLAM/NHMandSM/NHM_Galleries > > I think this offers a few key things: > > 1. It offers people the chance to work together to curate a museum (or any > other space) in their own way with many different perspectives, e.g > > * Butterfly identification guide for our butterfly house > > * Mammals Gallery gives Wikipedia articles for species in every display case > > * Browse a book that's on display (The Birds of America) with links to each > species in your own language > > * Being able to augment displays already in the museum: > > * Add to the Giant Sequoia cross section we have that has names and dates > on it's rings with links to those events and people > > * To link to Wikipedia articles from text on museum signs and to offer > people the change to do translations of that museum sign in their own > language. > > * Tree identification guide (not finished yet) > > 2. It has very low technology barriers compared to something using an app, it > works on any browser on most devices, including odd things like Blackberrys > and microsoft tablets > > 3. You don't have to put QR codes everywhere, sometimes it's not possible. > > 4. Anyone can do it for their local museum, art gallery, nature reserve etc > at no cost and without permission. > > I played around with quite a few things before coming up with this including > using Layar (image recognition) and google maps with geocoding, both being > hard work and clumsy and not working very well when things are close together > or 3d objects or the software doesn't allow collaboration. > > I have no idea where this sort of thing should live, any ideas? > > At QRpedia's heart it's a Wikipedia language redirect that happens to have a > QR code as the way of getting there, perhaps it just needs a different name, > Wikipedia language link?. I'd love to see QRpedia being used in a lot of > places, for instance I would love to see QRpedia being used on OpenStreetMap > links to Wikipedia and also for other projects like Wikivoyage. > > Cheers > > John > > > On 4 January 2014 23:12, Kerry Raymond <kerry.raym...@gmail.com> wrote: > I agree that object recognition is still hard, but much more straightforward > thing to do is to map from a location to the set of Wikipedia articles > geocoded as being "near" that location, which is useful in itself. Then it's > easy to have a mobile device app that shows Wikipedia articles "near me > now". > > In fact, this is all so straightforward I'm guessing someone's already built > it, although a quick google search hasn't turned it up for me. > > Obviously "near" is something that needs to be configured according to the > specific type of app you are dealing with. "Near" in an art gallery is > different to "near" while travelling on a train. > > Practical problems are the accuracy of geocodes both in terms of: > 1. the geocodes on Wikipedia articles (particularly for regions as opposed > to points) - how do I get the Wikipedia article for the city as a whole if I > am out in the suburbs? > 2. the geocodes generated by mobile devices, which will be the limiting > constraint for indoor applications > > All of the above equally applies to Commons and other projects where > geocodes are used. > > Kerry > > > > _______________________________________________ > Cultural-Partners mailing list > cultural-partn...@wikimedia.ch > https://intern.wikimedia.ch/lists/listinfo/cultural-partners > Please treat emails sent to this list as confidential.Ask senders for > permission before forwarding emails off-list. > > > _______________________________________________ > Cultural-Partners mailing list > cultural-partn...@wikimedia.ch > https://intern.wikimedia.ch/lists/listinfo/cultural-partners > Please treat emails sent to this list as confidential.Ask senders for > permission before forwarding emails off-list. > _______________________________________________ > http://wikimedia.ch Wikimedia CH website > Wikimediach-l mailing list > https://lists.wikimedia.org/mailman/listinfo/wikimediach-l
_______________________________________________ http://wikimedia.ch Wikimedia CH website Wikimediach-l mailing list https://lists.wikimedia.org/mailman/listinfo/wikimediach-l