Photosynth Presentation at TED

Thanks to Mateo for sending me this great presentation of the PhotoSynth work emerging from Microsoft’s Live Labs.

What’s interesting to me about this is that way back when I read David Siegel’s excellent Futurize your Enterprise (from 1999) he discusses many of these kinds of applications made possible by fully described semantic data. This led me to the concept of being able to search the web to pull up all photos from a physical location, say around the Empire State Building, and then stitch them all together into a single interaction point blending everyone’s photos. What really threw me about this presentation is that the way they zoomed round Notre Dame was almost identical to how I’d thought it might work. Now the great thing about this is that they don’t rely on positioning data, but instead visual reference points that they automatically detect and stitch together, so I’m certainly looking forward to the model they create of everyone’s photos from the top of the Empire State.

Update: Having played around with the online demo (which is stunning and intuitive) I was struck by the similarity to the sequence in Blade Runner where the main protoganist Rick Deckard zooms into a slightly holographic photograph of one of the replicants he’s chasing. The way the screen panned as he focussed on sections and ‘looked round’ shower curtains is totally how PhotoSynth feels in use. The future is here again.