Microsoft searched flickr for all photos tagged “Notre Dame.” From the partially overlapping segments across a multitude of photos, they deduced the position of the photographer for each photo (the yellow triangles show the calculated place, zoom, and angle of the shot). Then they reconstruct the 3D space, here seen as a white scatter-plot outline. A composite image, made from overlapping potions of all of the flickr images, can then be examined and zoomed in to.
Some of the images were partial scenes of a photograph of a poster with Notre Dame on it, but they contributed none-the-less.
Blaze from Microsoft says that this is a way of automatically creating geospatial hyperlinks between images.
It was a very cool splash from flickr-space.
Microsoft introduced this as an example of how they are innovating, but once again, it was an acquisition. =)
Leave a Reply