(Canceled Due To Volcanic Ash) Visual Recognition, the Future of Mobile AR browsers

The unarguable success of certain mobile devices comes from the user experience. On top of that, people are getting used to quick access to relevant information. The sum of both obviously determines those applications that have an impact on the mobile market.
The first point is fully covered in the first time somebody uses a Mobile AR browser. Once this wow effect is passed what happens is that one discovers the other side of the coin. One can be sitting at home, seeing the walls of the room and, overlaid, a picture of the major touristic attraction of the city. "Well, how is that linked to my room and relevant to the wall my mobile is pointing at?" one could say. Some might argue that it is valuable to know in what direction and distance that attraction is. In general it is, but what is the real benefit as compared to a map?
We covered a rather controversial point, the utility of browsing content unrelated to the view. Let us now cover another aspect, in this case, a fact. GPS antennas have an error of several meters (depending on many factors, such as weather or sky visibility). Such error introduced into the point of view of the user is non-relevant for distant points of interest (POI). What happens with closely located places is another story. An error of 2 or 3 meters is enough for the display to show information about a POI behind us, just in the middle of the screen (i.e. as if it was in front of us). Imagine we are looking for a review of a restaurant that is in front of us,... should we be seeing the review of another restaurant one block away? Here is where the user experience breaks.

On a separate track but naturally linked to AR, the computer vision community has been working on visual recognition for many years now. This technology enables to identify an object by taking a picture of it and querying a database of reference images. The step from academic research to commercial products has in this case been much shorter. There are several mobile applications that perform visual recognition of objects such as posters, cd/dvd covers or books.
Mobile Visual Search applications provide a natural interaction with the environment. Firstly, because the information or services provided are specific to THAT object that we are interested in; secondly and also important, because it skips the sometimes cumbersome step of typing text on a mobile device.

The future of Mobile AR goes naturally through Mobile Visual Recognition. An enhanced mobile view will be available through services that are relevant for the user. Identifying what the user is seeing is the key factor for an advanced Mobile AR user experience.