I’m the kind of person who gets annoyed when Google takes my picture without asking. So the implications of these two technologies has staggering possibilities to me. First, UC San Diego researchers managed to use a photo taken with a telephoto lens to make a functional set of duplicate keys. Combine that with the incredible new image recognition technology in the presentation at left, and the possibilities are amazing. The clip is nine minutes long, but worth a watch (and probably much better full size). The software being discussed is called Photosynth, and what it does is analyze photos for similarities to others, then uses that data to build a model of where the photos were taken. This can theoretically be done with photos from all over the web, the info being gathered much like text is gathered by Google. Now that image recognition is becoming so sophisticated, is it possible that soon those automated help voices on the phone will understand what we’re saying?