The paper describes a system, not available in Google labs as of yet, that Google has developed using a complex indexing system that draws on data already available on the web to identify famous landmarks.
Currently their project can identify up to 50,000 landmarks with an accuracy rate of 80%. Impressive stuff eh?
It’s early days but Google’s announcement hints that the area of photographic recognition could develop incredibly over the next few years. They say: “We expect the insights we’ve gained will lay a useful foundation for future research in computer vision.”