How does Google Street View know the numbers on houses? By using a crazy neural network of course. Because finding the number on a house is quite a task, Google engineers built a "deep convulutional neural network" that operates on pixels of the images themselves. It combines them for number imagery and logging for what it finds.

According to Google, "We can, for example, transcribe all the views we have of street numbers in France in less than an hour using our Google infrastructure," write the engineers in a new Arxiv paper about the project.

If the numbers are too blurry for them, they are identified by humans as part of a second generation CAPTCHA program. The paper is an interesting read. Click through to read it here: [Arxiv; MIT Technology Review]