This question already has an answer here:
I found out today about Google +'s image search feature. It allows users to perform a text-based search for objects within their picture library. I was instantly curious to know how this feature works. Does anyone know? Are Google using anonymous usage patterns from Google Images to build a computer vision library of CBIR object descriptors?
Possibly related is Google's 'Search by image' feature - see What algorithm does Google use for its "Search By Image" site?.