I'm trying to compile a database containing images of exemplars for various categories. Given cultural bias I would start with and American perspective.
I would also like to include images for broad categories. A tabby cat would be an exemplar for "cat", and could also be used from "mammal".
This database would be used primarily for language instruction, but tying it to the ImageNet API would be helpful to others.
Does anyone know where to start on such a project? Certainly ImageNet is a great start. I'd also be interested in replicating the methods used for Google Image Swirl. It is sad that both Image Swirl and
Bing Visual Search have both fallen by the wayside.