3D search engine (reverse 3D search)

As 3D scanning becomes more common, it would be useful to be able to take a 3D scan of an unknown object, submit this scan to a 3D search engine, and then receive information about it, or links to similar 3D objects. This is an extrapolation into the 3D world of a concept called reverse image search (popularized by the web service TinEye).

I would anticipate that such a search engine would need to rely on machine learning technology due to the wide range of shapes and sizes of functionally-equivalent objects. Consider, for example, the wide variety of objects such as pop-up toasters. Would a 3D scan of any given toaster result in the correct classification as a toaster?

If a library of scans (or photos) of objects is maintained, it may be possible to search for the exact make and model of a man-made object, facilitating ordering spare parts or maintenance guides for those objects that no longer bear their original labels.

The technology would not be limited to identification of man-made objects, but also identification of objects in the natural world, for example identifying invertebrates and fungi are often difficult due to the vast number of species.