This particular new service has been described in Amazon’s patent application as the “User guided object identification”. Should this particular service be implemented, we do wonder just how widespread its use will be, and will this idea actually catch on in the long run? No longer do you need to make some sort of lettered query, as a photo query of a particular object would do.
As for the other items, the process would be a little bit more complicated. In order to locate and match a shoe, a handbag or other real world item, you will need to do your bit when snapping a photo. For instance, you will need to align a sneaker or a boot in the camera viewfinder/screen so that it will match an outline provided by Amazon, in order to increase its accuracy level.