It was pretty cool when GMail got the avatar capability but is there really a different motive for the feature? The Google Operating System blog is speculating that Google could use the large database of images for their face recognition software that they are supposed to be working on.
When you upload a photo in GMail it wants you to crop it so that it contains the specific portion of the image that you would like. Most people who are uploading an image of a person will choose to keep the face and scrap the rest. This will leave Google with an enormous database of faces that it can work with. Then, of course, it can associate the face with the name on the account…pretty clever, huh?
I guess that could work in an ideal situation but everyone that I know who uses GMail does not have a picture of themselves. They’ll use something like a flower or a car, but not a picture.
It will be interesting to see where this one goes but in the mean time you can have some fun with Riya.
something tells me a bunch of 100×100 pixel icons are not going to be sufficient data to fuel a face recognition system, especially with no uniformity of size, angle, or lighting. this sounds pretty bogus.
Maybe they are saving the original photo as well and using the portion that you crop to help pinpoint the location of the face. Just a thought.
-Ryan
can i see how it works? the faces??