Friday 15 March 2013

AnthropoloG+

Hi all,
this post is nothing serious, maybe just fun, but somehow interesting. 
Today I was reading last post of Cicero Moraes in G+; it regards an exposition about about evolutionary anthropology which will soon take place in Brazil. I will not write about this exposition here, because I hope that Cicero will do a post in the next days, but I want to report what I noticed when I loaded the poster of the exhibition in G+ (you can see it in the image below)...

Poster of the exhibition (by Cicero Moraes)
... I knew that G+ added a facial recognition application, so that it is now simpler to find human faces in the user's uploaded pictures, what I did not expected was that G+ tried to recognize the faces inside the exhibition's poster. The result is what you can see in the short clip below:




the Taung Child (Australopithecus africanus), the Australopithecus afarensis and the Homo habilis are not considered human faces by G+. On the contrary, the facial recognition application is able to find the Turkana Boy (Homo erectus), the Homo floresiensis, the Homo erectus pekinensis, the Homo heidelbergensis, the Homo neanderthalensis and, of course, the Cro-Magnon (early Homo sapiens sapiens). 
To verify that the facial recognition was not due to Cicero tags, I downloaded the image and uploaded again on my account and... I get the same result. Moreover, when I uploaded the picture, the software asked me to tag the people of the "photo", recognizing the same six faces, as you can see in the image below.

The facial recognition during uploading process

I do not know if this report can be interesting for an anthropologist; maybe Moreno Tiziani could better analyze this phenomenon, but I just wanted to inform you about this strange anthropological attitude of a social network, let's say this AnthropoloG+ application :).

2 comments:

  1. Well, it's very strange, maybe we have to know how the recognition of faces "really" works: what are the parameters taken by the script?

    ReplyDelete
  2. Hi Moreno, I guess it is something related with the nose... If I find the time I will do some test ;)

    ReplyDelete

BlogItalia - La directory italiana dei blog Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.