Good Evening
Running late again, so here we go.
Both the redaction art, and the face finding, made me think of other ways in which clever scripts can glean information from databases. Could programs search collections to help identify items? The idea being that the image of the item is being used to supply information, rather than curatorial data or tags. I’m thinking of physical collections rather than archives, however, I doubt that would be necessary, as most items photographed for collections should have descriptive information attached prior to photography right? I wonder what collections WOULD find that useful. I guess that would make it a form of datamining, where the program is looking for images/shapes rather than text. I think it would be useful in some of the crowdsourcing projects we saw earlier in the semester on Zooniverse.
I work at the Australian War Memorial, where there is a running project to find portraits of all the first world war service personnel. I wonder if the collection has been searched with facial recognition software? Perhaps something to ask the curators tomorrow.
Also, here is my tweet for the Vintage Face Depot. Hmmmm
I just tried out the IBM demo system, and it’s pretty cool! Here is what it had to say about the same photo I used for Trove
Pretty good, but I’m not particularly wow’d by how old Watson thinks I am.
What sort of half-rate super computer can be thrown off by twelve years by just a little pre-mature balding?
Anyway, I couldn’t get the demo to work on the redaction data, I’ll have to try again later.
Goodnight all!
