Sunday, January 31, 2010

Uncanny valley

I was thinking about causes the uncanny valley effect and I think it has to do with a couple of things working in unison. First is the human’s brain ability to compare things and find dissimilarities. The second is the ability to compress faces. In terms of compression, the amount of data that a face has is huge. Think of how large the file would be if you took a high resolution JPG of someone’s face. As we encounter hundreds of people on a daily basis it would be impossible to store all of that information at full resolution. As a efficient compromise the brain compresses these images so that you can still recognize individual faces but do not have to store full resolution data. The final component is human’s wariness of disease. This is manifested as how we have unease when we see someone with a physical abnormality even though we understand that the underlying cause is not due to a contagious disease. I think that we are processing images of CG generated people to be “diseased” people.
I guess one way to test this would be to read fMRI scan of the brains of people who look at these images.

No comments:

Post a Comment