Hacker News Reader

Comment by tmalsburg2 | Hacker News Reader

Not an expert for face perception but I think faces are believed to be stored in memory in terms of their deviation from the standard face (the average face in your environment). Since beautiful faces happen to be precisely the faces that are close to the average, it's completely unsurprising that they are hard to capture because their deviations from the standard face are really subtle. Storage of faces as differences from the standard face also explains why white people find it hard to distinguish Asian faces (and the reverse): they all deviate from a white standard face in a similar way. Caricatures can also explain by this because a caricature is what you get when you amplify someone's deviations from the standard face.

View on HN

Nice explanation. Is there any literature that explains how faces are stored in memory? Fascinating.

The CP is discussing basically prototypes/schemas (essentially; schemas are debunked now but they're useful for understanding by analogy how connectionist cognitive models work); I don't know of any pop science books ala Thinking Fast and Slow about this topic but it's not hard to research. I would start with schema theory and other derived memory representations, the TRACE word recognition model, semantic networks, and neural networks.

PsychInfo is probably the best database for this, if you can't get access to it I'd recommend using google scholar and libgen.

https://en.m.wikipedia.org/wiki/Cross-race_effect

Faces are processed in an entirely different part of the brain than regular vision, which is why you can be blind but still recognize people. (Or vice versa if you're face blind.)

While I'm inclined to agree with you, do you have any supporting evidence?

I like this explanation too. It also fits with my thinking that when you think someone looks just like someone else, ask another person and they can't see it. I believe this is down to the way our brain recognises matching features against the stored 'standard face' of an person, but other people don't match in the same way as their 'standard face' is different.

Is that similar to the concept of the Eigenface?

https://en.wikipedia.org/wiki/Eigenface

Beauty is in the eye of the beholder, though. I would say it is something similar to a distinct symmetry.

Nice model, but unfortunately it's not true. Neocortex doesn't have 'special face diff storage mechanism'. It doesn't work like this.

I doubt scientists know that much on how brain works.

Really? How do you know this? I thought that our knowledge about how things are stored in the brain was extremely poor.

It seems that we do have some pretty specialised mechanisms in our heads for dealing with human faces, given how good we are at recognising them compared to other things. Even tiny changes in face shape (ie expression) are instantly recognisable and can cause major emotional responses.

Reply on news.ycombinator.com