Not really much to add to the question itself.
It is quite established that neurotypical humans are particularly good at DETECTING* faces. Of course there are false positives, i.e. we sometimes see faces where there are none, in the so-called paraeidolieae. But these false positives (together with studies on infants) teach us something important, i.e. that the main heuristic for human face detection is finding sme T-shaped contrast.
As far as I know, many machine vision algorithms work in a similar fashion.
Based on this, I expect that sunglasses should NOT dampen face detection. If anything, they can even INCREASE it.
... but is it the case? Any data on that? I'll be glad if you could point me to some references!
Thank you
* * * *
* DETECTING is different from RECOGNISING or 'READING' faces. It's not about claiming "this face is X's face" or "this face suggests X's mental state (e.g. anger)", but simply "there is some face here"