> It does not support the claim that corners are in any way special for human vision

Your binocular field of vision is a round rect: https://openbooks.lib.msu.edu/neuroscience/chapter/vision-ce...