Some evidence for a visual slant averaging mechanism