Shape vs. Context: Examining Human--AI Gaps in Ambiguous Japanese Character Recognition

Re-assign community