Do computational models of vision need shape-based representations? Evidence from an individual with intriguing visual perceptions.