Text this: Image region semantic enhancement and symmetric semantic completion for text-to-image person search