أرسل هذا في رسالة قصيرة: VIIDA and InViDe: computational approaches for generating and evaluating inclusive image paragraphs for the visually impaired