Minesh Mathew, Dimosthenis Karatzas and C.V. Jawahar. 2021. DocVQA: A Dataset for VQA on Document Images. _IEEE Winter Conference on Applications of Computer Vision_.2200–2209.