The dataset was introduced as dataset for the task3 of 2021 DocVQA challenge, that deals with VQA on infographics.
Similar to typical VQA task, task is to answer questions asked on a given infographic image. Similar to extractive QA framework popular in NLP, and the DocVQA dataset, here question-answers are primarily extractive type. But there are a small percentage of questions where answers arr not extractive.
There are 30 K questions and 5K Images in the dataset. Images are collected from the Internet. Questions and answers are manually annotated.
The dataset can be downloaded from the challenge page in RRC portal, Go to the "Download" tab in the challenge page and use the links under "Infographics VQA"
Minesh Mathew, Viraj Bagal, Ruben Perez Tito, Dimosthenis Karatzas, Ernest Valveny and C.V. Jawahar - InfographicVQA - arXiv preprint [ PDF ]
Ruben Tito*, Minesh Mathew*, C.V. Jawahar, Ernest Valveny and Dimosthenis Karatzas - ICDAR 2021 Competition on Document Visual Question Answering - ICDAR 2021 (Competition Session) [ PDF ]