Workshop on Document Visual Question Answering (DocVQA 2021) Call for Papers

Workshop on Document Visual Question Answering (DocVQA 2021) 


Visual Question Answering has become a key task in the vision and
language field, while it has also become clear lately that there are
numerous questions of common interest which cannot be answered unless
written information in the image could be read and understood in the
context provided by the visual information.

Document Visual Question Answering (DocVQA) aims to bring VQA to the
Document Image Analysis field, that focuses on understanding written
communication in images. DocVQA is proposed as a generic paradigm for
purpose-driven document analysis and recognition, where natural
language questions drive the information extraction and document
understanding processes. This challenges current practice in Document
Image Analysis Recognition where research has historically focused on
generic bottom-up information extraction tasks (character recognition,
table extraction, word spotting), largely disconnected from the final
purpose the extracted information is used for.


Invited Speakers 

Amanpreet Singh (Facebook AI Research)  Towards models that
can read and reason about scene text

Brian Price (Adobe Research Labs)  Understanding Data
Visualizations via Question Answering

Yijuan Lu (Microsoft Azure AI) Scene Text-Aware Pre-training
for Text-VQA and Text-Caption


Challenge Session

Winners of the 2021 edition of DocVQA challenge will be presenting
their winning submissions at the workshop



The DocVQA 2021 workshop will take place at the Int. Conf. on Document
Analysis and Recognition (ICDAR) on September 6, 2021, in the
afternoon session as a virtual event.


Minesh Mathew,  IIIT Hyderabad, India 

Ruben Perez, Computer Vision Centre, Spain 

Dimosthenis Karatzas,  Computer Vision Centre, Spain 

C.V. Jawahar, IIIT Hyderabad, India 

R. Manmatha, Amazon, USA