Resources
Public datasets, code, and presentations from our research.
Datasets
Publicly available datasets introduced in our works.
- 2021 HW-SQuAD and BenthamQA — QA over document images — Download · Paper
- 2021 Infographic VQA — Downloads
- 2020 DocVQA — Downloads
- 2020 RoadText1K — Project page
- 2018 LectureVideoDB — Project page
- 2018 IIIT Handwritten words (Devanagari & Telugu) — Project page
- 2018 IIIT Urdu OCR dataset — Project page
- 2017 IIIT Arabic dataset — Project page
- 2017 Synthetic scene text (Hindi, Telugu, Malayalam) — Download · Project · Paper
- 2017 IIIT-ILST (Hindi, Telugu, Malayalam) — Download · Project · Paper
- 2016 Hindi 100 pages (printed text) — Mirror 1 · Mirror 2 · Paper
Code
Demos & Presentations
- 2022 InfographicVQA at WACV 2022 — video
- 2021 DocVQA workshop at ICDAR 2021 — all talks
- 2021 Asking questions on HW document collections (ICDAR 2021) — video
- 2021 DocVQA at WACV 2021 — video
- 2020 Text and Documents workshop — DocVQA challenge — video
- 2016 OCR project at CVIT — video
- 2016 Capture and read (OCR + TTS Android app) — video 1 · video 2