Document and Pattern Recognition Lab
( dprl@RIT )

We research systems that recognize and retrieve information
in documents, images, and videos.

News: 2023

  • (Nov, 2023) A paper describing the dprl's ChemScraper parser for molecular diagrams in PDF drawing instructions ('born-digital') is available on arXiv here. The system can also generate annotated training data for visual parsers that recognize raster images (i.e., pixel-based, such as PNG). A link to associated code is provided in a footnote. Congratulations to Ayush Kumar Shah, Bryan Manrique Amador, Abhisek Dey, Ming Creekmore, and Blake Ocampo (PhD candidate, UIUC Dept. of Chemistry) on a job well done.
  • (Sept, 2023) Congratulations to former dprl PhD student Wei Zhong, who successfully defended his dissertation on math-aware search at the University of Waterloo (advisor: Jimmy Lin). Wei had to switch schools and countries due to visa restrictions during COVID. This past summer summer he also worked as a research intern at Microsoft research.
  • (Aug 24, 2023) I gave the keynote talk at GREC 2023 in San Jose, which was held as part of ICDAR 2023.  The talk was an overview of MathDeck and related work in math formula recognition and search. My thanks to everyone who attended, it was a very good experience!
  • (July 11, 2023) The lab's poster and longer video demonstration for our MathDeck demo at SIGIR 2023 are now available. Live demo page link is here.
  • (June 27, 2023) The first initial release of the ChemScraper tool developed by the dprl, Denmark Lab, and NCSA has been released.  There are plans to regularly update the tool moving forward; look for updates in the coming months.
  • (June 27, 2023) Abhisek Dey and Ming Creekmore presented a poster on our lab's work on extracting and recognizing chemical diagrams for the MMLI ChemScraper system for an NSF Site visit with the MMLI AI Center at UIUC (in Champaign, Illinois). Ayush Kumar Shah and Bryan Amador, who also contributed to the system participated in the meeting as well.
  • (April 19, 2023) Ayush Kumar Shah published a paper on an improved math formula parsing model using line-of-sight graphs (the Line-of-sight with Graph Attention Parser (LGAP), previously named QD-GGA). The paper will be presented at ICDAR 2023.
  • (April 1, 2023) A demonstration paper for the MathDeck system searching text and formulas in PDF files from the ACL Anthology will appear at SIGIR 2023. Bryan Amador, Matt Langsenkamp, Abhisek Dey, and Ayush Kumar Shah created the demo. A number of past DPRL students made important contributions as part of the MathSeer project.

Room 70-3500, GCCIS
Dept. Computer Science
Rochester Inst. Technology
Rochester, NY, 14623-5608


Phone: +1 (585) 475-4536
Fax: +1 (585) 475-4935