About the Lab

About the dprl

The Document and Pattern Recognition Lab (dprl) started in the Summer of 2007. We focus on developing technologies for extracting and searching graphics and text in documents and videos.

Our work involves a number of Computer Science areas, including Pattern Recognition, Machine Learning, Information Retrieval, and Human-Computer Interaction. Our publications have appeared in top venues such as TPAMI, CVPR, and SIGIR, and some of our work is used by groups around the world (e.g., the Tangent formula search engine, and the CROHME handwritten math recognition benchmark).

Members. University students of all levels (BSc, MSc, and PhD) from the Computing and Imaging Science colleges at RIT have worked in the lab. See the Members page for pictures and a list of all students who have worked in the lab.

Support. The dprl has been supported by a number of organizations, including the NSF, Xerox, Google, and the Alfred P. Sloan Foundation. Please see the Support page for details.

Our Projects

Lately we are focused on the recognition and retrieval of mathematical notation for the MathSeer project. This includes our innovative MathDeck search interface based on a chips-and-cards metaphor taken from board games (Note: MathDeck is best with Google Chrome). Currently we are also running the first ARQMath task on math-aware search at CLEF 2020.

Example projects can be found on the Projects page. Pages with our publications by topic (with PDFs), software, and data are also available.