Department of Computer Science
Rochester Institute of Technology
Phone: (585) 475-4536
|[ Home ]||[ News ]||[ Members ]||[ Projects ]||[ Publications ]||[ Software ]||[ Support ]|
News (2017)(Earlier News, 2008-2016)
- (Dec.) In December, the National Science Foundation (USA) awarded a grant to myelf, Anurag Agarwal (RIT), Douglas Oard (Univ. Maryland, College Park), and Lee Giles (Penn State) in support of research into exploiting text-math relationships in the design of math-aware search engines, and integration with the CiteSeerX technical paper database.
- (Dec.) Congratulations to Huang Xuan, who has secured an internship with Facebook in summer 2018.
- (Nov.) Kenny Davila's work on extracting whiteboard contents from lecture videos is being presented as a poster at ICDAR 2017. Source code and data from the work are available online here.
- (Nov) The lab's work was featured in a meeting with the RIT Board of Trustees. Mahshad Madhavi and Wei Zhong represented the lab. We thank Nicholas Paulus for preparing the poster used in the session.
- (Sep) The lab welcomes Mahshad Mahdavi, a new PhD student in the Imaging Science program. Mahshad will be working on machine learning techniques for extracting and recognizing formulas in documents.
- (Sep) Recent DPRL alumnus Lakshmi Ravi (MSc, 2017) has secured a position as a Data Scientist on Amazon's Alexa Machine Learning group in Boston.
- (Aug) PhD student Thomas Choi (at INSA Rennes (France), co-advised by Prof. Zanibbi) had his paper on bootstrapping small samples for recognizing accidentals in music notation accepted for publication at GREC 2017. There are many talks on Optical Music Recognition (OMR) planned for the workshop, so we expect this to be an exciting GREC!
- (Aug) We are very happy to introduce the new members of the DPRL in Fall 2017:
Rahul Dashora (MSc student) will work on improving the lab's Hierarchical Contextual Parsing (HCP) technique for parsing mathematical notation in PDFs, images, and handwriting.
- Xuan Huan (MSc student) will work on creating a new web-based system for formula entry and search (based on the lab's 'min' system)
- Ritvik Joshi (MSc student) will work on extracting formulas from PDF documents.
- Wei Zhong (PhD student) is working on math-aware search engines.
- (Aug) Kenny Davila has accepted a post-doctoral fellowship with the well-known CEDAR/CUBS lab at the University of Buffalo. Among the lab's many other accomplishments, CEDAR/CUBS contributed to the first automated mail-sorting system for the US Postal Service.
- (July) The Tangent-S formula retrieval system has been released, and is available as open source. The changes for this version were made by Kenny Davila.
- (July) Kardo Aziz has successfully defended his MSc thesis, titled Better Text Detection through Improved k-means-based Feature Learning. In his thesis, Kardo proposes a technique called Visual Similarity Sampling (VSS), which selects training samples using the average similarity of image patches within and between text and non-text classes. He has found evidence that this can improve results over pure uniform sampling when the sample size is fixed, by exploiting representative and discriminative patch characteristics when constructing a training set.
- (July) Kenny Davila has successfully defended his PhD dissertation, titled Symbolic and Visual Retrieval of Mathematical Notation using Formula Graph Symbol Pair Matching and Structural Alignment. Kenny's dissertation was co-advised by Prof. Zanibbi and Dr. Stephanie Ludi (Univ. North Texas). Kenny created a formula search engine capable of using visual structure, semantics, or just the symbol layout of formulae. From this, he is able to use this to do cross-modal searches to locate formulae in lecture videos from rendered LaTeX formulae. He also created a system that can be used to search for formula, and then jump to a video frame where selected ink from the whiteboard is first drawn (YouTube demo).
- (June) The Alfred P. Sloan Foundation has funded a proposal by Prof. Zanibbi and C. Lee Giles (Penn State) to integrate math-aware search into the CiteSeerX platform.
- (June) Michael Condon successfully defended his MSc thesis, Applying Hierarchical Contextual Parsing with Visual Density and Geometric Features to Typeset Formula Recognition. Michael made changes to formula structure representations for punctuation and refined other techniques from Lei Hu's HCP parsing algorithm, applying these to recognizing typeset math. His parser obtains very high expression recognition rates for formulas from the infty dataset (almost 91% parsing from connected components). Source code and data used for his work will be made available in the coming weeks.
- (June) Kenny Davila's paper Whiteboard Video Summarization via Spatio-Temporal Conflict Minimization has been accepted for publication at ICDAR 2017 (the leading document analysis and recognition conference). There is a YouTube video demonstrating his new system for summarization and navigation of lectures using whiteboard content summaries As far as we know, this is the first system that supports lecture video navigation using whiteboard contents directly.
- (May) Congratulations to Lakshmi Ravi, who received 2nd place for her MSc Project Poster, titled "Parsing Handwritten Math Formulas." Lakshmi will soon be starting work in the Alexa group at Amazon in Seattle. She will be doing work in Natural Language Processing (NLP).
- (May) Prof. Zanibbi received the Golisano Computing College Outstanding Scholar Award for a strong track record of research and scholarship that is "integral to, and not separated from, all aspects of a student's educational experience at RIT." Thanks from Prof. Zanibbi to all dprl students from the last ten years who made this possible. This award primarily reflects your hard work and accomplishments - thank you.
- (Apr.) Prof. Zanibbi gave a brief presentation at the RIT GCCIS Research Showcase about the lab, dprl@10: The Document and Pattern Recognition Lab's First 10 Years. A sincere thanks and congratulations to all of the dprl students from these first 10 years!
- (Apr.) Chinmay Jain has secured a Software Engineer position involving machine learning for customer recommendations with BBVA Compass.
- (Apr.) Congratulations to Kenny Davila whose paper Layout and Semantics: Combining Representations for Math Formula Search" has been accepted for poster presentation at SIGIR 2017, the leading Information Retrieval conference.
- (Apr.) Kenny Davila will be giving a public talk at RIT about his research on math formula retrieval in documents and videos Monday, April 17th at noon in the Bamboo Rooms (Campus Center 003), as part of the Move 78 Seminar Series.
- (Apr.) The lab gives a warm welcome to Wei Zhong, who will be joining the DPRL as a PhD student in Fall 2017. Wei will be doing work in math-aware search engines. For a glance of his current work in this area, you can try his Approach0 system for searching Math StackExchange posts here.
- (Mar.) Le Duc Anh from the Nakagawa Lab in Japan visited the lab for a week in early March, to talk with us about his work in handwritten math recognition. His picture can be found under the "Members" link. Anh recently defended his PhD, and will begin working on Computer Vision and Machine Learning at a company in Tokyo next month.
- (Feb.) Michael Condon has secured a job working as a Pattern Recognition Software Engineer at Apple in Cupertino, CA. He is joining a team that worked on handwriting recognition for the Apple Watch.
- (Jan.) The web page for ICFHR 2018 is now online, including a call for papers. Please consider submitting your work to the conference, and joining us in Niagara Falls!
- (Jan.) Prof. Zanibbi will be the new Communications Officer for the IAPR Technical Committee No. 11 ('Reading Systems').
- (Dec.) Congratulations to Chinmay Jain, whose poster received 2nd place in the Best MS Poster competition! Chinmay's poster is available here.