nl-uiuc AT lists.cs.illinois.edu
Subject: Natural language research announcements
List archive
- From: "McHenry, Kenton Guadron" <mchenry AT illinois.edu>
- To: "vision AT cs.uiuc.edu" <vision AT cs.uiuc.edu>, "nl-uiuc AT cs.uiuc.edu" <nl-uiuc AT cs.uiuc.edu>, "aivr AT cs.uiuc.edu" <aivr AT cs.uiuc.edu>, "aiis AT cs.uiuc.edu" <aiis AT cs.uiuc.edu>, "aistudents AT cs.uiuc.edu" <aistudents AT cs.uiuc.edu>
- Cc: "mchenry AT uiuc.edu" <mchenry AT uiuc.edu>
- Subject: [nl-uiuc] R. Manmtha, TODAY @ 3pm
- Date: Mon, 13 May 2013 18:41:27 +0000
- Accept-language: en-US
- List-archive: <http://lists.cs.uiuc.edu/pipermail/nl-uiuc/>
- List-id: Natural language research announcements <nl-uiuc.cs.uiuc.edu>
Digging into Data Sponsored Talk - R. Manmatha
Title: Searching Handwritten Manuscripts and Mining Scanned Books
When: Monday, May 13th at 3:00 pm
Where: NCSA 1040
Abstract: In recent years a lot of effort has gone in to scanning manuscripts and printed material and less in to the question of what we should do with this scanned material. For handwritten material even recognition is a challenge. I will discuss research on searching handwritten archives such as the manuscripts of George Washington. This will include work on word spotting (finding word images in handwritten material).
The second part of the talk will focus on mining scanned printed books. The focus so far has largely been on viewing these as single items for searching or reading. However, by viewing books as a collection one can try to find relationships between books. In this talk I discuss how one can find partial duplicates, translations and OCR errors in such collections efficiently in large collections of printed and scanned books. This is based on representing each book as a sequence of unique words and then aligning these sequences using a longest common subsequence algorithm.
Bio: R. Manmatha is a research associate professor in the Dept. of Computer Science at the University of Massachusetts, Amherst. His research is on retrieving images/videos and scanned handwritten documents and printed documents. He has worked on the automatic annotation and retrieval of images/videos and is currently involved in a project on event detection in videos. He proposed the idea of word spotting for handwritten documents (using word image matching to search handwritten documents). He and his students built the first automatic demonstration system for retrieving historical handwritten documents (a portion of George Washington’s handwritten documents). In addition he is involved in a project to datamine collections of scanned book archives. He was a co-founder of SnapTell, a mobile image search company (acquired by A9/Amazon) and is a consultant to A9/Amazon. He spent a summer as a visiting research scientist at Google working on their book project. He is an associate editor for IEEE Trans. PAMI and a program chair for ACM Intl. Conf. for Multimedia Retrieval (ICMR), 2014 and the 14th Intl. Conf. on Frontiers in Handwriting Recognition (ICFHR), 2014.
Title: Searching Handwritten Manuscripts and Mining Scanned Books
When: Monday, May 13th at 3:00 pm
Where: NCSA 1040
Abstract: In recent years a lot of effort has gone in to scanning manuscripts and printed material and less in to the question of what we should do with this scanned material. For handwritten material even recognition is a challenge. I will discuss research on searching handwritten archives such as the manuscripts of George Washington. This will include work on word spotting (finding word images in handwritten material).
The second part of the talk will focus on mining scanned printed books. The focus so far has largely been on viewing these as single items for searching or reading. However, by viewing books as a collection one can try to find relationships between books. In this talk I discuss how one can find partial duplicates, translations and OCR errors in such collections efficiently in large collections of printed and scanned books. This is based on representing each book as a sequence of unique words and then aligning these sequences using a longest common subsequence algorithm.
Bio: R. Manmatha is a research associate professor in the Dept. of Computer Science at the University of Massachusetts, Amherst. His research is on retrieving images/videos and scanned handwritten documents and printed documents. He has worked on the automatic annotation and retrieval of images/videos and is currently involved in a project on event detection in videos. He proposed the idea of word spotting for handwritten documents (using word image matching to search handwritten documents). He and his students built the first automatic demonstration system for retrieving historical handwritten documents (a portion of George Washington’s handwritten documents). In addition he is involved in a project to datamine collections of scanned book archives. He was a co-founder of SnapTell, a mobile image search company (acquired by A9/Amazon) and is a consultant to A9/Amazon. He spent a summer as a visiting research scientist at Google working on their book project. He is an associate editor for IEEE Trans. PAMI and a program chair for ACM Intl. Conf. for Multimedia Retrieval (ICMR), 2014 and the 14th Intl. Conf. on Frontiers in Handwriting Recognition (ICFHR), 2014.
Kenton McHenry, Ph.D.
Research Scientist, Adjunct Assistant Professor of Computer Science
Image and Spatial Data Analysis Division (http://isda.ncsa.illinois.edu)
National Center for Supercomputing Applications
Research Scientist, Adjunct Assistant Professor of Computer Science
Image and Spatial Data Analysis Division (http://isda.ncsa.illinois.edu)
National Center for Supercomputing Applications
- [nl-uiuc] R. Manmtha, TODAY @ 3pm, McHenry, Kenton Guadron, 05/13/2013
Archive powered by MHonArc 2.6.16.