Handwriting Recognition Workshop Sponsored by UB to Focus on Forensic Uses, Including Anti-Terrorism

BUFFALO, N.Y. -- Investigators working to identify the source of the anthrax-containing envelopes that terrorized the American public last fall concluded they were hand-addressed by the same person. Messages penned in Urdu that allegedly belonged to Daniel Pearl's kidnappers were introduced by prosecutors during their trial.

Expanding the forensic use of computer processing of handwriting to solve such high-profile criminal cases will be among the topics discussed at the Eighth International Workshop on Frontiers in Handwriting Recognition, sponsored by the University at Buffalo, to be held Aug. 6-8 in Niagara-on-the-Lake, Ontario.

The workshop will be attended by more than 100 computer scientists working in industrial, academic and government labs in 20 countries.

Interest in the use of computer processing of handwriting, once a field attractive to a small group of academic computer scientists and their counterparts in industry, has been on the increase for several reasons, said Sargur Srihari, Ph.D., SUNY Distinguished Professor in the UB Department of Computer Science and Engineering and director of UB's Center of Excellence in Document Analysis and Recognition (CEDAR).

"In addition to continued expansion in the market for handheld computing devices, such as personal digital assistants, and improvements in offline recognition of handwriting for postal and financial applications, forensic uses for the computer processing of handwriting, especially in the wake of Sept. 11, are stimulating new areas of research," he said.

The workshop will feature a panel discussion on Aug. 8 on computer processing of handwriting in forensic document examination. Panelists will include Srihari and representatives from the U.S. Secret Service and Federal Bureau of Investigations.

The workshop also will include sessions on using computational techniques to detect forgery; to recognize characters in foreign languages, such as Chinese and Arabic, and to determine the identity of a writer of a document, including work conducted in an independent analysis by Srihari that determined that the addresses on the envelopes containing anthrax likely were written by the same person.

"This is the first time this workshop is turning its attention to the computer processing of handwriting not just for recognition, to read what has been written, but also for analysis, to determine who wrote it," said Srihari.

"In law enforcement in general that work still is being done by human analysts, but we now are beginning to use computers to do it. Teams at CEDAR and at other institutions that will be represented at the workshop are beginning to prove that automated analysis techniques can be quite successful."

Srihari and his colleagues at CEDAR published a paper this month in the Journal of Forensic Sciences providing the first scientific proof that handwriting is unique to individuals.

Other work being presented at the workshop by CEDAR scientists will include preliminary reports on new techniques for recognizing medical words so that handwritten forms filled out by paramedics could immediately be fed into a database, revealing potential patterns of public health emergencies that could be related to a terrorist attack or an epidemic.

More information on the workshop, including titles of papers, is available at http://www.cedar.buffalo.edu/IWFHR8/.

CEDAR is the largest research center in the world devoted to developing new technologies that can recognize and read handwriting. In the U.S., it is the only center in a university where researchers in artificial intelligence apply pattern-recognition techniques to the problem of reading handwriting.

Over the past decade, CEDAR has worked with the U.S. Postal Service developing and refining the software now in use in postal distribution centers across the nation that allow up to 70 percent of the handwritten addresses on envelopes to be read by sorting machines.

Conference sponsors include UB, CEDAR, Hitachi Ltd., Microsoft Corp., Motorola, Inc., Siemens AG.