Ching Suen given award for pattern recognition work

By Shira Katz

Ching Y. Suen has spent 30 years improving machine recognition of handwriting in documents. His work was recognized Aug. 30 in Seoul, when he was presented with the 2005 Outstanding Achievement Award at the International Conference on Document Analysis and Recognition.

“It’s a very challenging field,” Suen admitted later. The challenge is to achieve a level of speed and accuracy in recognition that can be used by outside industries.

He feels his breakthrough was to combine the human recognition process with machine recognition techniques to produce more accurate results, and combining different types of recognizers together to recognize with a superior performance.

Cheques

“For example, most kinds of handwriting on cheques can be read accurately. Machines can now read all kinds of handwriting, but not writing that is sloppy. Humans can’t read sloppy handwriting, either.

Psychologists have done many studies on writing habits, styles, speed and legibility. Researchers have also considered how humans distinguish similar characters from one another — differences that might be misrecognized by computers.

Suen said that identification is much more accurate than it once was, and these improvements have significant applications in banking.

There are also implications for medicine (e.g., detecting abnormalities in x-rays), in police work (e.g., license plate detection), among others.

At the conference in Korea, he gave the keynote address, which focused on how humans recognize handwritten characters with distinctive features, and how this knowledge may be transferred to the computer and extracted mathematically with an ever- higher degree of speed and accuracy.

In addition, a total of 15 papers were presented by researchers associated with Concordia: Professors Tien Bui and Robert Sabourin, members of CENPARMI; postdoctoral fellow Richard Zannibi, doctoral student Chunlei He, and doctoral graduates Jie Zhou and Jianxiong Dong.

Suen is the Director of the Centre for Pattern Recognition and Machine Intelligence, and holds the Concordia Research Chair in Artificial Intelligence and Pattern Recognition. More information about CENPARMI is at http://www.cenparmi.concordia.ca/.

Suen founded ICDAR as well as several other conference series. He has been appointed deputy editor of Pattern Recognition, the most popular journal in the field.