September 25, 2001

NC State Researcher’s Work with Electronic Shape Recognition and Analysis Finds Practical Applications

The interpretive powers of the human brain are amazing — we spend our lives assigning meaning to the images we see every day.  Recognition of shapes and interpretation of their significance is a capability reserved for animal intelligence; artificial intelligence (AI) refers to the capability of a machine to learn such functions.  If a computer could be programmed to learn shape recognition and interpretation tasks, the useful applications would be many.  Examples could include a computer program that can tell you whether the vehicle at a certain remote location is a tank or a pickup truck, or a computer sophisticated enough to examine a piece of lumber for quality control grading.  These possibilities may not be far away, according to Dr. Hamid Krim, associate professor of electrical and computer engineering at NC State University.

Krim has been working on projects funded by the U.S. Navy, the U.S. Air Force and the National Science Foundation (NSF) to refine computer imaging capabilities.  This creative use of AI involves establishing intelligent algorithms programmed to characterize the shapes of objects and interpret the resulting data. 

For example, these imaging techniques could help workers in the North Carolina lumber industry grade wood more efficiently.  Currently, a quality control worker must examine each board as it comes by on a conveyer belt — a tedious and high-pressure job.  The wood must be screened for knots and other variations that reduce the quality of a board, then assigned a grade.  Imaging could allow this grading to be done automatically by a computer, freeing workers for more productive tasks.  Krim is setting up a demonstration bench with a conveyer belt in his laboratory to see how well the computer compares with human grading.  If successful, this technique would benefit the North Carolina lumber industry, an important part of the state economy.

-- rudd --

Media Contacts:  Dr. Hamid Krim, 919/513-2270, ahk@eos.ncsu.edu, Linda E. Rudd, 919/515-3848, linda_rudd@ncsu.edu