ODU Vision Lab Director Awarded NSF Grant for Computer Pattern Recognition Research
September 23, 2013
Khan Iftekharuddin, electrical engineering professor and director of the Old Dominion University Vision Lab, has been awarded a $250,000 grant from the National Science Foundation to study how to make groundbreaking computer vision pattern recognition algorithms work faster, better and more simply.
ODU's Machine Vision and Computational Intelligence Laboratory is one of the country's leading research centers for developing state-of-the-art applications for pattern recognition, signal and image processing, and machine learning.
The field of machine vision pattern recognition has many possible uses, from spotting security threats through facial recognition, to tumor detection in MRI, to rapid assessment of performance in things like vehicle engines.
The Vision Lab and Iftekharuddin have been tasked through the NSF grant to continue to improve pattern recognition technology, giving it more applications and making it more broadly available.
"We're seeking better, more flexible architecture, both software and hardware," Iftekharuddin said. "The existing techniques for face recognition, for example, are usually handcrafted and very complex. We're trying to see if we can simplify the process, and make it more effective at the same time."
The $250,000 award Iftekharuddin received is ODU's share of a three-year grant totaling more than $400,000. Iftekharuddin and Vision Lab researchers are partnering with electrical engineer Tarek Taha of the University of Dayton, an expert in hardware implementation of complex algorithms.
The grant is an effort to take a "step back" and look at the science underpinning pattern recognition, Iftekharuddin said. "We're trying to understand how visual processing in the brain is done and replicate some of its functionality in a generalizable pattern recognizer."
The increasingly complex needs for the pattern recognition technology - such as scanning every face as people go through a security-screening checkpoint at an airport - require computer architectures that can be replicated more simply than they are now. "For us to be able to process this amount of data, we need to be able to do it in a more simple platform," Iftekharuddin said.
Ultimately, the researchers hope to develop a real-time feedback component to the technology, known as a High Performance Cellular Simultaneous Recurrent Network, aimed at developing the kind of real-time feedback known to be crucial to the power of the human brain, Iftekharuddin said.
Iftekharuddin's expertise in the area of signal processing and computer vision has also been recognized by the U.S. Air Force. For eight weeks this summer, he was at Wright Patterson Air Force Base in Ohio, participating in the U.S. Air Force Summer Faculty Fellowship Program. The program, sponsored by the Air Force Office of Scientific Research, enlists math, science and engineering faculty at U.S. colleges and universities to help solve research challenges for the Air Force.
The project Iftekharuddin worked on involved inserting his research expertise in signal processing and machine learning into the USAF human performance directorate, with the goal of analyzing human movements in a new way.
"Using our machine learning technique, we're able to extract biomechanical information to do things such as gender detection and human detection and tracking," Iftekharuddin said.
Vision Lab master's student Jeff Flora also participated in the eight-week program.