Since 2000, the National Science Foundation has bestowed Early Career Development Awards on 20 of RPI’s young faculty members. The CAREER Award is the NSF’s most prestigious award for faculty in their early professional careers. The award includes financial support for research over the course of five years, the amount based on a budget request submitted by the honoree, with the minimum total award amount set at $400,000.
Dr. Vera Kettnaker has received a CAREER Award for her proposal to develop stochastic models that determine the daily routine of someone monitored with video cameras. Stochastic models are probabilistic models that show how something changes over time. Kettnaker plans on developing algorithms that learn the behavior of an individual and can detect when the individual has deviated from normal behavior.
She is currently working on solutions to problems such as low light and shadows. Eventually, after developing the algorithms and better ways to separate the person from the rest of the video image than current methods, she will need sample data. To create sample data, she will partition her lab into rooms and have a person perform various actions. She will not have to record data over a long time span, as the amount of time that passes can be simulated.
One application of this research is in the home of a senior citizen. A camera would be placed in each room to monitor the person’s behavior. The system would then be able to learn that its subject watches television from 8 to 10 pm and then usually goes to sleep. Then if the person is still watching television at 1 am, the system would ask the person if he is okay, and if there is no response, a signal could be sent to someone for help.
Kettnaker likes being involved in research that could be beneficial to society. She said she gets “a sense of challenge to see how far you can push things.”
Analyzing video using computers is nothing new to Kettnaker. The field of computer vision is an integration of computer science and her interest of the eye. She went through a computer science undergraduate program at Kaiserslautern in Germany and then received her Ph.D from Cornell. Her Ph.D research involved analyzing traffic patterns using video.
