Funded for 28 consecutive years by the National Science Foundation (NSF), the University of Central Florida’s Research Experience for Undergraduates (REU) in Computer Vision is the nation’s longest-running REU site. NSF recently extended the site to 2018 with a $400,000 grant.
Since 1987, about 270 undergraduates from 65 institutions have performed computer vision research here, including participant Elizabeth Cole, an electrical engineering major at the University of Massachusetts Amherst. Cole (pictured above) is exploring how computer vision is used in medical imaging; specifically, how it can automatically identify fat cells in CT scans.
“I applied to UCF because of the topic and to see if graduate school is right for me,” she said.
Malcolm Collins-Sibley, a sophomore at Northeastern University in Boston studying computer engineering, added computer science as his second major because of his UCF experience. He applied to the REU program intrigued by the topic and UCF’s round-the-clock mentoring.
“The subject matter was very interesting to me — the idea that a computer can interpret the world visually in the same way as any person can,” he said. “I was able to apply without any specific experience in computer vision,” he added. He was one of 10 participants selected from about 200 applicants.
Sibley partnered with UCF doctoral student Shervin Ardeshir on a project involving geo-semantic segmentation, describing it as “breaking an image from an urban setting into ‘meaningful’ sections.” He praised his lessons at UCF in code writing using MATLAB and weekly progress reports. “They helped me with my skills in presenting very technical work in a way that everyday people can understand.”
Ultimately, a paper co-authored with Ardeshir was approved for publication in Computer Vision and Pattern Recognition, the premier annual event in the field. The 2015 conference is ongoing this week in Boston.
“The research is intensive,” said UCF Site Director Mubarak Shah, who for the past decade has been among the top five cited experts worldwide in computer vision. “Some students complain that it’s too much, but we say, ‘that’s the life.’ We believe in theory and practice. We are interested in mathematical modeling and analysis of difficult vision problems and developing algorithms, while building real systems for demonstrating those solutions in real-life situations.”
Dr. Shah is UCF’s trustee chair professor of Computer Science. Niels da Vitoria Lobo, Ph.D., associate professor, Computer Science, assists with the REU site. Research topics include video surveillance, new computer vision algorithms, visual geo-localization, object detection, tracking, activity and event recognition, facial recognition, automated scrutiny of extensive video footage, and more.
NSF’s goal for REUs is to attract talented undergraduates into research careers. UCF’s full-time, 12-week summer REU in Computer Vision has become a national model that features mentoring from faculty and graduate students, and a streamlined short course and numerous practical challenges to test academic ability and communication skills.
Young researchers typically advance to graduate school and become faculty, like UNC Chapel Hill Assistant Professor Tamara L. Berg, who participated at UCF’s REU site in 2000-01. Others start companies, such as University of South Florida graduate Maha Sallam, who was at UCF’s REU in 1989-90. Her Tampa-based company VuEssense, Inc., employs 100 people and generates $20 million in revenue.
— UCF —
Pictured Above: Sarfaraz Hussein, UCF computer science doctoral student, mentors Elizabeth Cole, visiting undergraduate from the University of Massachusetts at Amherst.
Explore UCF’s REU in Computer Vision (includes a sample funding proposal).
Read about Dr. Mubarak Shah and the future of computer vision in UCF’s Pegasus magazine.
Kimberly J. Lewis contributed to this story.