Team receives high marks in PASCAL grand challenge
Tony Han, assistant professor in electrical and computer engineering, and his team of graduate research assistants have been perfecting their object detection software for over three years, but 2009 marks the first year they submitted their work to the international PASCAL grand challenge in object detection for feedback.
The challenge, which began in 2005, invites researchers from universities and institutes around the world to submit their software to a competition to detect objects in 20 different categories. Han’s research team ranked third overall, following the University of Chicago and Toyota Institutes joint team which tied for first and the University of Oxford who placed second. Twelve groups submitted detection results to be evaluated by organizers from various universities and companies in England and Switzerland.
“This competition is a great way to receive feedback from competitors and see how our work compares to other algorithms,” Han said.
The Mizzou team placed first in the chair and potted plant categories, and second in human, car, horse and bottle detection. They plan to compete in the competition annually because participants can build on previously submitted work.
“The teams that placed before us have attended for many years. If we work harder, we can continue to do better with more resources,” Han said.
Object detection is a field with many potential applications. Its capabilities may change video surveillance and Internet image searching. Han said leading Internet and software companies are working on searching images by content instead of tagged text. Researchers are focusing on object detection as a future tool for a diverse array of applications.