Media Release

Email this article

Tuesday, 8 January 2008

The world's largest search engine - Google - has awarded two research grants totalling $104,000 to the University of Adelaide to improve students' problem-solving skills and also develop novel 3D techniques.

Google is committing more than $46,000 to help the University create a computer-based curriculum with a focus on problem-solving skills.

The School of Computer Science will develop the curriculum, incorporating puzzle-based learning and computational problem solving, which becomes more specialised at each level.

"Some puzzle-based learning pilot classes have recently been presented and proved very popular with students," according to Associate Professor David Munro, Acting Head of School. "These classes will be developed and refined over the next 18 months, with the first course offered in 2009."

A second year course designed to teach problem solving using computers is already being trialled with Masters students and is also expected to be offered from 2009.

"The curriculum aims to produce graduates who are better at solving problems and therefore more valuable to the workforce," says Associate Professor Munro.

"University graduates with good problem-solving skills are a rare and highly-prized resource but with Google's support we hope to address this," he says. "Obviously, any increase in this pool of talent is of potential benefit to them."

The Australian Centre for Visual Technologies will also benefit from a $58,000 Google Research Award to develop a novel technique for automatically recognising objects in video.

"The amount of video available to us all, either through the web or our own video cameras, is immense", says Centre Director Associate Professor Anton van den Hengel. "However, finding the video we're interested in can be very difficult."

The proposed technique aims to solve this problem by analysing video and looking for some specific kinds of objects, such as cars, trees, people or particular buildings. People can then search for videos that contain these objects.

"Unlike other object recognition systems, this technique uses recently developed methods from computer vision to resolve the 3D shape of objects from their 2D appearance in the video. It is believed that this will lead to more reliable and accurate object recognition than is currently possible," Associate Professor van den Hengel says.