Two Oregon teens took home the nation’s top high school science prize for developing technology that could help children with autism identify emotion in human speech.

Akash Krishnan and Matthew Fernandez, both high school juniors from Portland, Ore., will share a $100,000 college scholarship for winning the team portion of the 2010 Siemens Competition in Math, Science & Technology last weekend. Their project dubbed “The Recognition of Emotion in Human Speech” relies on an emotional speech database with 18,215 files to distinguish among five emotions — anger, positive, neutral, emphatic and rest.

“The team built a computer algorithm that allows us to listen to an auditory signal from a human, analyze it and assess the emotional state of the speaker,” said competition judge Gert Lanckriet of the department of electrical and computer engineering at the University of California, San Diego. “It can help identify, for example, if the speaker is angry, sad, bored, anxious or happy. They came up with a strong creative idea and brought it from theory into practice.”

Krishnan and Fernandez were able to achieve 60 percent accuracy with their technology. Ultimately the teens’ research could help those with autism who struggle to interpret emotions. The technology could also have applications for computer games and lie detection. Previous research in this field had only resulted in 41 percent accuracy.

The teens said they were inspired by the movie “I, Robot” which features a robot that’s able to sense stress.

Krishnan and Fernandez — who have already garnered interest for their work from the U.S. government — are continuing to improve their technology and have filed for a patent. They say they are working to apply the emotion-recognition software to create a device specifically for children with autism.