Communications of the ACM
Cell Phones That Listen and Learn
The SoundSense system is a first step in a system that can learn user behavior on the go, says Dartmouth professor Tanzeem Choudhury.
Credit: Joseph Mehling / Dartmouth
A cell phone would be able to track the behavior of its user with SoundSense, new software developed by Dartmouth College researchers. SoundSense automatically classifies sounds as "voice," "music," or "ambient noise," but the user also can train it to recognize unfamiliar sounds. When a sound is frequently picked up via the microphone on a cell phone, SoundSense gives it a high "sound rank" and asks the user whether it is significant and if he or she wants to label the sound. In tests, the software correctly determined when the user was in a particular coffee shop, walking outside, brushing her teeth, cycling, and driving a car.
"The SoundSense system is our first step in building a system that can learn [user behavior] on the go," says project leader and Dartmouth professor Tanzeem Choudhury.
Monitoring everyday sounds via cell phones has the potential to provide people with much information on their daily activities, which could be used to improve their personal healthcare needs or time-management skills. SoundSense does not use a lot of power, sends data elsewhere for processing, stores raw audio clips, and can be told to ignore certain sounds.
View a SoundSense video on YouTube.
From Technology Review
View Full Article
Abstracts Copyright © 2009 Information Inc., Bethesda, Maryland, USA 
No entries found