Communications of the ACM
Math Tutoring Software Reads Visual Clues
University of Massachusetts in Amherst (UMA) researchers are testing a math tutoring program that uses a built-in camera to identify visual cues from the student user about what in the lesson is working. "It looks at things like frowns, the relative position of the eye and mouth, at smiles, at the direction of the look and other gestures," says UMA computer scientist Ivon M. Arroyo. The camera locates key features on the face and body and calculates how they move in relation to one another, which provides insight into the emotions of the user.
The program was designed to help girls in particular, who often show diminishing interest in math as they reach high school. Using the visual cues, the program's learning companion character will give hints or encouraging words to the user. The researchers say the software is currently 70 percent to 80 percent accurate in determining a user's emotion.
From The Republican (MA)
View Full Article
Abstracts Copyright © 2010 Information Inc., Bethesda, Maryland, USA 
No entries found