Facial Expression Classifiers

October 28, 2013 6:00 pm
Published by

An MIT Technology Review article by David Talbot concerned facial expression recognition software, including Affectiva’s Affdex software (http://www.affectiva.com/).

The article briefly concerns the application of such technology for education. Two of these studies were from UCSD researchers who founded another facial expression recognition company, Emotient (http://www.emotient.com/). Emotient is the “leader in emotion detection and sentiment analysis based on facial expressions”, offering marketing solutions and analytic products.

One report (Littlewort et al., 2011 IEEE International Conference on Automatic Face & Gesture Recognition and Workshops) tracked the preferred viewing speed and the perceived difficulty of video lectures watched by subjects, and found that both variables were significantly correlated with Emotient’s measures of facial expression. Another study by Emotient researchers (Whitehill et al., 2014 IEEE T Affect Comput) found that student engagement during an iPad tutoring session could be classified (as “high” or “low” engagement) with similar accuracy when performed by human judges or their facial expression recognition software. Another paper, by Arroyo et al. (2009 Proc AIED), concerns the use of emotional state monitoring by intelligent tutors.

Tags: , , ,

Categorised in:

This post was written by Syngli