Ask the Scientist:
Featured Scientist: Marian Bartlett
CERT, or Computer Expression Recognition Toolbox, is a software package for automatically detecting facial expressions. CERT was developed at UC San Diego, originating from a collaboration between Ekman and Sejnowski (Bartlett et al., 1996, 1999, 2006; Donato et al., 1999; Littlewort et al., 2006). The software performs real-time expression detection via web-cam input, and can automatically detect and code such expressions as anger, disgust, fear, joy, sadness, surprise, contempt, head pose, and 30 facial action units from the Facial Action Coding System. (Note: CERT is being commercialized by Machine Perception Technologies, at http://mpt4u.com; it offers free academic licenses).
Automatic facial expression recognition tools such as CERT are beginning to bring about paradigmatic shifts in a number of fields by making facial expression more accessible as a behavioral measure. They enable new experiments into facial behavior that were previously infeasible. Automated systems can be applied to much larger quantities of video data than human coding. Statistical pattern recognition on this large quantity of data can reveal emergent behavioral patterns that previously would have required hundreds of coding hours by human experts, and would be unattainable by the non-expert. Moreover, automated facial expression analysis is enabling investigations into facial expression dynamics that were previously intractable by human coding because of the time required to code intensity changes.
Examples in Research
CERT, in particular, is being used to enhance facial expression recognition and production in children with autism (http://tdlc.ucsd.edu/research/highlights/rh-facial-expression-2010.html). The new project integrates the computer-based intervention known as Let’s Face It! (LFI!) with CERT. Let’s Face It! (LFI!) is a training program designed to improve the face skills of children with autism in a variety of face processing domains. The integration of the LFI! program and CERT allows the child to receive immediate feedback on their facial expression productions. These exercises developed by TDLC researchers engage the production system, and may aid children with autism in learning nonverbal behaviors essential for social functioning.
While the accuracy of automated facial expression recognition systems is still below that of human experts, automated systems already enable new experiments into facial behavior that were previously infeasible. Automated systems can be applied to much larger quantities of video data than human coding. Statistical pattern recognition on this large quantity
For more information about automated facial expression recognition systems, please see: Automated facial expression measurement: Recent applications to basic research in human behavior, learning, and education (written by Marian Stewart Bartlett and Jacob Whitehill, TDLC members: http://tdlc.ucsd.edu/research/publications/Bartlett_Whitehill_Automated%20Facial_Expression_2010.pdf)