The self-reported liking ratings of orange juices correlate significantly with such basic emotions as anger, disgust, and happiness. Three scientists from University of Natural Resources and Life Sciences (Department of Food Science and Technology), Vienna, Austria have used FaceReader™ software  to test people’s reactions to tasty and not-so-tasty orange juices . The program – automatically analyzing facial expressions – provides instantaneous, precise reading of the participant’s emotional state. The reactions can be discretely recorded and give insight to participants’ more “sincere” opinions about the tasted products.
The researchers carried out two related experiments - the participants were either aware (explicit condition) or unaware (implicit condition) of the video recording. In the explicit condition, the least liked orange juice (as measured by self-report) – orange juice syrup - elicited the strongest “disgusted” facial expressions as in comparison to other types of juice. Additionally, in this condition, the most liked orange juice – freshly squeezed – elicited the happiest expressions. Surprising results came from the implicit condition – where people were unaware of the video recording. The least liked juice still caused most angry and disgusted expressions but the most liked juice did not elicit significantly more happy facial expressions. The scientists concluded that it is feasible to test consumer products with this new method – computerized facial coding system.
These experiments show that expressing emotions serve social and communication functions for humans. If people are told that they are watched then they express emotions that are more positive and suppress some of the negative reactions. The social context shapes the way we react. Scientists researching consumer behavior are interested in automatic solutions that interpret human emotional reactions to commercial products. They try to detect “real” and implicit behavior as opposed to socially-mediated reactions. The results of this study were presented during the “5th European Conference on Sensory and Consumer Research” and contribute to the rapidly developing field of market research - neuromarketing . The autonomous spontaneous and posed expressions can be smoothly and reliably analyzed using state-of-the-art artificial intelligence software such as FaceReader, provided by behavioral research tools and solution providers like Noldus Information Technology .
FaceReader has recently been released in its 5.0 version  providing updated algorithms that detect (a) six basic emotions plus neutral; (b) emotional valence; (c) gaze direction and head orientation. The software is trained with a set of more than 10.000 photos of facial expressions of emotions; its classification is computed through a 3-layer neural network and constantly tracks the face with almost 500 unique virtually-imposed key-points on the face. So far, it has been used to analyze as diverse issues as - the presidential debate between Romney and Obama in 2012 ; - astronauts’ facial behavior during Mars-500 Isolation Experiment  or - empathetic feedback in e-learning applications .
By Peter Lewinski
Marie Curie Research Fellow
VicarVision, Amsterdam, NL
FaceReader methodology white paper
Request the FREE FaceReader methodology note to learn more about facial expression analysis theory.
- Learn what FaceReader is and how it works
- Learn how the calibration works
- Get insight in quality of analysis and output
 Danner, L.; Sidorkina, L.; Duerrschmid, K. (2012). Make a face! Implicit and explicit measurement of facial reactions elicited by model foods using FaceReading Technology. Proceedings of the 5th European Conference on Sensory and Consumer Research, P9-15.