Scientific publications
Read about the research that supports the FaceReader Ecosystem
Over the past 20+ years, our facial coding platform and its embedded technologies have been the subject as well as the preferred instrument for numerous accredited scientific studies. Below we present a comprehensive overview of the literature that has emerged from these studies, highlighting and validating the cutting-edge technology of FaceReader Online.
2014
117 citations
Facial expressions and autonomous nervous system responses elicited by tasting different juices
Danner, Haindl, Joechl, Duerrschmid
The study aimed to understand reactions elicited by tasting different juices by examining self-reported liking, autonomic nervous system responses, and both implicit and explicit facial expressions. Eighty-one participants tasted banana, grapefruit, mixed vegetable, orange, and sauerkraut juices. Measurements included skin conductance level , skin temperature , heart rate , pulse volume amplitude , and facial expressions. Results showed significant differences in SCL and PVA responses, as well as in the intensity of several facial expressions across the juices. A moderate correlation was found between these physiological responses and self-reported liking, allowing differentiation between liked, disliked, and neutral samples. Notably, in the implicit approach, participants displayed minimal positive emotions for liked samples, whereas in the explicit approach, they exhibited strong positive emotions. Negative emotions were more pronounced for disliked samples in both approaches. The findings suggest that self-reported liking, ANS responses, and facial expressions provide distinct information about taste experiences.
The article was accepted on 4 June 2014.
2014
209 citations
Predicting advertising effectiveness by facial expressions in response to amusing persuasive stimuli
P. Lewinski, M. L. Fransen, E. S. H. Tan
We present a psychophysiological study of facial expressions of happiness produced by advertisements using the FaceReader system for automatic analysis of facial expressions of basic emotions . FaceReader scores were associated with self-reports of the advertisement’s effectiveness. Building on work describing the role of emotions in marketing research, we examined the relationship between the patterns of the FEBE and the perceived amusement of the advertisements, attitude toward the advertisement and attitude toward the brand . Differences were observed between FEH scores in response to high-, medium-, and low-amusing video advertisements . Positive correlations were found between FEH and AAD and FEH and AB in high- and medium- but not in low-AVAs. As hypothesized, other basic emotions did not predict advertisement amusement or advertisements’ effectiveness. FaceReader enabled a detailed analysis of more than 120,000 frames of video-recordings contributing to an identification of global patterns of facial reactions to amusing persuasive stimuli. For amusing commercials, context-specific FEH features were found to be the major indicators of advertisement effectiveness. The study used video-recordings of participants in their natural environments obtained through a crowd-sourcing platform. The naturalistic design of the study strengthened its ecological validity and demonstrated the robustness of the software algorithms even under austere conditions. Our findings provide first evidence for the applicability of FaceReader methodology in the basic consumer science research.
2014
194 citations
Make a face! Implicit and explicit measurement of facial expressions elicited by orange juices using face reading technology
Danner, Sidorkina, Joechl, Duerrschmid
This study examined consumers’ facial reactions elicited by the flavor of orange juice products using both implicit and explicit measurement approaches. The objectives were to assess whether facial expressions measured with Noldus FaceReader technology could accurately differentiate between various orange juice samples, to explore the relationship between implicit and explicit facial reactions to these juices, and to determine if these facial reactions could explain introspective liking ratings on hedonic scales. Participants tasted different orange juices, including diluted syrup, nectar, 100% juices, and not-from-concentrate juice. In the implicit approach, participants were unaware of being recorded, and their automatic facial reactions during and after tasting were analyzed. In the explicit approach, participants intentionally displayed facial expressions after tasting, which were also recorded and analyzed. Both measurement methods revealed significant differences in facial expressions elicited by the different samples. Explicit measurements correlated well with liking ratings, particularly with “happy” and “disgusted” expressions indicating liked and disliked samples, respectively. In the implicit measurements, “neutral,” “angry,” and “disgusted” expressions correlated with liking ratings, though discrimination between samples was better in the explicit condition. The study concluded that facial expression analysis using FaceReader technology is a viable method for differentiating between orange juice samples and can provide additional insights beyond traditional acceptance tests. However, further research is needed to evaluate this technology in more complex testing scenarios and real-life environments.
2014
66 citations
Measuring Software Screen Complexity: Relating Eye Tracking, Emotional Valence, and Subjective Ratings
Goldberg
2014
132 citations
Risk Aversion and Emotions
Nguyen & Noussair
We consider the relationship between emotions and decision-making under risk. Specifically, we examine the emotional correlates of risk-averse decisions. In our experiment, individuals’ facial expressions are monitored with face reading software, as they are presented with risky lotteries. We then correlate these facial expressions with subsequent decisions in risky choice tasks. We find that a more positive emotional state is positively correlated with greater risk taking. The strength of a number of emotions, including fear, happiness, anger and surprise, is positively correlated with risk aversion.
2014
417 citations
Automated facial coding: Validation of basic emotions and FACS AUs in FaceReader
P. Lewinski, T. M, Den Uyl, C. Butler
In this study, the authors validated the automated facial coding software, FaceReader , using two publicly available datasets of human expressions of basic emotions. They reported matching scores for the recognition of facial expressions and the Facial Action Coding System index of agreement. While previous research in 2005 reported matching scores of 89% for FaceReader, it utilized an older version without FACS classifiers. In this study, the authors tested the latest version and found that FaceReader recognized 88% of the target emotional labels in the Warsaw Set of Emotional Facial Expression Pictures and the Amsterdam Dynamic Facial Expression Set . The software achieved a FACS index of agreement of 0.69 on average across both datasets. These results are significant when compared to human performance rates for both basic emotion recognition and FACS coding. Human emotion recognition for the two datasets was 85%, indicating that FaceReader performs comparably to humans in recognizing emotions. To receive FACS certification, a human coder must reach an agreement of 0.70 with the master coding of the final test. Although FaceReader did not attain this score, action units 1, 2, 4, 5, 6, 9, 12, 15, and 25 may be used with high accuracy. The authors conclude that FaceReader has proven to be a reliable indicator of basic emotions over the past decade and has the potential to become similarly robust with FACS.