FaceReader is a tool to track the user affective state while using products or software without resorting to self-report.
Real-time analysis of facial expression from a video.
FaceReader constructs a model of the face from the video and automatically evaluate several elementary facial movements (action units). Based on these movements it calculates the likeliness that each of six basic emotions (joy, anger, sadness, surprise, fear and disgust) is felt at any given time. Tests from the provider of the software indicate a success rate of up to 90% with frontal face images.
Data limited to six basic emotions.
The video has to be captured during a test with a real product, limiting the usefulness of FaceReader in the early stages of the design process.
References describing the method
Uyl, M. den, Kuilenburg, H. van & Lebert, E. (2005). FaceReader: an online facial expression recognition system. Measuring Behavior 2005, 5th International Conference on Methods and Techniques in Behavioral Research, 30 August – 2 September 2005, Wageningen, The Netherlands.
References about quality of the method
No independent test available. (?)