Human perception and biosignal-based identification of posed and spontaneous smiles

摘要: Facial expressions are behavioural cues that represent an affective state. Because of this, they unobtrusive alternative to self-report. The perceptual identification facial can be performed automatically with technological assistance. Once the have been identified, interpretation is usually left a field expert. However, do not always felt affect; also communication tool. Therefore, expression measurements prone same biases as Hence, automatic measurement human affect should make inferences on nature instead describing movements only. We present two experiments designed assess whether such automated inferential judgment could advantageous. In particular, we investigated differences between posed and spontaneous smiles. aim first experiment was elicit both types expressions. contrast other studies, temporal dynamics elicited were constrained by eliciting instruction. Electromyography (EMG) used discriminate them. Spontaneous smiles found differ from in magnitude, onset time, offset speed independently producer's ethnicity. Agreement type EMG-based detection reached 94% accuracy. Finally, agreement video coders showed although labels fairly good, worsens labels. A second confirmed layperson's accuracy regards distinguishing poor. would beneficial terms assessments further research this topic.