Multimodal Affect Detection from Physiological and Facial Features during ITS Interaction

Facial affect analysis is perceived as one of themost complex and challenging areas for humanisation ofrobots. Several Facial Expression Recognition (FER)systems apply generic machine learning algorithms toextract facial features. This results in an erroneous classification of previously unseen data. This paper improviseson previous research done on emotion detection and im-plements techniques to leverage the potential of Convolu-tional Neural Networks (CNNs) effectively to classify aset of grayscale images of human faces into seven distinctemotion categories. We experiment with some populartransfer learning models to achieve a maximum accuracyof 98% for the seven-class classification task. To eke outfurther precision and reduce the value loss, we incorpo-rate the Squeeze and Excitation Network to theResNet-50 model, which resulted in a validation accuracyof 99.36%.

Download Full-text

The Impact of Visual Stereotypes on Judgments about Rape

Swiss Journal of Psychology ◽

10.1024/1421-0185/a000178 ◽

2016 ◽

Vol 75 (3) ◽

pp. 133-140

Author(s):

Robert Busching ◽

Johannes Lutz

Keyword(s):

Mental Representations ◽

Irrelevant Information ◽

Clear Distinction ◽

Rape Myth ◽

Facial Features ◽

Reverse Correlation ◽

Correlation Technique ◽

Specific Effects ◽

Legal Judgments ◽

The Impact

Abstract. Legally irrelevant information like facial features is used to form judgments about rape cases. Using a reverse-correlation technique, it is possible to visualize criminal stereotypes and test whether these representations influence judgments. In the first step, images of the stereotypical faces of a rapist, a thief, and a lifesaver were generated. These images showed a clear distinction between the lifesaver and the two criminal representations, but the criminal representations were rather similar. In the next step, the images were presented together with rape scenarios, and participants (N = 153) indicated the defendant’s level of liability. Participants with high rape myth acceptance scores attributed a lower level of liability to a defendant who resembled a stereotypical lifesaver. However, no specific effects of the image of the stereotypical rapist compared to the stereotypical thief were found. We discuss the findings with respect to the influence of visual stereotypes on legal judgments and the nature of these mental representations.

Download Full-text