Emotion recognition in the wild

For inferring the affective state of a person from data, captured in real-world conditions, methods which can perform emotion analysis ‘in the wild’ are required. Here, the term ‘in the wild’ signifies different environments/scenes and background noise, illumination conditions, head pose and occlusion. Automatic emotion recognition has made a significant progress in last two decades. However, such developed frameworks have been strictly employed to data collected in controlled laboratory settings with frontal faces, perfect illumination and posed expressions. On the contrary, images and videos on the WWW have been captured in different, unconstrained environments and this poses a big challenge to automatic facial emotion recognition methods. This special issue addresses the problem of emotion recognition in challenging conditions and is based on the recent series of Emotion recognition in the Wild (EmotiW) challenge.