Multimodal Affective Behaviour Analysis for Groups

During Human-Human and Human-Computer Interaction, the affective behaviour of persons in a group can change and impact other group members. Hence, from the perspective of Human-Computer Interaction, it is important to develop automatic methods for affective behaviour analysis for groups. However, the lack of datasets and computational representation of groups for diverse group settings still needs improvement. This research addresses these challenges by building a group modelling framework for different tasks and using the model for real-world applications. These research contributions have a wide range of applications in affective computing, human-computer interaction and social-signal processing.