Scope of the Special Issue

By Human-Human interaction we mean dyadic interactions between two people or, more generally, multiple people interactions that inform us of their activities, relation to one another and the cultural and social setting. Some examples are: handshake as an interaction
that can be part of an activity such as an agreement or a greeting; a dialogue between several people in the context of a working meeting; etc. This topic is very important with many impacts both from computer science or societal point of view: understanding natural
human interactions in a social context opens up to numerous and important applications, from more intelligent video indexing to smart surveillance, from ageing people monitoring to smart internet browsing, etc.

In Computer Vision, whilst single person understanding (face, emotion or action recognition, etc.) is widely addressed, social interactions understanding is quite new and challenging. Human-human interaction has attracted increasing attention in recent years due to its wide
applications and advancements in computer vision, human-computer interaction and social behaviour analysis, to mention just a few of the most relevant areas covered by this topic.

However, despite significant research progress, the automated understanding of human interactions or human activities from visual cues, or generic multimodal data, still remains a challenging topic. Most of the difficulties in revealing the true nature of collective behaviors come from the need for analyzing and understanding interactions among humans in time and space or within “in-the-wild” environments, where advanced modelling is required.

Further remarkable shortcomings arise from dealing with the considerable variability appearing in natural dataset recordings, such as the appearance of the people and the performance of their interaction, fostering the development of highly robust methods as well
as learning-based techniques trained on large datasets.

This special issue mainly focuses on last advances in theory, methods and applications aimed at providing deep insight into the meaning of human-human interactions. We solicit original contributions of experts in this multidisciplinary topic wishing to address relevant
issues ranging from fundamentals to real-world applications.

Topics include but are not limited to
● Multimodal pattern analysis
● Affective computing and interaction
● Emotion detection, analysis and recognition
● Cognitive interaction understanding
● Visual Attention models and systems
● Crowd behaviour understanding
● Human and group action recognition
● Human and group behavioural understanding
● Human-virtual agent interactions
● Collective activity analysis
● Databases for human-human interaction
● Scene understanding
● Tracking in crowded environments
● Articulated multi-person tracking
● Verbal and non-verbal communication perception
● Human communication dynamics and complex interaction networks
● Structural video analysis
● Modeling and simulation of human interactions
● Multimodal dyadic interaction
● Machine learning for human interaction analysis
● Bayesian modeling in multimodal interaction
● Affective and social signal processing
● Mobile multimodal systems
● Pervasive computing for human interaction understanding
● Human-human versus human-agent interaction
● Human interaction in games and sports