09.00Invited talk: Jean-Marc Odobez (Idiap Research Institute and Ecole Polytechnique Federale de Lausanne)From the recognition of Visual Focus of Attention in group conversation to the extraction of head and body pose cues in open spaces
10.00Coffee break
10:30Violent Flows: Real-Time Detection of Violent Crowd BehaviorTal Hassner, Yossi Itcher and Orit Kliper-Gross
11:00Parameterizing Interpersonal Behaviour with Laban Movement Analysis - A Bayesian ApproachKamrad Khoshhal Roudposhti, Luis Santos, Hadi and Jorge Miranda Dias
11:30A Dynamic Curvature Based Approach For Facial Activity Analysis in 3D SpaceShaun Canava, Yi Su, Xing Zhang and Lijun Yin
12.30Lunch Break
14.00 Invited talk: Elisabeth Oberzaucher (University of Vienna) Intelligent space - evolutionary considerations of human-environment interactions
15.00 Coffee break
15.30 Understanding dyadic interactions applying proxemic theory on videosurveillance trajectories Simone Calderara and Rita Cucchiara
16.00 Urban Tribes: analyzing group photos from a social perspective. Ana Murillo, Iljung Kwak, Lubomir Bourdev, David kriegman and Serge Belongie
16.30Invited talk: James M. Rehg (Georgia Institute of Technology)Behavior Imaging and the Study of Autism
17.30 Discussion
18:00 Closing

Invited Talks

Jean-Marc Odobez (Idiap Research Institute and Ecole Polytechnique Federale de Lausanne)

From the recognition of Visual Focus of Attention in group conversation to the extraction of head and body pose cues in open spaces

Non-verbal behaviors are essential cues for the understanding of people activity and in particular of their interaction with each other and with their environment. Amongst them, gaze, or its discrete version the visual focus of attention (VFOA) is of major importance. It conveys a wealth of information about a person, how she explores an environment or react to different stimuli, and plays a fundamental role in communication with functions such as establishing relationships, regulating the discourse, or exercising social control. Recognizing gaze is however a very difficult task, especially if one looks for non invasive approaches that do not interfere with normal activities.
In this talk, we will first address the VFOA recognition in group conversation such as meetings. Using head pose as primary cue, we will introduce VFOA Bayesian models and discuss important factors contributing to its recognition, like how to model the interplay and timing information between people VFOA, the structure of the spoken conversation and contextual cues related to group activity, or the need for gaze models and unsupervised adaptation. In a second step, i will present some works towards the extension of VFOA analysis in more open settings, and in particular different unsupervised and coupled adaptation approaches for joint head and body pose estimation.

James M. Rehg (Georgia Institute of Technology)

Behavior Imaging and the Study of Autism

In this talk I will describe current research efforts in Behavior Imaging, a new research field which encompasses the measurement, modeling, analysis, and visualization of social and communicative behaviors from multi-modal sensor data. Beginning in infancy, individuals acquire the social and communicative skills which are vital for a healthy and productive life, through face-to-face interactions with caregivers and peers. However, children with developmental delays face great challenges in acquiring these skills, resulting in substantial lifetime risks. Autism, for example, affects 1 in 110 children in the U.S. and can lead to substantial impairments, resulting in a lifetime cost of care of $3.2M per person. The goal of our research in Behavior Imaging is to develop computational methods that can support the fine-grained and large-scale measurement and analysis of social behaviors, with the potential to positively impact diagnosis and treatment. I will present an overview of our research efforts in Behavior Imaging, with a particular emphasis on the use of computer vision techniques. Specifically, I will describe a new approach to video analysis based on the concept of temporal causality, which leverages a novel representation of video events as multiple point processes. Our method provides a new bottom-up approach to video segmentation based on the temporal structure of video events. I will present results for retrieving and categorizing social interactions in collections of real-world video footage. I will also highlight our recent efforts in the semi-supervised recognition of objects and activities from egocentric video.

Elisabeth Oberzaucher (University of Vienna)

Intelligent space - evolutionary considerations of human-environment interactions

Urbanization brings about a number of challenges for the humans inhabiting cities. Evolutionary history has not prepared us to live in built environments almost completely lacking natural elements, and living in close proximity to anonymous masses of people. The anonymity of settings has made surveillance systems necessary, to compensate for the lack of social control. The main motivation why surveillance systems are installed, is related to terrorism, vandalism and crime. But once in place, these systems afford exploitation in a much broader context. Surveillance material could be used to help guide the maintanance of public services, through crowd monitoring and the like. But it can also be used to gain insights about the behavioral affordances of the settings. Specific environments ask for specific behaviors, while other behaviors are unusual or forbidden. This phenomenon was described in the Behavior-Settings-Theory. Through the analysis of the occurence of behaviors in space, a classification of urban areas can be undertaken that could help inspire those who design them. In behavior research, usually the quality of a certain area is described along a set of criteria. Then behavior is observed and linked to the characteristics of the location. Through a combination of computer vision methodology with ethological methods, we are now addressing this question from a different angle: What does the behavior tell us about the quality of an area - can we differentiate areas of humane design from adverse surroundings merely through behavioral patterns? This talk will give some ideas how computer vision and ethology can prove mutually inspiring and how the interaction between the two fields can lead to new insights.