Entity

Time filter

Source Type


Hung H.,University of Amsterdam | Odobez J.-M.,Idiap Research Institute | Gavrila D.,University of Amsterdam | Gavrila D.,Daimler Research and Development
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2011

In the past years, efforts in surveillance and open space analysis have focused on traditional computer vision problems like scene modeling or object detection and tracking. Research on human behavior recognition have tended to work on predefined simple activities such as running, jumping or left luggage, and single-person trajectory analysis. The goal of the workshop is to bring together experts and researchers from different fields to share their experience and expertise about the opportunities on the development of tools for automated social analysis in open and public spaces. Humans exhibit a rich range of behaviors, from their interaction with the environment such as how groups of people occupy the space or how they manipulate or use objects within it, to the way they communicate with each other. Such behaviors can be captured from multiple sensors. Automatically interpreting interactive behavior provides a richer foundation for ambient intelligent environments. © 2011 Springer-Verlag.


Hofmann M.,University of Amsterdam | Gavrila D.M.,University of Amsterdam | Gavrila D.M.,Daimler Research and Development
Computer Vision and Image Understanding | Year: 2011

We present a novel approach for 3D human body shape model adaptation to a sequence of multi-view images, given an initial shape model and initial pose sequence. In a first step, the most informative frames are determined by optimization of an objective function that maximizes a shape-texture likelihood function and a pose diversity criterion (i.e. the model surface area that lies close to the occluding contours), in the selected frames. Thereafter, a batch-mode optimization is performed of the underlying shape- and pose-parameters, by means of an objective function that includes both contour and texture cues over the selected multi-view frames. Using above approach, we implement automatic pose and shape estimation using a three-step procedure: first, we recover initial poses over a sequence using an initial (generic) body model. Both model and poses then serve as input to the above mentioned adaptation process. Finally, a more accurate pose recovery is obtained by means of the adapted model. We demonstrate the effectiveness of our frame selection, model adaptation and integrated pose and shape recovery procedure in experiments using both challenging outdoor data and the HumanEva data set. © 2011 Elsevier Inc. All rights reserved.


Hung H.,University of Amsterdam | Odobez J.-M.,Idiap Research Institute | Gavrila D.,University of Amsterdam | Gavrila D.,Daimler Research and Development
Communications in Computer and Information Science | Year: 2012

Human behaviour, and in particular interactive or social behaviour in public spaces is rich and highly varying. It is a great source of information about people's attitudes towards strangers, friends and family, and how they chose to navigate through and familiarise themselves with the urban environment. This paper provides a summary of our workshop on interactive human behaviour analysis in open or public spaces, and in particular, highlighting the future applications, challenges, and goals that such an area of research should have. We discuss the outcomes of the discussions, talks and presentations of the day. © 2012 Springer-Verlag.


Liem M.C.,University of Amsterdam | Gavrila D.M.,University of Amsterdam | Gavrila D.M.,Daimler Research and Development
Computer Vision and Image Understanding | Year: 2014

We present a system to track the positions of multiple persons in a scene from overlapping cameras. The distinguishing aspect of our method is a novel, two-step approach that jointly estimates person position and track assignment. The proposed approach keeps solving the assignment problem tractable, while taking into account how different assignments influence feature measurement. In a hypothesis generation stage, the similarity between a person at a particular position and an active track is based on a subset of cues (appearance, motion) that are guaranteed observable in the camera views. This allows for efficient computation of the K-best joint estimates for person position and track assignment under an approximation of the likelihood function. In a subsequent hypothesis verification stage, the known person positions associated with these K-best solutions are used to define a larger set of actually visible cues, which enables a re-ranking of the found assignments using the full likelihood function. We demonstrate that our system outperforms the state-of-the-art on four challenging multi-person datasets (indoor and outdoor), involving 3-5 overlapping cameras and up to 23 persons simultaneously. Two of these datasets are novel: we make the associated images and annotations public to facilitate benchmarking. © 2014 Elsevier Inc. All rights reserved.


Keller C.G.,University of Heidelberg | Keller C.G.,Daimler Research and Development | Gavrila D.M.,Daimler Research and Development | Gavrila D.M.,University of Amsterdam
IEEE Transactions on Intelligent Transportation Systems | Year: 2014

Future vehicle systems for active pedestrian safety will not only require a high recognition performance but also an accurate analysis of the developing traffic situation. In this paper, we present a study on pedestrian path prediction and action classification at short subsecond time intervals. We consider four representative approaches: two novel approaches (based on Gaussian process dynamical models and probabilistic hierarchical trajectory matching) that use augmented features derived from dense optical flow and two approaches as baseline that use positional information only (a Kalman filter and its extension to interacting multiple models). In experiments using stereo vision data obtained from a vehicle, we investigate the accuracy of path prediction and action classification at various time horizons, the effect of various errors (image localization, vehicle egomotion estimation), and the benefit of the proposed approaches. The scenario of interest is that of a crossing pedestrian, who might stop or continue walking at the road curbside. Results indicate similar performance of the four approaches on walking motion, with near-linear dynamics. During stopping, however, the two newly proposed approaches, with nonlinear and/or higher order models and augmented motion features, achieve a more accurate position prediction of 10-50 cm at a time horizon of 0-0.77 s around the stopping event. © 2013 IEEE.

Discover hidden collaborations