Bernard Boulay

Applying 3D human model in a posture recognition system

This paper proposes an approach to recognise human postures in video sequences, which combines a 2D approach with a 3D human model. The 3D model is a realistic articulated human model which is used to obtain reference postures to compare with test postures. Several 2D approaches using different silhouette representations are compared with each other: projections of moving pixels on the reference axis, Hu moments and skeletonisation. We are interested in a set of specific postures which are representative of typical video understanding applications. We describe results for recognition of general postures (e.g. standing) and detailed postures (e.g. standing with one arm up) in ambiguous/optimal viewpoint with good/bad segmented silhouette to show the effectiveness of our approach.

A Generic Framework for Video Understanding Applied to Group Behavior Recognition

This paper presents an approach to detect and track groups of people in video-surveillance applications, and to automatically recognize their behavior. This method keeps track of individuals moving together by maintaining a spacial and temporal group coherence. First, people are individually detected and tracked. Second, their trajectories are analyzed over a temporal window and clustered using the Mean-Shift algorithm. A coherence value describes how well a set of people can be described as a group. Furthermore, we propose a formal event description language. The group events recognition approach is successfully validated on 4 camera views from 3 datasets: an airport, a subway, a shopping center corridor and an entrance hall.

Improving person re-identification by viewpoint cues

Re-identifying people in a network of cameras requires an invariant human representation. State of the art algorithms are likely to fail in real-world scenarios due to serious perspective changes. Most of existing approaches focus on invariant and discriminative features, while ignoring the body alignment issue. In this paper we propose 3 methods for improving the performance of person re-identification. We focus on eliminating perspective distortions by using 3D scene information. Perspective changes are minimized by affine transformations of cropped images containing the target (1). Further we estimate the human pose for (2) clustering data from a video stream and (3) weighting image features. The pose is estimated using 3D scene information and motion of the target. We validated our approach on a publicly available dataset with a network of 8 cameras. The results demonstrated significant increase in the re-identification performance over the state of the art.

Probabilistic recognition of complex event

This paper describes a complex event recognition approach with probabilistic reasoning for handling uncertainty. The first advantage of the proposed approach is the flexibility of the modeling of composite events with complex temporal constraints. The second advantage is the use of probability theory providing a consistent framework for dealing with uncertain knowledge for the recognition of complex events. The experimental results show that our system can successfully improve the event recognition rate. We conclude by comparing our algorithm with the state of the art and showing how the definition of event models and the probabilistic reasoning can influence the results of the real-time event recognition.