Constructing 3D models from video is one of the most important problems in computer vision. We propose a novel 3D face modeling approach from monocular video captured by a conventional camera. An algorithm is proposed to estimate the head pose by comparing the edges of video frame, and the contours extracted from a generic face model. A generic 3D face model is assumed to be the initial estimate of the true 3D model. The generic face model is adapted to the actual 3D face model by global and local deformations. An affine model is used for global deformation. The 3D model is locally deformed by computing the v 4.1 Subject A: Video frames used for 3D model reconstruction. .. . 4.2 (a): Generic mesh after global affine deformation (b): Optimal perturbations applied to each control point to obtain the final adapted model. Green: Perturbation along X-axis (width), Red: Perturbation along Y-axis (height), Blue: Perturbation along Zaxis (depth).

No abstract

There is an increasing demand for detecting complex events from heterogeneous sensor networks to support the Global War On Terror (GWOT). This paper presents a framework addressing several vital aspects of a multimodal complex event detection system for wide area surveillance. The paper discusses the key modules for an effective system including complex event specification, cross-camera tracking for wide area surveillance, simple and complex event detection from video and network data, and high-level complex event inference from multimodal data using Markov Logic Networks. A novel testing framework is also discussed for data collection, system debugging and performance evaluation.

No abstract