Search

In this paper, we propose a novel hybrid shape descriptor that combines the 2D view and transform-based features. The 2D view features are extracted from six orthogonal directions of a 3D model by using the Scale Invariant Feature Transform (SIFT) method. In order to capture the six orthogonal 2D views, Continuous Principal Component Analysis (CPCA) is used to implement pose alignment. Meanwhile, the eigenspace is computed and stored to reduce the 2D view feature vector dimension. Then, the radial integral transform and the spherical integration transform are used to extract transform-based features. The similarity between the query model and models in the database is computed by using the weighted sum of 2D view and transform-based feature similarity. Experimental results show that the proposed hybrid shape descriptor can achieve satisfactory retrieval performance for both the articulated models in the McGill Shape Benchmark and the rigid models in the Princeton Shape Benchmark. [ABSTRACT FROM AUTHOR]

The 3D human body tracking from videos in unconstrained scenes is a challenging problem and has widespread applications. In this paper, we introduce a novel framework that incorporates the graph-based human limbs detection into the articulated Bayesian tracking. The 3D human body model with a hierarchical tree structure can describe human΄s movement by setting relevant parameters. Particle filter, which is the optimal Bayesian estimation, is used to predict the state of the 3D human pose. In order to compute the likelihood of particles, the pictorial structure model is introduced to detect the human body limbs from monocular uncalibrated images. Then the detected articulated body limbs are matched with each particle using shape contexts. Thus the 3D pose is recovered using a weighted sum of matching costs of all particles. Experimental results show our algorithm can accurately track the walking poses on very long video sequences. [ABSTRACT FROM AUTHOR]

Biometric technology is critical to the rapidly growing suite of civilian applications. In this paper, we offer a broader scope of biometrics by a novel concept, the biometrics grid. The architecture of biometrics grid aims to overcome/resolve three main problems of existing biometric technology using the grid computing: 1) the need for cooperation in different fields, 2) duplicated works and 3) the consideration of application demands in different levels. The proposed architecture of biometrics grid provides a Web portal based virtual environment as well as features like biometric applications meeting QoS goals. Experimental results are given in order to show the feasibility to deploy the idea of grid computing in biometric applications. [ABSTRACT FROM AUTHOR]