Video Search: New Challenges in the Pervasive Digital Video Era

The explosion of multimedia digital content and the development of technologies that go beyond traditional broadcast and TV have rendered access to such content important for all end-users of these technologies. While originally developed for providing access to multimedia digital libraries, video search technologies assume now a more demanding role. In this paper, we attempt to shed light onto this new role of video search technologies, looking at the rapid developments in the related market, the lessons learned from state of art video search prototypes developed mainly in the digital libraries context and the new technological challenges that have risen. We focus on one of the latter, i.e., the development of cross-media decision mechanisms, drawing examples from REVEAL THIS, an FP6 project on the retrieval of video and language for the home user. We argue, that efficient video search holds a key to the usability of the new ”pervasive digital video” technologies and that it should involve cross-media decision mechanisms.

An Empirical Study of Non-Rigid Surface Feature Matching of Human from 3D Video

This paper presents an empirical study of affine invariant feature detectors to perform matching on video sequences of people with non-rigid surface deformation. Recent advances in feature detection and wide baseline matching have focused on static scenes. Video frames of human movement capture highly non-rigid deformation such as loose hair, cloth creases, skin stretching and free flowing clothing. This study evaluates the performance of six widely used feature detectors for sparse temporal correspondence on single view and multiple view video sequences. Quantitative evaluation is performed of both the number of features detected and their temporal matching against and without ground truth correspondence. Recall-accuracy analysis of feature matching is reported for temporal correspondence on single view and multiple view sequences of people with variation in clothing and movement. This analysis identifies that existing feature detection and matching algorithms are unreliable for fast movement with common clothing.

Frequency of PRRS live vaccine virus (European and North American genotype) in vaccinated and non-vaccinated pigs submitted for respiratory tract diagnostics in North-Western Germany

The frequency of PRRSV corresponding to live vaccines and wild-type was determined in 902 pigs from North-Western Germany submitted for post-mortem examination. Overall, 18.5% of the samples were positive for the EU wild-type virus. EU genotype vaccine virus was detected in 1.3% and the NA genotype vaccine virus in 8.9% of all samples. The detection of the EU vaccine was significantly higher in pigs vaccinated with the corresponding vaccine (OR=9.4). Pigs vaccinated with NA genotype had significantly higher detection chances for the corresponding vaccine virus when compared to non-vaccinated animals (OR=3.34) animals, however, NA vaccine was also frequently detected in non-vaccinated pigs. Concluding, the dynamics of NA genotype vaccine and EU wild-type virus corresponds with studies on PRRSV spread in endemically infected herds. The potential of spontaneous spread of the NA genotype vaccine should be considered in the planning of eradication programs.

Real-time Human Motion Capture with Simple Marker Sets and Monocular Video

In this paper we present a hybrid method to track human motions in real-time. With simplified marker sets and monocular video input, the strength of both marker-based and marker-free motion capturing are utilized: A cumbersome marker calibration is avoided while the robustness of the marker-free tracking is enhanced by referencing the tracked marker positions. An improved inverse kinematics solver is employed for real-time pose estimation. A computer-visionbased approach is applied to refine the pose estimation and reduce the ambiguity of the inverse kinematics solutions. We use this hybrid method to capture typical table tennis upper body movements in a real-time virtual reality application.