Spatio-temporal clustering of probabilistic region trajectories

We propose a novel model for the spatio-temporal clustering of trajectories based on motion, which applies to challenging street-view video sequences of pedestrians captured by a mobile camera. A key contribution of our work is the introduction of novel probabilistic region trajectories, motivated by the non-repeatability of segmentation of frames in a video sequence. Hierarchical image segments are obtained by using a state-of-the-art hierarchical segmentation algorithm, and connected from adjacent frames in a directed acyclic graph. The region trajectories and measures of confidence are extracted from this graph using a dynamic programming-based optimisation. Our second main contribution is a Bayesian framework with a twofold goal: to learn the optimal, in a maximum likelihood sense, Random Forests classifier of motion patterns based on video features, and construct a unique graph from region trajectories of different frames, lengths and hierarchical levels. Finally, we demonstrate the use of Isomap for effective spatio-temporal clustering of the region trajectories of pedestrians. We support our claims with experimental results on new and existing challenging video sequences. © 2011 IEEE.

Highly efficient spin transport in epitaxial graphene on SiC

Spin information processing is a possible new paradigm for post-CMOS (complementary metal-oxide semiconductor) electronics and efficient spin propagation over long distances is fundamental to this vision. However, despite several decades of intense research, a suitable platform is still wanting. We report here on highly efficient spin transport in two-terminal polarizer/analyser devices based on high-mobility epitaxial graphene grown on silicon carbide. Taking advantage of high-impedance injecting/detecting tunnel junctions, we show spin transport efficiencies up to 75%, spin signals in the mega-ohm range and spin diffusion lengths exceeding 100μm. This enables spintronics in complex structures: devices and network architectures relying on spin information processing, well beyond present spintronics applications, can now be foreseen. © 2012 Macmillan Publishers Limited. All rights reserved.