Inlier Clustering based on the Residuals of Random Hypotheses.

We present an approach for motion clustering based on a novel observation that a signature for putative pixel correspondences can be generated by collecting their residuals with respect to model hypotheses drawn randomly from the data. Inliers of the same motion cluster should have strongly correlated residuals, which are low when a hypothesis is consistent with the data in the cluster and high otherwise. After evaluating a number of hypotheses, members of the same cluster can be identified based on these correlations.

Ultra-long magnetization needle induced by focusing azimuthally polarized beams with a spherical mirror.

Based on extended Richards-Wolf theory for axisymmetric surfaces and the inverse Faraday effect, we propose the generation of a purely longitudinal magnetization needle by focusing Gaussian annular azimuthally polarized beams with a spherical mirror. The needle obtained has a longitudinal length varying hundreds to thousands of wavelengths while keeping the lateral size under 0.4λ, and the corresponding aspect ratio can easily reach more than 2000.