No abstract

Person re-identification is a challenging problem in computer vision due to large variations of appearance among different cameras. Recently, metric learning is widely used to model the transformation between cameras. However, traditional metric learning based methods only learn one metric for the whole feature space, which cannot model different kinds of appearance variations well. In this paper, we introduce bagging into metric learning, and propose a baggingbased large margin nearest neighbor (LMNN) method for person re-identification. That is, multiple LMNN predictors are generated on sub-regions of the feature space and leveraged to obtain an aggregated predictor for performance improvement. Two bagging strategies, sample-bagging and featurebagging, are proposed and compared. Extensive experiments on three benchmarks demonstrate the superiority of proposed approach over state-of-the-art methods.

The electromagnetic force generated by a pulsed magnetic field within a metal melt leads to changes in the internal temperature and flow fields of the molten metal, thus improving the solidification of the metal structure. Using the combination of a solidification test, experimental simulation and theoretical analysis, this study simulated the distribution of both electromagnetic force and the flow field in a metal melt under wide-spectrum pulse conditions, and studied the influence of a wide-spectrum pulsed magnetic field on the solidification structure of pure aluminium with a constant flow velocity. The results of this study show that the structural refinement of the solidification of pure aluminium can be different, in spite of equal flow velocity. Furthermore, this study shows that an applied time-averaged electromagnetic force causes crystal nuclei to pass through the solid–liquid interface boundary layer and promotes the growth of crystal grains. These grains flowed with the melt flow field to achieve both refinement and homogenization of the solidified structure.