Spatial-Spectral Weighted and Regularized Tensor Sparse Correlation Filter for Object Tracking in Hyperspectral Videos

Hyperspectral video camera captures spatial, spectral, and temporal information of moving objects. Traditional object tracking methods developed for color videos have been applied to hyperspectral videos after compressing hundreds of spectral bands into three, which does not fully use the wealth spectral information. To address this issue, we present a tensor sparse correlation filter (CF) with a spatial-spectral weighted regularizer for object tracking. First, tensor processing is used to reduce the spectral differences in homogeneous background, thereby producing robust spectral structure features. Second, a spatial-spectral weighted regularizer is designed in the CF framework to penalize filter template by suppressing the spectral features dissimilar to the center pixel in tracking. Third, a sparse constraint term and tracking context information are incorporated to suppress unexpected peaks in the response map. Finally, a reformulated stacked histogram of oriented gradient (HOG) feature extractor and a 2-D adaptive scale search strategy are developed to further improve the tracker's feature discrimination and scale adaptation capability. Experimental results demonstrate that the proposed method achieves superior tracking performance than the traditional CF-based trackers.