A more stable and robust approach is proposed for visual tracking relying on Maximally Stable Extremal Regions (MSERs), sparse random projection and online Least Squares SVM classifier (LS-SVM) learning.
With the fusion of MSERs and sparse random projection, the stable adaptive object appearance is modeled to adapt the variation of appearance.
An online closed-form LS-SVM is employed to quickly and robustly predict the target object location.