Robust correlation filter tracking with multi-scale spatial view

摘要: Abstract With extensive applications, visual tracking has already become one of the most important research focuses in computer vision. Due to such interference as serious occlusion or severe illumination change and so on, appearance model target tends vary heavily, posing great challenges on tracking. However, a majority existing methods have difficulties detecting above under single spatial view, affecting performance method apparently. In this paper, robust correlation filter with multi-scale view (RCFMSV) is proposed which group filters different areas established. There are two models RCFMSV, detection (DMMSV), responsible for help sensitivity scales. The other on-line location (On-line LMMSV), mainly used perform collaborative by introducing pre-location adopting around reference realize more accurate method. Extensive experiments been conducted algorithm object benchmark detailed comparative analysis between state-of-the-art also made. It confirmed that RCFMSV our work competitive performance.