Occlusion-aware Visual Tracker using Spatial Structural Information and Dominant Features

To overcome the problem of occlusion in visual tracking, this paper proposes an occlusion-aware tracking algorithm. The proposed algorithm divides object into discrete image patches according to pixel distribution by means clustering. avoid drifting tracker false targets, extracts dominant features, such as color histogram or oriented gradient orientation, from these patches, and uses them cues for tracking. enhance robustness tracker, employs implicit spatial structure between another cue tracking; Afterwards, incorporates components particle filter framework, which results a robust precise tracker. Experimental on sequences with different resolutions show that outperforms comparison algorithms handling