Tracking algorithm based on color correlation histogram using particle filter

A novel tracking algorithm based on the color correlation histogram using particle filter was proposed in this paper. By using the color correlation histogram as target feature, spatial information is introduced into the target representation, which gives a reliable likelihood description of the observation and the prediction for tracking the targets accurately. The capability of the algorithm to tolerate appearance changes like orientation changes, small scale changes, partial occlusions and background scene changes is demonstrated using some real video sequences. Experimental evidence shows that the color correlation histogram is more effective than the traditional color histogram for objects tracking.