Abstract：According to the poor tracking ability adopting static feature model

an adaptive feature generating model based tracking program is present. In this program

the object is valid tracking signal

on the contrary

the background is noise constructing the likelihood maps a local SNR(Signal Noise Ratio) is computed to evaluate the tracking ability of current feature space

and the feature space with maximal SNR is selected as the optimal tracking feature space. Object tracking results based on mean shift demonstrated that the proposed method is more robust and feasible than the classical one.