We propose an algorithm addressing the problem of template drift in object tracking.Template drift is a common phenomenon in which the target gradually shifts away from the template and the template is gradually occupied by background objects due to template update.In order to solve this problem

this paper derives expressions that accurately evaluate the influence of template drift on the template-updating filter

so that the filter achieves a quasi-optimal balance between reducing template drift and keeping track of target appearance variations in a spatiotemporally adaptive manner.Experimental results based on a wide range of real-world video sequences containing diverse degrees of target appearance variations confirm the effectiveness of our algorithm.