Automated tracking of subviral particles in fluorescence image sequences opens new opportunities for the research of medicines to fight Ebola and Marburg viruses. Based on tracking algorithms the motion and distribution of subviral particles in image sequences can be automatically analyzed. For this, an accurate tracking is mandatory. A method to generate a weighted mean of a two-sided Kalman filtering on interrupted tracks to recover lost data is presented. The method is extensively tested on one real track with a simulated interruption. The results show clear advantages of this novel adapted method over one-sided Kalman filtering and unweighted mean calculation.