Human tracking system for automatic video surveillance with particle filters

The algorithm presented is a very efficient and robust object tracking algorithm based on particle filter. The aim is to deal with noisy and bad resolution video surveillance cameras. The main feature used in this method is multi-candidate object detection results based on a background subtraction algorithm combined with color and interaction features. This algorithm only needs a small number of particles to be accurate. Experimental results demonstrate the efficiency of the algorithm for single and multiple object tracking.