Occluded targets tracking using improved GM-PHD tracker

The closed-form solution for Probability Hypothesis Density (PHD) filter is Gaussian Mixture PHD (GM-PHD) filter which is applied for multiple-target tracking in noisy observation set. The main drawback of GM-PHD filter is its failure in keeping trajectories of targets. To solve the problem of GM-PHD filter, we propose Improved GM-PHD (IGM-PHD) tracker which is a simple and efficient approach to detect occlusion time and correctly keep the trajectories of occluded targets using a weighted history of targets distance. Experimental results obtained on real and simulated data sets show that the IGM-PHD tracker outperforms other powerful multi-target trackers such as GM-PHD Tracker.