Particle swarm optimized particle filter and its application in visual tracking

Particle swarm optimization is proposed to optimize the particle filter in order to Travel out the well-known particle impoverishment and dependency problem. Through particle swarm optimization, particle samples are moved to neighbor higher likelihood areas. In this way, it can obtain more approximate to the true posterior probability density function. Meanwhile, the number of particle sample reducing significantly, make it the better choose to apply the real-time estimation and tracking problem.