Anti-occlusion tracking algorithm based on LSSVM prediction and Kalman-MeanShift

An embedded Least Squares Support Vector Machines prediction anti-occlusion Kalman-MeanShift tracking algorithm was proposed to solve the poor tracking ability problem in occlusions. The proposed technique employed MeanShift algorithm iterations to derive the object candidate which was the most similar to a given object model, then used Kalman filter to estimate the real states of the object. When the object was occluded seriously, the observation could not be used for updating by Kalman filter, LSSVM was employed to trend prediction of object moving and produce predictive a new object position, searched object by MeanShift algorithm nearby predicted position. The results showed that the moving objects occlusion tracked by the new technique was more robust than traditional methods. In addition, the results seemed to suggest that the new method be able to provide efficiency and accuracy advantages in computer version.