Object Tracking Method Based on Semi Supervised Extreme Learning

A moving object tracking method based on semi supervised extreme learning machine is proposed, which transforms the moving object tracking into a dichotomous problem, to solve the problem that the moving object tracking scene is occluded and the moving object is deformed and can easily lead to the loss of the moving object. Firstly, the semi-supervised limit-based learning machine network training set is established by sampling the object area. Secondly, the discriminant model is set up by learning training to locate the optimal moving object. At the same time, the similarity proportion and update threshold of the moving object are introduced to judge whether the semi supervised extreme learning machine network model, with updated reasoning can improve the prediction accuracy of the model. Compared with the other six target tracking methods, the proposed algorithm can greatly improve the center location error (CLE) in the simulation experiment, and can accurately locate the moving object and the speed can meet the real-time requirements.