A new spectrum optimization method for variable structure object tracking

In order to overcome the disadvantages of soft feature method, such as object lost in the process of moving object with variable structure, a novel approach of object tracking based on spectrum optimization is put forward in this paper. Space spectrum and edge spectrum are being extracted from moving object with effective coefficient, and then the precursor impact strength is revised and improved in the neighbor space spectrum with different effective coefficient. Soft feature is predicted and updated for the purpose of precise object tracking and effective and active object prediction using fewer amount of data. Experimental results show that this new kind of spectrum optimization approach can overcome the problem of object lost in the process of object moving with variable structure, and achieve better tracking results than soft feature method as before.