Object Tracking Algorithm Based On Continuous Correlation Filtering

Object tracking technology is a research hotspot in the field of computer vision and image processing.It is widely used in military and civilian fields.In the direction of precision guidance,reconnaissance and early warning,intelligent transportation and medical imaging,object tracking technology provides tremendous support.Object tracking is the process of estimating the state of a particular object as the scene in the video sequence changes.The object is the tracking target.Due to the complexity of the actual tracking scenario,object tracking still faces many challenges.This paper improves on the most popular correlation filtering methods in the object tracking field based on continuous correlation filtering.Through observation,it is found that there are many similar samples in the training set,which will have a certain negative impact on the training of the relevant filtering,and the sparse representation method has the effect of dimension reduction.Therefore,this paper reduces the dimension of the training set by sparse representation,removes redundant samples,achieves the effect of streamlining the training set,and improves the accuracy of the tracking algorithm to some extent.We also found that each feature channel does not contribute exactly the same to the object tracking.Therefore,based on the continuous correlation filtering algorithm,each feature channel is weighted to enhance the influence of important features and weaken the influence of the secondary features.An efficient model update strategy is adopted to make the performance of the algorithm more excellent.Finally,we compare the two improved strategies and find that the advantage of the training set dimension reduction method based on sparse representation is more obvious.