Studies Of Object Tracking Algorithms In Image Sequences

This paper mainly discusss the studies of object tracking algorithms in imagesequences. Object tracking exist questions such as object sheltered from others and objectitself change in color or shape. For those questions, ombine the image scale space theory,this paper introduce the object tracking algorithm using scale invariant feature andmeanshift.The detection of image feature is quite important in image tracking process. Thispaper detailed introduce the image scale space thory and feature detection algorithm，suchas harris, SURF, SIFT. At the same time, this paper introduces the implemented method ofSIFT based on FPGA and SIFT. This paper presents a novel fast single-pass contourtracing algorithm in a binary image, The proposed algorithm is viewed as follow steps:firstly a set of contour segments of all object contours can be generated and traced in atop-down line scan fashion; then all contour segments are employed to be integrated intorespective intact contours; finally all results are converted into the chain code form as thefinal output. This algorithm can extract multiple contours of an image in one pass andnever lose any outer and inner contour of object region. It is faster on implementation.Experiments results prove those advantages. Then this paper presents a novel fast templatematching algorithm based on context prediction. The predicted regions are those windowsthat contain the current entire sub-window. Comparison skipping or comparisonterminating is executed when a low bound of distance which has been calculated betweenthe template and the window exceeds the threshold. Experimental results and theoryanalyses prove the proposed method is faster than the conventional fast template matchingmethod, strictly guaranteeing the same accuracy and up to maximal twenty times fasterthan the SSDA.For real time processing of the hige resolution image, this paper introducethe combine the low resolution and high resolution to detect and track object.Then this paper discuss the meanshift tracking algorithm, describe the meanshift sshortage and limitation, introduce the tracking algorithm based on meanshift and SIFT andthe tracking algorithm based on meanshift and SURF, this method effects enhancing therobust and stability of tracking. The result prove the method s advantage. At the same time, we have implemented experiments related of meanshift and introdueced object trackingalgorithms.