基于SVM与MeanShift的非刚性目标跟踪框架.pdfVIP

第31 卷 第9 期 系统工程与电子技术 Vol.31 No. 9 2009 年9 月 Systems Engineering and Electronics Sep. 2009 :(2009) 0922266205 SVM Mean2Shift , ( 东北电力大学自动化工程学院, 吉林 吉林 132012) : , , ( support vector machine, SVM) (Mean2Shift)。, 。SVM 。 ,( confidence map) ,Mean2Shift , , 10% 。 SVM ,, 。,, 。 : ; ; ; ; : TP 391 : A Nonrigid target tracking framework based on SVM and Mean2Shift HOU Yi2min, LUN Xiang2min ( School of Automation Engineering , N orthea st Dianli Univ . , J ilin 1320 12 , China Abstract: To solve the problem oftracking nonrigid target s for image sequence in dynamic scene, a tracking framework based on support vector machine( SVM) and Mean2Shift is proposed. The rectangle in which the tracked target lies in is selected in the intial image, and the pixels in the given region around the target are taken as scene data. Then both the target and the scene pixel data are applied to train a two2value SVM classifier. The obtained SVM classifier is employed to classify the pixel data lying in the same region of the next image, thus the confidence map including two pixel values is got. The Mean2Shift algorithm is performed to get theposition of the current target within the area of the confidence map. Similarly, the new SVM classifiers are trained for tracking the images to be followed until the whole image sequence tracking is complet ed. The experiment result s show that the proposed method is suitable for tracking nonrigid target s and active scenes, and has the real2time characteristic. Keywords: image processing; target tracking; support vector machine; mean2shift; confidence map 0