Adaptive Bandwidth Mean Shift for Video Image Segmentation
WU Ti-hua · Acta Simulata Systematica Sinica · 2005
A new nonparametric estimation method called adaptive bandwidth mean shift (ABMS) algorithm is applied to realizing image smoothing and segmentation,which removes noises and trivial texture and preserves the structural features of the image objects.The intersection of confidence intervals (ICI) rule and the sample point density estimator in ABMS algorithm were used to determine the smoothing bandwidth and obtain the mean shift vector of each pixel in the video image.Along the direction of this vector,image was smoothed and the pixels were clustered to several blocks.At the same time,the motion features of the image blocks were fused and the object was extracted from the video sequence.The experiments and simulations verify the efficiency of the proposed method.