Mean shift

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: IEEE CVPR 2000 Real-Time Tracking of Non-Rigid Objects using Mean Shift Dorin Comaniciu Visvanathan Ramesh Peter Meer Imaging & Visualization Department Siemens Corporate Research 755 College Road East, Princeton, NJ 08540 Electrical & Computer Engineering Department Rutgers University 94 Brett Road, Piscataway, NJ 08855 Abstract 2 Mean Shift Analysis A new method for real-time tracking of non-rigid objects seen from a moving camera is proposed. The central computational module is based on the mean shift iterations and nds the most probable target position in the current frame. The dissimilarity between the target model (its color distribution) and the target candidates is expressed by a metric derived from the Bhattacharyya coe cient. The theoretical analysis of the approach shows that it relates to the Bayesian framework while providing a practical, fast and e cient solution. The capability of the tracker to handle in real-time partial occlusions, signi cant clutter, and target scale variations, is demonstrated for several image sequences. We de ne next the sample mean shift, introduce the iterative mean shift procedure, and present a new theorem showing the convergence for kernels with convex and monotonic pro les. For applications of the mean shift property in low level vision ( ltering, segmentation) see 6]. 2.1 Sample Mean Shift Given a set fxi gi=1:::n of n points in the ddimensional space Rd, the multivariate kernel density estimate with kernel K (x) and window radius (bandwidth) h, computed in the point x is given by n X f^(x) = 1 d K x ; xi : (1) 1 Introduction nh i=1 The e cient tracking of visual features in complex environments is a challenging task for the vision community. Real-time applications such as surveillance and monitoring 10], perceptual user interfaces 4], smart rooms 16, 28], and video compression 12] all require the ability to track moving objects. The computational complexity of the tracker is critical for most applications, only a small percentage of a system resources being allocated for tracking, while the rest is assigned to preprocessing stages or to high-level tasks such as recognition, trajectory interpretation, and reasoning 24]. This...
View Full Document

Ask a homework question - tutors are online