FEATURE TRACKING BY MULTIFRAME RELAXATION

This paper describes a novel feature tracking method. It is based on a n interframe relaxation technique. This method combines intra- and int er-frame constraints on the behaviour of acceptable contour structure. The intra-frame information is represented by both a dictionary of lo cal contour structure and a statistical model of the response of a set of directional feature detection operators. The inter-frame ingredien t represents the novel modelling component; it is encapsulated by an i mplicit model of the underlying surface structure of 3D feature points . The model is represented in terms of a series of unimodal probabilit y densities whose single parameter is the inter-frame distance. The in itial probabilities in our relaxation scheme effectively combine distr ibutions describing the statistical uncertainties in the position and feature characteristics of multiframe contours; these probabilities ar e refined in the light of the dictionary to produce consistent contour s. We present an experimental evaluation of the resulting feature dete ction method on cranial MRI data. Here the method significantly outper forms its single frame counterpart in terms of its ability to extract noise-free and smooth feature contours.