Tracking features in image sequences using discrete Morse functions

Abstract

The goal of this contribution is to present an application of discrete Morse theory to tracking features in image sequences. The proposed algorithm can be used for tracking moving figures in a filmed scene, for tracking moving particles, as well as for detecting canals in a CT scan of the head, or similar features in other types of data. The underlying idea which is used is the parametric discrete Morse theory presented in [13], where an algorithm for constructing the bifurcation diagram of a discrete family of discrete Morse functions was given. The original algorithm is improved here for the specific purpose of tracking features in images and other types of data, in order to produce more realistic results and eliminate irregularities which appear as a result of noise and excess details in the data.