Fuzzy-based person tracking in real-time

Abstract

The implementation of a system for person tracking is often a compromise between sophisticated tracking algorithms and simplicity (to achieve the desired frame rate on standard PCs). We propose a fuzzy based rule system for tracking people using simple segmentation algorithms. The system allows integration of heuristics in the tracking algorithm, adjusted by means of weighting functions. It also combines several segmented features in an intuitive way. The result is a stable system that can not be achieved using the single extracted features alone.