Tracking system using texture cue based on wavelet transform

Abstract

This paper presents an approach for tracking objects whose principal discriminate characteristic is its texture. The presented system extracts texture features based on the wavelet transform and uses a fuzzy grammar classifier. The feature vector consists of 6 characteristics extracted from the wavelet detail images. The overall system was integrated on the platform developed by sony – AIBO robot. This application ensures a real time tracking approach and can be parameterized in order to be flexible in face of different types of texture.