Output details

Self-organizing maps for texture classification

<24>This work is a follow-up of previous research, investigating further our intelligent machine vision system. The pattern classification problem assumes no a priori expert knowledge available (presenting closer to reality premise), hence Self-Organizing Maps are used for solving it. In the preprocessing and feature extraction stages of the classification, supervised (LDA) and unsupervised (PCA) analysis techniques are also considered for comparison purposes. The significance of this research is in incorporating unsupervised learning and proving the usability of our intelligent machine vision system for solving real world problems, widening its applicability to automated process and product quality control of practical applications.