A design process for texture segmentation

Abstract

Classification technique has become the necessary tool to detect products which has texture patterns in machine vision. The design used for the process of texture discrimination is critical to the performance of the outcome. In this study we propose a design process model which can systematically maintain and improve on the performance of texture segmentation in terms of speed and segmentation accuracy. The overall accuracy and time tested for the synthetic textures segmentation has shown significant improvement throughout the parameters tuning of the chosen methods.Keywords -Pattern recognition, image segmentation, system design, texture analysis, feature extraction, clustering algorithm 978-1-4244-4870-8/09/$26.00 ©2009 IEEE Proceedings of the 2009 IEEE IEEM