Textile Printing Pattern Image Segmentation Based on Algorithm of MRF

摘要

In order to improve the accuracy of textile image segmentation, this paper presents a method of applying the MRF algorithm to the fabric image segmentation. Segmentation results could be obtained by three steps: First step is pretreating the original images. Second step is establishing the MRF model by way of the ICM algorithm. The last step is combining the bayesian theory and MAP. Several printing images are applied and perfect results are received in this experiment.