A novel segmentation method using improved PCNN for fabric defect image

Fabric defect image segmentation is not only a key stage on real-time visual detection but also a very difficult problem. A novel method for fabric defect image segmentation using improved Pulse Couple Neural Networks (PCNN) is proposed. According to different gray intensity between the field of defects and the field of no defects, PCNN neuron cell is fired to implement segmentation. The iteration index of PCNN is controlled by the minimum cross entropy. And, segmentation evaluation criteria is also presented in this paper. The validity tests on the developed algorithms have been performed with some fabric defect images. Experimental results show that the proposed method can segment common fabric defect quickly and correctly. It is more effective than other methods using performance evaluation.

[1]  Robert M. Hodgson,et al.  Texture Measures for Carpet Wear Assessment , 1988, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  Wu Cheng,et al.  Researches on Evaluation Methods for Image Segmentation , 2000 .

[3]  Reinhard Eckhorn,et al.  Feature Linking via Synchronization among Distributed Assemblies: Simulations of Results from Cat Visual Cortex , 1990, Neural Computation.

[4]  Qi Fei.  PCNN IMPLEMENTATION AND APPLICATIONS IN IMAGE PROCESSING , 2005 .

[5]  Randall R. Bresee,et al.  Fabric Defect Detection and Classification Using Image Analysis , 1995 .

[6]  Harsh Potlapalli,et al.  Texture characterization and defect detection using adaptive wavelets , 1996 .

[7]  K. Srinivasan,et al.  FDAS: A Knowledge-based Framework for Analysis of Defects in Woven Textile Structures , 1992 .

[8]  Xue Jing.  lmage Segmentation Algorithms Based on Cross Entropy and Fuzzy Divergence , 1999 .

[9]  Harsh Potlapalli,et al.  Image Analysis of Mispicks in Woven Fabric , 1995 .

[10]  C. H. Li,et al.  An iterative algorithm for minimum cross entropy thresholding , 1998, Pattern Recognit. Lett..

[11]  Richard W. Conners,et al.  A Theoretical Comparison of Texture Algorithms , 1980, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[12]  J. Y. Dong,et al.  Image segmentation based on modified pulse-coupled neural networks , 2007 .

[13]  Bugao Xu,et al.  A Simplified pulse-coupled neural network for adaptive segmentation of fabric defects , 2009, Machine Vision and Applications.

[14]  John L. Johnson,et al.  PCNN models and applications , 1999, IEEE Trans. Neural Networks.