Fabric Defect Detection Based on Regional Growing PCNN

This paper presents an adaptive image segmentation method based on a new Regional Growing Pulse Coupled Neural Network (PCNN) model for detecting fabric defects. In this method, the pixels of analyzed image are mapped on the neurons in a pulse coupled neural network. Improved PCNN model and regional growing theory are combined in the light of the requirements for fabric defect detection. And the mean and variance value of the defect-free images are introduced into this model. The validation tests on the developed algorithm were performed with fabric images from TILDA database and results showed that the proposed method is feasible and efficient for fabric defect detection.