Fuzzy neural network approach to classifying dyeing defects

Classification of seven kinds of dyeing defects is proposed using image pro cessing and fuzzy neural network approaches. The defects include filling ba nd in shade, dye and carrier spots, mist, oil stain, tailing, listing, and uneven dyeing on selvage. The fuzzy neural classification system is constru cted by a fuzzy expert system with the neural network as a fuzzy inference engine. The neural network is trained to become the inference engine using sample data. This fuzzy neural network system possesses merits of both fuzz y logic and neural networks, and thus is more intelligent in handling patte rn recognition and classification problems. Region growing is adopted to di rectly detect different defect regions in an image. In all, seventy samples , ten samples for each defect, are obtained for training and testing. The r esults demonstrate that the fuzzy neural network approach can precisely cla ssify these samples by the features selected.