Application of machine learning methodology for textile defect detection

This study investigates the use of artificial intelligence (AI) and machine learning (ML) technologies in the textile industry, particularly emphasising how they improve operational efficiency and enhance product quality. Using a comprehensive dataset obtained from textile manufacturing operations, a specially tailored convolutional neural network (CNN) model and a long-short-term memory (LSTM) model are implemented for the classification of fabric defects. After undergoing intensive training and validation, our model showed significant improvements in performance over a large number of epochs. The CNN model started with 61.15% accuracy initially and reached 92.91% accuracy after training. The validation accuracy increased from 72.44% to 92.05%. On the same dataset, the LSTM model resulted in 86.11% training accuracy and 87.80% validation accuracy. The significant improvements in accuracy highlight the power of AI and ML to not only improve classification accuracy but also boost overall operational performance by continuously learning from fresh data inputs. Moreover, this research highlights the impact of AI and ML breakthroughs on textile production as they optimise procedures, enhance efficiency, and strengthen competitive advantage. The findings demonstrate that these technologies are a substantial advancement for the textile sector, providing powerful tools to reduce faults, streamline production processes, and ultimately provide goods of superior quality. Therefore, the study promotes the wider use of AI and ML technologies in the textile manufacturing industry, emphasising their crucial role in driving future advancements and sustainable growth.