Intelligent recognition of the intricate layout of colored yarn in yarn-dyed fabrics based on stepwise clustering method

Yarn-dyed fabrics (CY-Layout) play a crucial role in the aesthetic appeal and marketability of textile products. Traditional analysis methods rely on manual inspection, which is labor-intensive and prone to errors, especially in large and intricate fabrics. To address these challenges, we propose a novel stepwise (coarse-to-fine) clustering methodology that enhances both accuracy and efficiency in recognizing yarn layouts. Our approach consists of two key steps: first, coarse clustering of yarn colors within individual regions, followed by fine clustering of regional colors across the entire fabric. By utilizing a pretrained convolutional neural network for yarn segmentation, we introduce a warp-and-weft separation strategy tailored for large-scale fabric images. This method not only improves yarn color recognition but also effectively accommodates complex color patterns, providing a reliable solution for quality control in textile manufacturing. Our results demonstrate the effectiveness of this approach, marking a significant advancement in the automated detection of CY-Layout in large and intricate fabrics.