Enhancing Fabric Defect Detection Using Efficient Pyramid Split Attention in a Lightweight YOLOv5 Framework

Fabric defect detection is a fundamental quality control process in textile manufacturing, yet achieving accurate and reliable automated inspection remains difficult because of complex background textures, subtle defect patterns, and substantial variation in defect scale and shape. Although deep learning–based detectors have improved inspection performance, many lightweight models still suffer from limited feature discrimination, particularly in real-time industrial environments where computational efficiency is critical. To address this limitation, this study proposes an enhanced fabric defect detection framework by integrating an Efficient Pyramid Split Attention (EPSA) mechanism into a YOLOv5-based convolutional network. The EPSA module is designed to adaptively recalibrate multi-scale feature responses, enabling the network to emphasize defect-relevant information more effectively while preserving inference efficiency. A quantitative experimental design was employed using a labeled fabric defect image dataset, and the proposed model was evaluated through comparative and ablation analyses against baseline and alternative attention-based configurations. Experimental results indicate that the EPSA-enhanced model achieves superior detection performance in terms of mean Average Precision while maintaining real-time processing capability. The improvement is especially evident for small, low-contrast, and irregular defects embedded in repetitive fabric textures. These findings confirm that pyramid-based attention can substantially improve feature representation without imposing significant computational overhead. The proposed approach offers a practical and efficient solution for automated textile inspection and provides a useful foundation for future research on lightweight attention modeling for industrial vision systems.