ML-Based Approach for Enhancing Sewing Operator Training in the Apparel Industry Using Hand Movement Recognition

In the apparel industry, the training of sewing operators plays a pivotal role in ensuring the production of top-quality garments. This research presents a novel approach to improve training methods through the implementation of a real-time hand movement recognition system. This system is designed to identify omissions and incorrect hand actions, providing immediate alerts based on Garment Standard Data (GSD) for prompt corrective actions. Leveraging advanced computer vision techniques and a graph neural network (GNN), the framework achieves an impressive 85.7% accuracy in monitoring and analyzing sewing operators' hand movements. By comparing detected movements with predefined standards, the system identifies deviations and offers instant feedback to operators. Experimental results underscore the system's effectiveness in pinpointing incorrect steps and hand movements, highlighting the potential of GNNs to elevate training in the apparel industry. The developed system significantly enhances sewing operator efficiency and productivity, ultimately leading to the production of higher-quality garments.