Convolutional Neural Networks for the Binary Classification of Textile Materials
Convolutional Neural Networks for the Binary Classification of Textile Materials

Summary/Abstract: This study aims to improve the binary classification of textile materials-specifically distinguishing between acceptable items and those with defects using convolutional neural networks (CNNs). A quantitative, applied approach with a pure experimental design was employed, incorporating the Mobile-D methodology. Data collection was conducted through direct observation, and two groups were compared: The experimental group (EG), used the mobile application, and the control group (CG), did not. Results, validated through the Mann–Whitney U test, showed that the EG outperformed the CG with a 35-second reduction in classification time, a 10% increase in accuracy, and a 17% improvement in the timely detection of non-conformities. In conclusion, the use of convolutional neural networks enhances the effectiveness and efficiency of binary classification in textile inspection processes.

• Details
• Contents