Fabric Texture Reconstruction Via Multilayer Dictionary Learning

Fabric texture reconstruction is a critical task in the textile industry, with applications in computer-aided design, quality control, and production optimization. However, traditional methods suffer from limitations such as poor accuracy, slow convergence, and low robustness. To address these issues, this study proposes a novel approach that combines multilayer dictionary learning-based encoding and variational autoencoder (VAE) for fabric texture reconstruction. The proposed method utilizes multilayer dictionary learning-based encoding to extract deeper fabric texture features and generates new fabric images with similar texture features using VAE. Experiments on a publicly available dataset of fabric images demonstrate the superiority of the proposed method in terms of accuracy, robustness, and efficiency compared to state-of-the-art methods. The proposed method has significant implications for the textile industry.