Transfer of Traditional Dyeing Art Style and Innovation of Dyeing Wastewater Treatment Technology of Hainan Li Nationality Based on Deep Learning

Abstract

This study explores the fusion of traditional dyeing and weaving art of the Hainan Li nationality with modern textile design using a deep learning–based style transfer approach. The objective is to develop a reliable digital method for preserving traditional textile patterns while supporting sustainable and efficient design applications. Image data of traditional Li patterns and contemporary textile designs were systematically collected and processed to ensure cultural authenticity and visual diversity, forming a high-quality image dataset. A deep learning style transfer model based on adaptive feature transfer and hierarchical fusion (AFTH) was developed to integrate traditional dyeing and weaving patterns with modern textile design. Experimental results indicate that the proposed method demonstrates improved performance in pattern structure preservation, style representation accuracy, and overall visual quality when compared with representative existing methods. In addition, this study incorporates intelligent wastewater treatment and environmentally friendly dyeing technologies to evaluate the environmental and economic benefits of sustainable textile production. By combining machine learning–based process optimization with intelligent sewage treatment systems, resource consumption, chemical usage, and pollutant emissions are effectively reduced. The integrated results demonstrate that digital design innovation and intelligent environmental management can be jointly promoted within the textile industry.