Design and application of garment art patterns based on visual sample generation
Year: 2026
Type: article
Abstract: Clothing pattern design plays an important role in clothing aesthetics and market sales. However, with the advancement of technology and changes in consumer aesthetic needs, traditional hand-designed methods have been unable to meet the personalized and diversified pattern needs of contemporary young people. Therefore, it is necessary to explore and develop more efficient and innovative pattern design methods to meet the modern clothing market's demand for personalized, creative, and diversified design. To address this challenge, a f-WGRAN model based on zero-sample image classification is proposed. This model generates image features through an improved generative adversarial network and trains a classifier using the generated features to achieve clothing image classification and feature extraction. Meanwhile, the study adopts the L-system generation method for pattern drawing, and obtains the desired artistic pattern outline by selecting appropriate axioms, achieving personalized and beautiful clothing pattern design. The results showed that the f-WGRAN model achieved the highest classification accuracy of 69.8% and 68.7% on the AWA and FLO datasets, respectively, in the traditional zero-sample learning case. In the generalized zero-sample academic image classification experiments, the classification accuracy of the visible category samples of the f-WGRAN model was as high as 63.2%, 75.3%, 58.0%, and 34.8% on the AWA, FLO, CUB, and SUN datasets, respectively. The mean accuracy reached 60.5%, 67.3%, and 50.6% on three datasets. The combination of f-WGRAN model and L-system method has strong feasibility in clothing pattern design. This method can automatically generate multiple styles of patterns according to different needs, achieving a faster design process and fully meeting the needs of young consumers for personalized and innovative patterns. This technology has promoted the intelligent and personalized development in the field of fashion design, providing effective methodological references for modern fashion design.
Source: Systems and Soft Computing
Author: Juntao Zhang
Institution: Zhengzhou Preschool Education College
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,870
OpenAlex ID: W7125939208