Xueqing Zhao

Traditional Chinese textile patterns are often used as decorative elements on garments, fabrics, and accessories, adding unique cultural charm and aesthetic value to textile products. Recently, the big advances in data-driven artificial intelligence (AI) technology offer new ways to study textile patterns, but collecting enough data for training is challenging. One-shot image synthesis technology can help in constructing datasets under data-scarce conditions. Nevertheless, existing one-shot image generation methods are inadequate in fulfilling the structural layout and efficiency requirements for the production of traditional Chinese textile patterns. This paper proposes a one-shot Chinese textile pattern image generation (OS-CTPIG) method, which facilitates the generation of Chinese textile patterns with central radial structures, faithful colors, and exquisite details. A multiscale framework is employed for the generation of details in a coarse-to-fine manner. The simultaneous control over color and structure, along with the respective optimization at the color space and spatial feature space, complemented by a color leakage constraint module, is designed to preserve the traditional aesthetic standards in structures. Moreover, the integration of an energy function-enhanced sliced Wasserstein distance optimization objective and a dynamic task-weighting mechanism is proposed to improve computational efficiency. The experimental results demonstrate that the OS-CTPIG method is both effective and efficient in generating traditional Chinese textile pattern images under one-shot conditions. Subsequently, a dataset comprising over 3000 images of traditional Chinese textile pattern images has been constructed successfully.

As research and applications of artificial intelligence generated content (AIGC) in the field of fashion design continue to advance, existing intelligent garment generation models commonly face challenges such as high computational costs, monotonous texture generation, and training instability, which often lead to blurred details when generating diverse garment forms (e.g., wide sleeves in Hanfu, pleats in dresses). To address these issues, an intelligent garment generation method based on progressive training and contrastive learning strategy is proposed in this paper, which we call PT-CLGAN. First, a lightweight progressive generator is designed to reduce the computational load while ensuring high-resolution details of the garment generation. Second, a contrastive learning strategy is used to enhance the decision boundary of the discriminator, and capture the latent space feature contrast, thereby improving the training stability of PT-CLGAN. Finally, in order to verify the effectiveness of the proposed method, the simulation experiments are simulated on GarmentDesign, ExpandFashion, and DeepFashion2, we use PSNR, LPIPS, and MS_SSIM metrics for generated garment images as the objective evaluation metrics, and compare our proposed method with existing methods such as generative adversarial network (GAN), DCGAN, and TexGAN. The comprehensive experimental evaluation confirms the superiority of the proposed approach over traditional generative models, with significant improvements in garment synthesis quality. This approach enables the generation of previously unexplored pattern, cut, and color combinations by designers, thereby transcending the limitations of human experiential constraints. For instance, GANs can effectively integrate traditional embroidery techniques with contemporary minimalism to produce distinctive visual lexicons.

Traditional Chinese patterns, as one of the core cultural heritage carriers, face inheritance challenges due to traditional design’s heavy reliance on professional expertise and lengthy cycles. The rise of Artificial Intelligence Generated Content (AIGC) has introduced a human-intelligence collaborative paradigm to address these issues—this article focuses on leveraging AIGC for intelligent design of traditional Chinese patterns. Specifically, a dataset of traditional Chinese patterns was first constructed, featuring 30,000 images represented by Tuanke patterns. Subsequently, ChiPa, an intelligent design method tailored to traditional Chinese patterns, was proposed. This method enables simultaneous control over both the content and color. ChiPa’s core is a novel Structure Projector (Feature Map Mixture + Latent Space Mapper with MCConv), which maps sketches to StyleGAN’s latent space to generate structurally accurate grayscale Tuanke patterns. For color customization, ChiPa employs an arbitrary style transfer module to render grayscale patterns via user-specified references. Extensive experiments validate ChiPa’s superiority: it outperforms SOTA methods in all metrics; 3 domain experts and 30 volunteers rated it highest in structural standardization, stylistic consistency, and symbolic accuracy. Case studies demonstrate applications in cultural (e.g., packaging) and commercial (e.g., sofa upholstery) products. ChiPa lowers public participation barriers, expands Tuanke patterns’ carriers, and promotes living conservation and dissemination of Chinese traditional cultural heritage.

The design innovation of textile and apparel is the core of enterprise competitiveness. By assessing the value of intellectual property rights such as textile and garment design copyrights, it can not only stimulate the innovation motivation of designers, but also realize the continuous value-added and market expansion of clothing brands, thereby promoting the development of innovative design in the textile and garment industry. In this paper, inspired by the artificial intelligence and computational model, a method based on a multimodal cultural product value assessment hybrid expert model (MCP-MOE for short) is proposed for textile and apparel design. First, a multimodal representation learning mechanism is designed to learn the multidimensional features of cultural products, by fusing multimodal data and extracting multidimensional features, to realize the complementary advantages of cultural product features. Second, we construct a MCP-MOE model to regressively train the above multidimensional features, which can dynamically obtain the key features of cultural products more objectively, without relying on subjective factors. Finally, in order to verify the effectiveness of the proposed method, the simulation experiments are simulated on the multimodal dataset, mean absolute error, mean squared error, and root mean squared error are used as the objective evaluation metrics, and compare our proposed method with existing methods such as XGBoost, LightGBM, and CatBoost. The extensive experimental results show the proposed method has a good performance in assessing cultural value, and makes sense in the study on the evaluation of cultural value.

The study of ancient textile patterns not only enriches historical material but also inspires modern costume art design. Recent developments in computer vision technology provide novel analytical methods for the study of ancient textile patterns. As contours are one of the most significant features in computer vision tasks, contour extraction serves as a prerequisite to accomplishing intelligent understanding and automatic design of textile patterns. In this paper, a perception-consistent contour extraction method (ViCo) that simulates the human visual system from the retina, the lateral geniculate nucleus (LGN), to the V1 area in the cortex is proposed, along with an ancient textile pattern dataset represented by TuanKe in the Chinese Tang dynasty. Experiments were conducted both qualitatively and quantitatively to verify the effectiveness of the proposed method on the TuanKe dataset. Analysis of the results implies the effectiveness of ViCo because the retinal and LGN filter enhances edge information and restrains tiny pixel changes, and the cortex simulation is direction selective. This study is among the initial efforts to employ computer vision technology for ancient textile pattern studies. The study presents a TuanKe dataset and a novel bio-inspired contour extraction method, which has significant historical research importance, fabric art design inspiration, and computer application value.