Abstract:

Objective Design-preserving virtual garment transfer can transfer a garment from one body to another, which is significant for garment customization. However, most current research on virtual garment transfer only focuses on generating personalized 3-D garments rather than 2-D patterns and cannot be used for 3-D scanned human bodies. To address this issue, we propose a garment customization solution for 3-D scans based on design-preserving virtual garment transfer. Given a source garment worn on a source body, this research aims to obtain the target garments and patterns for a 3-D scan and show the final virtual fitting result of the target garment.
Method Firstly, the principle of virtual garment transfer of CLO 3-D and Marvelous Designer software is clarified in detail. Secondly, a method for generating personalized 3D garments and 2-D patterns for the 3-D scanned human bodies is proposed in combination with virtual garment transfer and the "Auto-convert-to-avatar" of CLO 3-D. Finally, the feasibility of this method in actual garment customization is verified by performing virtual garment transfer experiments with scanned human bodies.
Results In order to explore the difference in results between the two virtual garment transfer methods, two experiments were conducted and the results demonstrated that the method of using the default virtual avatar in CLO 3D or Marvelous Designer worked better than "Create fitting suit" when performing virtual garment transfer (Fig. 4). For the 3-D scanned human body, we proposed a personalized garment customization solution (Fig. 6). The virtual garment transfer for the 3-D scan was divided into three main steps. First, we used the CLO 3D default avatar to fit the scan and obtained the converted avatar based on the "Auto-convert-to-avatar " function of CLO 3D. Second, we obtained the transferred garment by transforming the source garment from the source body to the converted avatar. Finally, we fitted the transferred garment onto the scan and obtained the final virtual fitting effect. To verify the effectiveness of our proposed personalized garment customization solution, we scanned four young males with 3-D scanning and then obtained the corresponding converted avatars. For convenience, we denoted the scans as S_A, S_B, S_C, S_D and the converted avatars as C_A, C_B, C_C, C_D, respectively. According to our solution, we quickly obtained the transferred 3-D garment and corresponding 2-D patterns. The experimental results showed that the garments obtained by virtual garment transfer could meet the personalized requirements for different body shapes and sizes of the 3-D scanned human body (Fig. 10). In addition, the transferred patterns can satisfy the needs of actual garment production (Fig. 11).
Conclusion In this paper, we have proposed a personalized garment customization solution for a 3-D scanned human body based on design-preserving virtual garment transfer. For the 3-D scanned human body, our method was available for generating personalized 3-D garments and the corresponding 2-D patterns while also showing the final wearing effect of the transferred garment on the scan. The scheme's feasibility for actual garment customization was verified by conducting virtual garment transfer experiments with 3-D scans. The transferred 3-D garments can fit the scans well, and the transferred 2-D pattern can satisfy the needs of actual garment production. The method described in this paper was quick and effective for creating personalized 3-D garments and 2-D patterns. As a result, it can significantly improve the efficiency of customized garment development and facilitate garment customization in the apparel industry. In practical applications, creating a new style of 3-D garment is only necessary. Our proposed method allows personalized 3-D garments and corresponding 2-D patterns to be quickly obtained for 3-D scanned bodies of different sizes and shapes. In future work, how to optimize the existing personalized garment customization solutions is a topic worthy of in-depth study.

Key words: virtual garment, virtual garment transfer, 3-D scanned human body, digital avatar, virtual fitting, garment customization