![[2511.18765] NI-Tex: Non-isometric Image-based Garment Texture Generation](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2511.18765] NI-Tex: Non-isometric Image-based Garment Texture Generation
Summary
The paper presents NI-Tex, a method for generating non-isometric garment textures using a new dataset and advanced techniques for cross-pose texture learning.
Why It Matters
This research addresses the limitations of existing texture generation methods in the fashion industry, providing a solution that enhances the realism and diversity of garment textures. By leveraging a novel dataset and innovative algorithms, it has the potential to significantly improve the quality of 3D garment design, which is crucial for applications in fashion, gaming, and virtual reality.
Key Takeaways
- NI-Tex utilizes a new dataset called 3D Garment Videos for better texture generation.
- The method overcomes challenges related to non-isometric image-geometry pairs.
- An iterative baking method is proposed for seamless PBR texture production.
- The architecture achieves high-quality, spatially aligned materials suitable for industry use.
- This approach enhances the flexibility and quality of garment texture generation.
Computer Science > Computer Vision and Pattern Recognition arXiv:2511.18765 (cs) [Submitted on 24 Nov 2025] Title:NI-Tex: Non-isometric Image-based Garment Texture Generation Authors:Hui Shan, Ming Li, Haitao Yang, Kai Zheng, Sizhe Zheng, Yanwei Fu, Xiangru Huang View a PDF of the paper titled NI-Tex: Non-isometric Image-based Garment Texture Generation, by Hui Shan and 6 other authors View PDF HTML (experimental) Abstract:Existing industrial 3D garment meshes already cover most real-world clothing geometries, yet their texture diversity remains limited. To acquire more realistic textures, generative methods are often used to extract Physically-based Rendering (PBR) textures and materials from large collections of wild images and project them back onto garment meshes. However, most image-conditioned texture generation approaches require strict topological consistency between the input image and the input 3D mesh, or rely on accurate mesh deformation to match to the image poses, which significantly constrains the texture generation quality and flexibility. To address the challenging problem of non-isometric image-based garment texture generation, we construct 3D Garment Videos, a physically simulated, garment-centric dataset that provides consistent geometry and material supervision across diverse deformations, enabling robust cross-pose texture learning. We further employ Nano Banana for high-quality non-isometric image editing, achieving reliable cross-topology texture ge...