Stitched Embeddings: A Unified Latent Space for 3D Garments and 2D Patterns

2026-07-01

AI Summary

This work addresses the lack of an efficient, simulation-free unified representation between 3D garment modeling and 2D patterns by proposing the first bidirectional latent space framework that operates without physical simulation. The method introduces BoxMesh as an intermediate geometric representation and leverages geometric priors from pretrained 3D foundation models to enable differentiable, bidirectional mapping between 3D garments and 2D patterns. The proposed architecture significantly improves computational efficiency, achieves state-of-the-art accuracy in pattern reconstruction, and enables novel applications such as recovering 2D patterns from 3D meshes and editing 3D garments through 2D pattern manipulation. This approach effectively bridges neural 3D vision with practical garment manufacturing pipelines.

Abstract

While garments are essential for realistic digital humans, their topological variety makes them much harder to model than parametric bodies. Traditional tailoring relies on 2D sewing patterns, yet bridging these patterns to 3D geometry currently requires physical simulations. We present Stitched Embeddings, the first simulation-free framework to unify 3D garment reconstruction and sewing pattern inference within a single bidirectional latent space. By leveraging the geometric priors of a pretrained 3D foundation model, our approach overcomes the data scarcity typically associated with high-quality garment modeling. We propose to use the BoxMesh as a critical intermediate representation to align 2D panels into 3D configurations without the computational overhead of a simulator. This architecture achieves state-of-the-art accuracy in pattern reconstruction while significantly improving efficiency. Furthermore, our differentiable pipeline enables novel applications, including pattern recovery from meshes and 3D editing from 2D patterns. Finally, this work provides a scalable link between neural 3D vision and the physical garment manufacturing pipeline. Project Page: https://andreus00.github.io/stitchedembeddings

Problem

• 3D garments
• 2D patterns
• latent space
• garment modeling
• simulation-free

Innovation

• Stitched Embeddings
• Unified Latent Space
• Simulation-Free Garment Modeling
• BoxMesh Representation
• Differentiable Pattern-to-Mesh Pipeline