The field of fashion technology is moving towards more personalized and interactive experiences. Researchers are exploring the use of multimodal models to enhance retailing through natural language and visual interactions. There is a growing interest in generating high-fidelity 3D garments and sewing patterns, with a focus on structured and editable parametric representations. Additionally, there is a trend towards developing more general models that can learn the dynamics of complex clothing and enable predictions regarding clothing movement. Noteworthy papers include: FashionM3, a multimodal fashion assistant that delivers contextually personalized suggestions; GarmentX, a novel framework for generating diverse and wearable 3D garments; and GarmentDiffusion, a generative model capable of producing centimeter-precise 3D sewing patterns from multimodal inputs.
Fashion Technology Developments
Sources
FashionM3: Multimodal, Multitask, and Multiround Fashion Assistant based on Unified Vision-Language Model
Masked Language Prompting for Generative Data Augmentation in Few-shot Fashion Style Recognition
GarmentX: Autoregressive Parametric Representations for High-Fidelity 3D Garment Generation
Learning a General Model: Folding Clothing with Topological Dynamics
YoChameleon: Personalized Vision and Language Generation
GarmentDiffusion: 3D Garment Sewing Pattern Generation with Multimodal Diffusion Transformers