« Todos los Eventos
From physical textiles to digital twins
15 marzo, 2023 @ 16:00 - 18:00
PONENTE |
Dra. Elena Garcés García |
AFILIACIÓN |
Head of AI & Optics at SEDDI |
RESUMEN |
Virtual design, online marketplaces, product lifecycle workflows, AR/VR, videogames, …, all require lifelike digital representations of real-world materials (i.e., digital twins). Acquiring these digital copies is typically a cumbersome and slow process that requires expensive machines and several manual steps, creating roadblocks for scalability, repeatability, and consistency. Among the many industries requiring digital twins of materials, the fashion industry is in a critical position; facing the demand to digitize hundreds of samples of textiles in short periods, which cannot be achieved with current technology. In this context, casual capture systems for optical digitization provide a promising path for scalability. These systems leverage low cost devices (such as smartphones), one or more different illuminations, and learning-based priors to estimate the material’s diffuse and specular reflection lobes. However, to train a machine learning-based solution, data is needed, which is not easy to obtain for textile materials. In this talk, I will discuss our end-to-end approach to providing a scalable solution to digitize textiles, including the design of an optical gonioreflectometer (capable of seeing yarns at the fiber level) for dataset creation and our deep learning-based solution that only requires a single image as input. |
BIO |
Elena es doctora en Informática desde 2016 donde se especializó en técnicas avanzadas de procesamiento de imagen y render inverso para captura de apariencia. Actualmente es investigadora en la URJC además de líder del equipo de IA y óptica de la empresa SEDDI, donde desarrolla métodos de captura de materiales, y modelado de humanos y ropa virtual. |