7458 – Computational Image Rendering Using Procedural Yarn Model

Image rendering technique using a procedural yarn model for generating computer graphical representation of clothing fabrics.

Abstract

Realistically rendering cloth and textile in computer graphics, for use in digital applications and representation, requires accurate modeling of optical properties and frequent re-adjustments for representing appearance under different conditions. Conventional fiber-based models work either at micro-appearances, which, although detailed, require programming expertise. Conventional yarn-based models are powerful and convenient but lack fiber-level details captured in micro-appearances.

The current invention from Cornell University bridges the gap between procedural and micro-appearance models creating convenient models containing fiber-level micro-detail to provide richly detailed fabric models. The system does this through utilization of Computed Tomography (CT) data of yarn samples as input model parameters feeding to an image rendering system capable of generating outputs, based on procedural yarn algorithms, as images. This system can generate detailed images capturing fly-away fiber behavior.

Advantages

  • Produces fiber-level structure in yarn based procedural models.
  • Reproducibility verified by CT measurements and photographs.
  • Uses considerably lower memory and renders compact models without compromising performance in comparison to traditional models which face rendering challenges due to burgeoning size.

Potential Applications

  • Fashion Design – To simulate renders of potential garments.
  • Video games and movies – Lending realism to virtual realities and visual effects
  • Online retail – As product images
  • Tactical Military Uniforms – Simulating appearance of fabrics and fibrous materials in different conditions.
  • Textile manufacturing – Simulating wear and tear on fabrics to adjust yarn and fiber compositions in clothing.

Contact Information

Name: Martin Teschl

Email: mt439@cornell.edu

Phone: (607) 254-4454