Depth image based cloth deformation for virtual try-on

  • Authors:
  • Muhammad Mobeen Movania;Farzam Farbiz

  • Affiliations:
  • Institute for Infocomm Research, A* Star, Singapore;Institute for Infocomm Research, A* Star, Singapore

  • Venue:
  • ACM SIGGRAPH 2013 Posters
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

The affordability of RGBD cameras has pushed the retail shopping to the next level with the Virtual Try-on (VTO) application. Existing approaches [Giovanni et al. 2012], [Kevelham and Magnenat-Thalmann 2012] use skinning whereas systems like FitNect augment skinning with cloth simulation. Alternatively, image-based approach [Zhou et al. 2012] utilizes video clips for virtual try-on. In all of the existing approaches, the user's depth profile is not taken into consideration. Therefore, the user does not feel a good fit. Using customized avatars [Yuan et al. 2012] solves the fitting issue but it requires an offline user-look-alike avatar creation stage which is improper for walk-through scenarios. Furthermore, accurate physics based simulation of a 3D garment on a detailed avatar model is very time consuming.