UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Garment modelling and visualisation

Spanlang, Bernhard; (2005) Garment modelling and visualisation. Doctoral thesis , UCL (University College London). Green open access

[thumbnail of Spanlang_thesis.Redacted.pdf]
Preview
Text
Spanlang_thesis.Redacted.pdf

Download (48MB) | Preview

Abstract

This thesis describes methods for modelling and visualising fabrics and garments on human body shapes by means of computer systems. The focus is on automatic real-time simulation but, at the same time, realistic appearance of garments is important. Very high physical accuracy in the simulation of mechanical properties of cloth is relaxed, however. We first introduce an image-based collision detection (IBCD) mechanism which harnesses the rasterisation, depth buffer and interpolation units of existing graphics hardware for robust and efficient collision detection and response between body-scan and garment surfaces. We introduce directional velocity modification (DVM), a numerical method to overcome a phenomenon called super-elasticity of conventional mass spring particle systems (MSPS). A reverse engineering technique is presented that enables us to extract assembly information and data for realistic visualisation from photographs of real garments. The images used in this technique are acquired in a set-up that includes a high resolution digital imaging device. Body landmark data is used automatically to pre-position garment panels around a body-scan before virtual sewing. We show that DVM generates visually pleasing garments on static and animated body-scans. The efficiency of the IBCD is demonstrated by a comparison with a traditional geometry based method. A system that exploits the developed techniques was created for Bodymetrics Ltd., at their request. This system is integrated with a 3D whole body scanner for fully automatic virtual try-on of clothes in a department store. The system presents a world first installation and was reported in several newspapers and fashion magazines. While it demonstrates the efficiency and robustness of the developed methods, clothes shoppers can benefit in that the system gives quicker fit-feedback than a traditional mirror in a changing room and can display their dressed body from any angle and distance. In collaboration with a garment designer we also tested the system in a design scenario. In an evaluation of the commercial potential of the developed methods the major players in markets closely related to virtual try-on technology are identified and market entry strategies are discussed.

Type: Thesis (Doctoral)
Title: Garment modelling and visualisation
Identifier: PQ ETD:602407
Open access status: An open access version is available from UCL Discovery
Language: English
Additional information: Thesis digitised by ProQuest. Third party copyright material has been removed from the ethesis. Images identifying individuals have been redacted or partially redacted to protect their identity.
UCL classification: UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Computer Science
URI: https://discovery.ucl.ac.uk/id/eprint/1446482
Downloads since deposit
51Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item