Login | Register

Hair modeling and rendering using Ray-tracing on GPU

Title:

Hair modeling and rendering using Ray-tracing on GPU

Sedaghat, Nasim (2008) Hair modeling and rendering using Ray-tracing on GPU. Masters thesis, Concordia University.

[img]
Preview
Text (application/pdf)
MR45502.pdf - Accepted Version
8MB

Abstract

Realistic human hair simulation, especially in real-time, is one the most challenging problems in Computer Graphics due to the unique nature of hair strands. The major obstacle is the large number of hairs and rather small diameter of individual hair strands. Specifically, we are dealing with four problems in hair simulation: modeling, rendering, collision detection, and animation. In this thesis, we focus our studies on hair modeling and hair rendering. For modeling hair, we based our algorithm on key strand hair modeling where hairs are considered as wisps of similar hairs with one key strand per wisp. We added various parameters to this method to achieve more realistically-looking hair. Also, we introduced a new technique to produce short human hairstyles easily. We proposed a new 3D geometric object called Continual Cylinders to represent hairs in 3D space based on the fact that the cross section of hairs can be approximated by ovals. Simulating hair strands with this rather simple geometric object helps us generate realistic hairstyles without affecting running time. Rendering a sufficient number of individual hairs to convey realism in real-time is a very hard task in Computer Graphics. To be able to do this we rendered hair using GPU in this thesis and we successfully rendered about 10,000 hairs in real-time. Moreover, we used the ray tracing method, along with GPU, which is a time-consuming rendering technique that is mostly used to produce very high quality images. Here as well, we introduced and used various methods to achieve realism such as hairs self-shadowing, hairs fineness, slightly different hairs color between strands and within a strand itself.

Divisions:Concordia University > Gina Cody School of Engineering and Computer Science > Computer Science and Software Engineering
Item Type:Thesis (Masters)
Authors:Sedaghat, Nasim
Pagination:xi, 112 leaves : ill. ; 29 cm.
Institution:Concordia University
Degree Name:M. Comp. Sc.
Program:Computer Science and Software Engineering
Date:2008
Thesis Supervisor(s):Fevens, Thomas
ID Code:976073
Deposited By: Concordia University Library
Deposited On:22 Jan 2013 16:19
Last Modified:18 Jan 2018 17:41
Related URLs:
All items in Spectrum are protected by copyright, with all rights reserved. The use of items is governed by Spectrum's terms of access.

Repository Staff Only: item control page

Downloads per month over past year

Back to top Back to top