Lego brick approximation of 3D models on a GPU using openCL

This research improved the execution time of the Lego brick approximation system implemented by (Bautista, 2007) using the Graphic Processing Units (GPU) parallel cores. The system is a variation of a Lego CAD (computer aided design) which is used in virtual Lego model design before it is tediously...

Full description

Saved in:
Bibliographic Details
Main Author: Guillergan, Paul Jonathan D.
Format: text
Language:English
Published: Animo Repository 2012
Subjects:
Online Access:https://animorepository.dlsu.edu.ph/etd_masteral/4303
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: De La Salle University
Language: English
Description
Summary:This research improved the execution time of the Lego brick approximation system implemented by (Bautista, 2007) using the Graphic Processing Units (GPU) parallel cores. The system is a variation of a Lego CAD (computer aided design) which is used in virtual Lego model design before it is tediously created by hand using actual Lego parts. It works by reading a 3D mesh model and automatically converts it to a 3D Lego model which is made up of Lego bricks. The algorithms used in the system runs in a sequential manner which can be time consuming when producing complex high-resolution Lego models. A new algorithm was developed called Triangle Mesh Shell Voxelization for parallel execution to replace one of the main processes of the system. The new Voxelization algorithm was designed such that it can be ported easily from sequential to parallel. The proponent used collision detection techniques in computer graphics to achieve portability. The parallel implementation was done using OpenCL which has a heterogeneous framework that enables computation on different processing units such as the GPU. Based on our experimental results we achieved a speedup of up to 5x from a CPU-only to a CPU+GPU implementation.