Mesh Extraction from Volumes

- Adelailson Peixoto -

Updated: 07/06/2001

Overview

This research intends to develop and implement a technique for extracting meshes from volumes. It's based on the work developed by Zoe Wood et al. and based on other approaches. The main goal is to get, from those works, a background about all necessary steps to reconstruct a semi-regular mesh from volumes.

The process of reconstructing a mesh from a volume dataset may involve several steps, like:

Topology Extraction: this involves extracting the correct topology of the object that's inside the volume.

Coarse Mesh Construction: this step involves extracting an initial coarse mesh, using available topological information.

Mesh Refinement: this involves the refinement of the coarse mesh, until one obtains a final desired mesh representation.

The research of the new technique involves:

Understanding and implementation of a basic approach (Zoe Wood et al.) and other related works, that may or not present all above steps.
Extensions - Future works, that will try to extend the above works for the development of a new technique.

Basic Approach

The basic approach used as support for this research is a Master Thesis titled Semi-Regular Mesh Extraction from Volumes , developed by Zoe Wood. In such work are applied the steps of topology extraction, coarse mesh construction, and mesh refinement for reconstructing a semi regular mesh from a volume dataset:

Topology: the correct topology of the object that's inside the volume is obtained by constructing a topological graph. More .

Coarse Mesh: the coarse mesh is constructed from information stored at topological graph. More .

Refinement: this step uses distance transform to bring the coarse mesh to the final desired representation. This is done by applying successively internal and external forces to triangles that make up the mesh. More .

Other Works

Here are listed the main works related with the development of this work. By now the more important ones are:

Fast Marching Methods:

Fast marching methods, developed by J. A. Sethian, are efficient way to evolve or propagate interfaces. These methods may be used to propagate distance information from an initial curve or surface. In the above basic approach, fast marching methods are used during the topology extraction step, to propagate distances and construct the topological graph. More .

Fast Extraction os Adaptive Multiresolution Meshes with Guaranteed Properties from Volumetric Data:

This work (developed by Marcel Gavriliu et al.) uses a unified technique for both coarse mesh extraction and refinement steps, obtaining adaptive multiresolution meshes, from volumetric dataset. This unified technique uses concepts and strategies similar to the work developed by Zoe Wood. More.

Dynamic Meshes for Accurate Polygonization of Implicit Surfaces with Sharp Features

This approach (developed by Yukata Ohtake et al.) is mainly worried about the refinement problem. It reconstructs an accurate poligonanization of implicit surfaces with initial sharp features. It's based on mesh evolution towards a given implicit surface F(x,y,z)=0, with simultaneous control of the mesh vertex positions and mesh normals. More.

Proposed Approach

The present research proposes a technique for mesh extraction from volumetric data. In this new technique the topology extraction and the coarse mesh construction steps are performed based on the work developed by Marcel Gavriliu et al. The refinement step combines the Zoe and Yukata approaches. The distance transform computation is performed by using fast marching methods.

More Extensions: Future Works

After implementing the above proposed approach, the work intends to implement new tools related as future works in the original base works, such as the use of multi resolution representation of the volume.