MEPP - 3D MESH PROCESSING PLATFORM

Guillaume Lavou

, Martial Tola

and Florent Dupont

Universit

e de Lyon, CNRS, INSA-Lyon, LIRIS, UMR5205, F-69621, Villeurbanne, France

Universit

e de Lyon, CNRS, Universit

e Lyon 1, LIRIS, UMR5205, F-69622, Villeurbanne, France

Keywords:

Mesh Processing, Open Source, Platform.

Abstract:

This paper presents MEPP, an open source platform for 3D mesh processing. This platform already contains a

large set of processing tools from classical ones (simpliﬁcation, subdivision, segmentation) to more technical

algorithms (compression, watermarking, Boolean operation, perceptual metrics, etc.). Its main objective is to

allow a quick start for both users and developers by providing highly detailed tutorials and simple integration

mechanisms, through a modular architecture where components are implemented as dynamic plugins.

1 INTRODUCTION

Technological advances in the ﬁelds of telecommuni-

cation, computer graphics, and hardware design dur-

ing the last decade have contributed to the emergence

of three-dimensional (3D) data, often represented by

polygonal meshes, in numerous industrial domains

like mechanical engineering, scientiﬁc visualization,

digital entertainment or cultural heritage. This emerg-

ing type of data is more complex to handle than other

media such as audio signals, images or videos, and

thus has brought new challenges to the scientiﬁc com-

munity, and has open new research domains like ge-

ometry processing.

To ease the development of new algorithms for

3D mesh processing, it is crucial to have some open

source tools and libraries available for the scientiﬁc

community. The library CGAL (The CGAL Project,

2011) (Computational Geometry Algorithms Library)

introduced for several years, offers powerful geome-

try processing algorithms and data structures like the

Polyhedron type for manipulating manifold polygonal

meshes, however it is based on template programming

and thus is not so easy to use for a beginner. Some

open source softwares have also been introduced like

MeshLab (Cignoni et al., 2008), Graphite (Graphite,

2003) or OpenFlipper (M

obius and Kobbelt, 2010),

based on their own data structure for 3D mesh ma-

nipulation; they offer interesting geometry processing

for user and also allow developers to add plugins with

their own code, more or less easily.

We introduce the open source platform MEPP (3D

MEsh Processing Platform) which can be seen as co-

mplementary to these existing works. It is available

online

and its strong points regarding existing works

are as follows:

• it focuses on making the use and installation very

easy, whichever the operating system, by provid-

ing highly detailed user and developer tutorials;

the objective is to allow a quick start for new de-

velopers,

• it simpliﬁes the use of CGAL data structures by

hiding the template programming,

• it provides not only standard modeling tools

(simpliﬁcation, subdivision, segmentation) but

also highly technical components like progres-

sive compression, watermarking, Boolean opera-

tion and perceptual metrics,

• it allows the visualization and processing of color

meshes and dynamic mesh-sequences.

Next sections present respectively the main architec-

ture of the platform, its manipulation and visualiza-

tion functionalities and a rapid description of cur-

rently available components.

2 ARCHITECTURE

This section describes successively the goals of our

platform, its kernel and related features and then the

principles of component / plugin development.

http://liris.cnrs.fr/mepp/

206

Lavoué G., Tola M. and Dupont F..

MEPP - 3D MESH PROCESSING PLATFORM.

DOI: 10.5220/0003928502060210

In Proceedings of the International Conference on Computer Graphics Theory and Applications (GRAPP-2012), pages 206-210

ISBN: 978-989-8565-02-0

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

2.1 Goals

MEPP provides a GUI and basic functionalities to al-

low development and fast integration and can also be

easily enriched with new features and new modules.

The platform provides a modular architecture through

the use of components available as dynamic plugins.

As it is important that all developments have a min-

imal impact compared to an operating system to

provide a great portability, we have directed our

choices to use robust, ﬂexible and effective C++

cross-platform open source libraries.

In order that users and developers can easily use and

enhance the platform, a major effort has focused on

the fact that it can be deployed quickly and easily,

thanks in particular to a detailed installation docu-

mentation. The integration of new developments is

facilitated by a step by step documentation on the use

and enrichment of the platform.

2.2 Kernel and Related Features

Based on CGAL

, Qt

, libQGLViewer

, OpenGL

Boost

and FFmpeg

, the MEPP platform, written

in C++ under the GNU GPL v3, runs on Windows,

Linux and Mac OS X. It provides a development

environment based primarily on the class Polyhe-

dron of CGAL, library based on the principles of

Object-Oriented Programming and especially para-

metric polymorphism, i.e. template.

Without any component, the MEPP kernel only al-

lows loading, viewing and saving mesh and mesh se-

quences. The loading of mesh can be done by using

the menu or by drag and drop from one application

window or from a ﬁle browser.

MEPP allows the management of multiple objects

in one or more windows and offers two types of pro-

cessing and display for the meshes:

• the ”space” mode in which several objects are

treated in the same scene, allowing to compare

them, to observe them by coupling their rotation,

or to assess the results of treatments (Figure 1),

• the ”time” mode in which several objects are seen

as a sequence of meshes and visualized using a

3D+t conﬁgurable video recorder (step by step,

loop, reverse, speed, etc.).

http://www.cgal.org/

http://qt.nokia.com/

http://www.libqglviewer.com/

http://www.opengl.org/

http://www.boost.org/

http://ffmpeg.org/

At any time, users can easily switch from one mode

to another.

The platform offers an OpenGL accelerated dis-

Figure 1: The interface of MEPP, here in ”space” mode.

play with ”Display Lists” in order to allow the use

of all graphics cards. The rendering is possible

as cloud of points, wireframe and conventional 3D

with or without reinforcement of points and edges.

Three color modes are available: ”vertex color”, ”face

color” or ”material”.

Meshes related to OBJ (Wavefront), OFF (Object File

Format), PLY (Polygon File Format), SMF (3D World

Studio) and X3D (XML-based free format) can be

read. For now, only the formats OBJ (Wavefront) and

OFF (Object File Format) can be written. Image and

video capture (H.264/MPEG-4 AVC) are also avail-

able.

The management of the mesh is also done within

the MEPP kernel, ie its memory storage and ac-

cess but also basic functionalities associated (num-

ber of components and boundaries, normals, bound-

ing box, degree, valence, genus, tags, etc.). The ker-

nel also provides to all components a ”Viewer” object

that manages the display and behavior of the object

”Scene” according to the mode (”normal”, ”space” or

”time”), itself containing the object or objects Polyhe-

dron (Figure 2).

The structure of underlying data used in the plat-

form to represent a mesh is based on the concept of

semi-oriented edges, ie half-edges, with relations of

incidence and adjacency. It restricts the class of rep-

resentable mesh with those of type manifold, with or

without borders; we call it a polyhedron.

One of the big advantages of MEPP, however, is

the fact that the deﬁnition of this polyhedron only ap-

pears rarely to the developer eyes, in order to best ab-

stract the CGAL library.

MEPP - 3D MESH PROCESSING PLATFORM

207

2.3 Component/Plugin Development

Through the principle of inheritance, the polyhedron

and its geometric items (vertices, edges, facets) can be

enhanced, new classes are derived from these existing

classes, which gives the developer the opportunity to

add his own data and associated functions.

Given this behavior and for modularity, the solution

that we have chosen for MEPP is the multiple ”con-

ditional” inheritance, each component can thus enrich

the geometric items (vertices, half-edges, facets) and

the polyhedron itself.

The resulting enriched polyhedron and its geometri-

cal items, therefore inherit of all the enhancements

brought by the inheritance of each respective compo-

nents implemented as plugins (.dll / .so / .dylib) with

selection at compilation and detection and automatic

loading at runtime.

This principle of plugin also allows us to enhance the

MEPP GUI (menus and toolbars) and at the same time

to resolve the problem of heterogeneous licenses (free

component, private component with more restrictive

license, etc.). In order to manage interactions, events

related to the interface (pre-draw, post-draw, mouse

movements and clicks, key press, etc.) are transmitted

from the kernel to components through a signal mech-

anism speciﬁc to Qt (Figure 2). MEPP is entirely gen-

Figure 2: MEPP architecture.

erated by the cross-platform open-source build sys-

tem CMake

. On Linux and Mac OS X, a prerequi-

site of packages (speciﬁc to the distribution for Linux,

Homebrew

or MacPorts

for Mac OS X) is required.

On Windows, these packages are provided as ”binary

kits” precompiled for 32 and 64 bit architectures. A

http://www.cmake.org/

http://mxcl.github.com/homebrew/

http://www.macports.org/

Linux virtual machine ”ready to start” is also avail-

able, offering to start even faster in developing.

3 COMPONENTS

This section provides an overview of currently avail-

able components of our platform.

3.1 Basic Processings

Some basic processing operations are provided:

• basic mesh manipulation (triangulation, noise ad-

dition, smoothing, scaling, rotation, translation),

• subdivision algorithms (Loop, Sqrt3, Doo-Sabin,

Catmull-Clark, Quad triangle),

• simpliﬁcation algorithms (Lindstrom-Turk imple-

mentation from CGAL and a canonical vertex re-

moval algorithm).

Figure 3 illustrates a result of simpliﬁcation.

Figure 3: Two objects in the same space: the original col-

ored mesh (1,2M vertices) and result after simpliﬁcation.

3.2 Curvature and Segmentation

Our platform provides the curvature calculation im-

plementation of Cohen-Steiner and Morvan (Cohen-

Steiner and Morvan, 2003), with a geodesic neighbor-

hood integration. It computes both min-max curva-

ture values and directions. A segmentation algorithm

is also provided, implementing the Variational Shape

Approximation from Cohen-Steiner et al. (Cohen-

Steiner et al., 2004). Results from curvature calcu-

lation and segmentation are illustrated in ﬁgure 4.

GRAPP 2012 - International Conference on Computer Graphics Theory and Applications

208

Figure 4: Segmentation of the Blade model and curvature

values and directions of the RockerArm model.

3.3 Boolean Operation

A fast and exact Boolean operation algorithm be-

tween 3D meshes is implemented (Leconte et al.,

2010); this algorithm is able to compute the union,

intersection and difference between two 3D meshes.

One of the key feature is the speed, indeed for in-

stance this algorithm is able to compute the intersec-

tion, union or difference between two 80K vertices

meshes in about 2.5 seconds on a 2GHz processor.

Figure 5 illustrates some Boolean operations between

two 3D models.

Figure 5: Results after union, difference and intersection

between the LionVase and the Fandisk objects.

3.4 Perceptual Quality Metrics

Recent perceptual metrics are implemented in the

platform (Lavou

e, 2011); given a distorted 3D shape

and a reference one, they compute a score that pre-

dicts the perceived distortion between them, as well

as a distortion map. These perceptual quality metrics

may be more relevant than classical root mean square

distance to evaluate or drive processing operations. To

our knowledge, no open platform currently proposes

such tools.

3.5 Compression and Watermarking

The MEPP platform proposes a recent progressive

compression algorithm applying for colored meshes

(Lee et al., 2011b), and a join watermarking scheme

(Lee et al., 2011a). The compression method provides

high compression ratio and allows a decompression

by Level-of-Details (LoD). When the watermarking

method is activated, a secret bit string is hidden in

each level-of-details, it allows the protection of the

decoded LoD.

3.6 Minkowski Sum

Finally, our platform provides an algorithm for

Minkowski sum of convex polyhedra (Barki et al.,

2009). The algorithm is exact and quite fast regarding

its state-of-the-art counterparts.

4 CONCLUSIONS

We have presented the Mesh Processing Platform

MEPP; its main objective is to ease the develop-

ment of new processing algorithms for 3D meshes.

This platform is mostly based on the Polyhedron type

from CGAL, however one of the main advantages of

MEPP, is the fact that the deﬁnition of this polyhe-

dron structure is hidden to the developer eyes; hence

this will allow us, in the next future, to integrate other

data structures, for instance supporting non-manifold

meshes; we also plan to integrate the management of

textured meshes.

REFERENCES

Barki, H., Denis, F., and Dupont, F. (2009). Contributing

vertices-based Minkowski sum computation of con-

vex polyhedra. Computer-Aided Design, 41(7):525–

538.

Cignoni, P., Callieri, M., and Corsini, M. (2008). Meshlab:

an open-source mesh processing tool. In Eurographics

Italian Chapter Conference.

Cohen-Steiner, D., Alliez, P., and Desbrun, M. (2004). Vari-

ational shape approximation. In ACM Siggraph, pages

905–914.

Cohen-Steiner, D. and Morvan, J. (2003). Restricted delau-

nay triangulations and normal cycle. In 19th Annu.

ACM Sympos. Comput. Geom.

MEPP - 3D MESH PROCESSING PLATFORM

209

Graphite (2003). http://www.loria.fr/ levy/Graphite/index.html.

Lavou

e, G. (2011). A Multiscale Metric for 3D Mesh Vi-

sual Quality Assessment. Computer Graphics Forum,

30(5):1427–1437.

Leconte, C., Barki, H., and Dupont, F. (2010). Ex-

act and Efﬁcient Booleans for Polyhedra.

http://liris.cnrs.fr/Documents/Liris-4883.pdf.

Lee, H., Dikici, C., Lavou

e, G., and Dupont, F. (2011a).

Joint reversible watermarking and progressive com-

pression of 3D meshes. The Visual Computer, 27(6-

8):781–792.

Lee, H., Lavou

e, G., and Dupont, F. (2011b). Rate-

distortion optimization for progressive compression of

3D mesh with color attributes. The Visual Computer.

obius, J. and Kobbelt, L. (2010). OpenFlipper: An Open

Source Geometry Processing and Rendering Frame-

work. In Curves and Surfaces.

The CGAL Project (2011). CGAL User and Refer-

ence Manual. CGAL Editorial Board, 3.9 edition.

http //www.cgal.org/.

GRAPP 2012 - International Conference on Computer Graphics Theory and Applications

210