Site hosted by Angelfire.com: Build your free website today!

Geometry Compression Resources

References to papers and sites for the general area of Geometry Compression.
This page is maintained byDinesh Shikhare

Papers

M. Deering, Geometry compression,
In SIGGRAPH'95 Conference Proceedings,
pages 13-20, August 1995.
This was the first paper on Geometry Compression.
Compression is achived by encoding geometry using
generalised triangle strips and by using
lossy techniques to encode vertex coordinates, colour
and vertex normals.

This work is now a part of Java3d
G. Taubin and J. Rossignac, Geometry
Compression through Topological Surgery
,
ACM Transactions on Graphics, July 1998.

Derives from the body of work on encoding
of planar graphs for traversal of manifold
meshes in a particular fashion to obtain a
vertex spanning tree. This traversal is
then encoded compactly. Connectivity encoding
is non-lossy. Geometry is predictively encoded.
The correction vectors are entropy encoded.
Normals, colour and texel are quantised.
G. Taubin, W.P. Horn, F. Lazarus and J. Rossignac,
Geometric coding and VRML, Proceedings of the IEEE,
July 1998. (Also, IBM Research TR RC-20925, July 1997).
Almost a rewrite of the "Surgery" paper.
G. Taubin, Andre Gueziec, William Horn and
Francis Lazarus, Progressive Forest Split Compression,
In SIGGRAPH'98, August 1998.

A mesh is split into a forest of spanning trees
of vertex runs. This decomposition is such that
the complete connectivity can be used to reconstruct
the original mesh. Geometry encoding is lossy.
Oliver Staadt, Markus Gross and Roger Weber,
Multiresolution Compression and Reconstruction,
Proceedings of IEEE Visualization '97,
pages 337-346, 1997.
Based on wavelet based encoding of a
regular grid of points in 2D or 3D.
This technique has two strategies for graph
surfaces and for general manifolds.
Graph surfaces are handled using a 2D grid
by intersecting the triangulation by the
grid and recording contours as iso-curves.
Encoding and reconstruction are quite involved.
For other meshes, they use 3D grids and
iso-surfaces.
Markus Gross, Lars Lippert and Oliver Staadt,
Compression Methods for Visualization,
Future Generation Computer Systems,
Vol 15, No. 1, pages 11-29, 1999.
Repackaging of the above paper.
Mike Chow,
Optimized Geometry Compression for Real-time Rendering,
In Proceedings of the IEEE Visualization '97, pp. 346-354 (November 1997). IEEE.
Edited by Roni Yagel and Hans Hagen. ISBN 0-58113-011-2..
A simple improvement over Deering's work.
Modifies the traversal of Generalised tri mesh
and removes the restriction of 16 vertex cache.
Stefan Gumhold and Reinhard Kline,
Compression of Discrete Multiresolution models.
ISSN 0946-3852, WSI-98-1.
Multi-resolution representation of tri-mesh models.
Nice, consice survey of simplification of meshes,
progressive mesh schemes. Contribution: new heirarchical
scheme for decomposition, representation and storage
of tri-mesh models. Storage compaction based on
detection of patterns in the heirarchical
representation.
S. Gumhold and R. Klein,
Compression of discrete multiresolution models
Technical Report WSI-98-1, Wilhelm-Schickard-Institut for Informatik,
University of Tuebingen, Germany, January 1998.


Rewrite of the above.
Jarek Rossignac
Edgebreaker: Connectivity compression for triangle meshes
Technical report GIT-GVU-98-35, GVU Center, Georgia Tech, 1998.
(Now, EEE Transactions on Visualization and Computer Graphics,
5(1), pp. 47-61 (January - March 1999). ISSN 1077-2626.)
Implementation is availablehere
Preprocessing renumbers the vertices of the mesh
and represents the model in half edge data structure.
Then traverses triangles, recording the history in
terms of 5 opcodes. Suitable for manifolds.
Connectivity is encoded in 1.5 to 2 bits per triangle.
The paper is very readable. Has a nice survey and
classification of the previous art.

Rossignac claims guarranteed upper bound on bits/vtx
in this encoding.

Stefan Gumhold and Wolfgang Strasser,
Real-time Compression of Triangle Mesh Connectivity,
In SIGGRAPH'98.
Very similar to Edgebreaker of Rossignac.
But independently developed and reported.
Renato Pajarola and Jarek Rossignac
Compressed Progressive Meshes
Technical Report: GIT-GVU-99-05, GVU Center,
Georgia Institute of Technology, January 1999.
An attempt to compactly encode Hoppe's
Progressive Meshes representation for tri-meshes.
Davis King and Jarek Rossignac
Optimal Bit Allocation Geometry Compression
Technical Report of GVU, Georgia Tech, 2000
This paper attempts to define metrics for measuring
"complexity" of a given polygonal model. Metrics are
developed for complexity of meshes in terms of
"shape factors". Relationships are proposed among
number of vertices, bit count, compression factor,
and shape factor.
F. Bossen,
On the art of Compressing Three-Dimensional Polygonal
Meshes and their Associated Properties.
PhD thesis,
Ecole Polytechnique Federale de Lausanne (EPLF),
June 1999.
This thesis is widely cited. I could not locate it
on the net. But I think the thesis of this work will
be close to what Taubin's group has done.
D. Cohen-Or, D. Levin and O. Remez,
Progressive Compression of Arbitrary Triangular Meshes.
In the Proceedings of Visualization 99.
pp. 67-72 (October 1999, San Francisco, California).
IEEE. Edited by David Ebert and Markus Gross and Bernd Hamann.
ISBN 0-7803-5897-X.
This scheme is based on decimation of triangle meshes.
For deletion of each vertex, the resulting hole is
filled with a new triangulation. Hierarchies of
the model are generated by at various resolutions.
Progressive transmission is achieved by sending coarse
model in the heirarchy first and following it up with
information for reconstructing details regarding deleted
vertices. A compact encoding of refinement information
achieves compression.
D. Cohen-Or, Y. Mann and S. Fleishman
Deep Compression for Streaming Texture Intensive Animations,
SIGGRAPH'99
Compression of streaming 3D animated geometry
alongwith textures. Coherence across the frames
in the animation is exploted to achieve compression.
Special treatment for textures is claimed with some quality factors for texture.
Jerome Edward Lengyel,
Compression of Time-Dependent Geometry,
1999 ACM Symposium on Interactive 3D Graphics,
pp. 89-96 (April 1999). ACM SIGGRAPH.
Edited by Jessica Hodgins and James D. Foley. ISBN
1-58113-082-1.
Compression of 3D animated geometry.
Exploits the coherence across frames using
a predictive encoder. Good description of
how animation is typically modelled in animation
packages.
L. Kobbelt and H-P. Seidel.
Efficient Storage, Compression and Transmission
of Complex Polygonal 3D Models,
Computer Graphics International 1998,
(June 1998, Hannover, Germany).
IEEE Computer Society.
... remarks pending ...
Costa Touma and Craig Gotsman.
Triangle Mesh Compression,
Graphics Interface '98, pp. 26-34 (June 1998).
Edited by Kellogg Booth and Alain Fournier.
ISBN 0-9695338-6-1.
This work reports improvements over the
Topological Surgery paper of Taubin and Rossignac.
The work is based on a slightly different traversal
of triangles in the mesh.
Gabriel Taubin, William Horn and Paul Borrel
Compression and Transmission of
Multi-resolution Clustered Meshes
IBM Research Report RC21398(96630),
February 1999.
This scheme of progressive transmission of
compressed meshes uses the Topological
Surgery scheme on parts of the model. The
partitioning of the model is based on
clustering of vertices and submeshes.
Andre P. Gueziec, Frank Bossen, Gabriel Taubin
and Claudio T. Silva.
Efficient Compression of Non-Manifold Polygonal Meshes,
IEEE Visualization '99, pp. 73-80
(October 1999, San Francisco, California).
IEEE. Edited by David Ebert and Markus Gross
and Bernd Hamann. ISBN 0-7803-5897-X.
This work extends the ealier efforts of Topological
Surgery (TS) to non-manifold models. This is done by
developing an algorithm to decompose a non-manifold
model into multiple manifold surfaces, which are then
compressed using TS algorithm.
Chandrajit L. Bajaj and Valerio Pascucci
and Guozhong Zhuang.
Progressive Compression and Transmission of Arbitrary Triangular Meshes,
IEEE Visualization '99, pp. 307-316
(October 1999, San Francisco, California).
IEEE. Edited by David Ebert and Markus Gross
and Bernd Hamann. ISBN 0-7803-5897-X.
Triangle-strip decomposition based scheme for
compression of meshes. Lossy predictive encoding
used for compact storage of geometry.
Simplification schemes are introduced for hierarchical
representation, storage and transmission.
Chandrajit Bajaj, V. Pascucci, G. Zhuang,
Single Resolution Compression of Arbitrary Triangular Meshes with Properties,
Computational Geometry: Theory and Applications,
vol 14: 167-186,
also in Proceedings of Data Compression Conference,
March 28-31, 1999 Snowbird, Utah,
TICAM Report Number 99-05
Triangle-strip decomposition based scheme for
compression of meshes. Lossy predictive encoding
used for compact storage of geometry.
Chandrajit Bajaj, V. Pascucci, G. Zhuang,
Compression and Coding of Large CAD Models
TICAM and Dept. of Comp. Sc.
The University of Texas at Austin.
Previous techniques have been extended
to handle quad and general polygonal models
and also non-manifold models through a new
polygon strip scheme.
Andrei Khodakovsky, Peter Schröder and
Wim Sweldens,Progressive Geometry Compression,
ACM SIGGRAPH 2000

This is a wavelet based scheme that works well
on regular and semi-regular connectivity meshes.
But it can't work effectively on arbitrary meshes.
Karni Z. and Gotsman C.
Spectral Coding of Mesh Geometry.
Proceedings of SIGGRAPH 2000,
Computer Science Department, Technion, Israel

Zachi Karni and Craig Gotsman
3D Mesh Compression Using Fixed Spectral Bases
Proceedings of Graphics Interface 2001
June 7th - 9th, 2001 Ottawa, Ontario, Canada.

The authors claim this scheme to be the JPEG
algorithm for arbitrary triangle meshes. This
signal processing based technique derives from
Taubin's SIGGRAPH'95 paper on surface fairing.
The methos decomposes the given model into smaller
patches and then encodes the patches using
the low frequency components of the spectral
analysis.
Martin Isenburg, Jack Snoeyink,
Face Fixer: Compressing Polygon Meshes with Properties.
Proceedings of SIGGRAPH 2000,
University of North Carolina at Chapel Hill.

An edge-based traversal and encoding scheme
which also handles meshes having arbitrary
polygons other than triangles.
Discusses encoding of attributes, connectivity.
Andrei Khodakovsky and Igor Guskov
Normal Mesh Compression
Caltech. Technical Report 2000.
(submitted for publication)

This work derives from their earlier work
on "Normal Meshes" published in SIGGRAPH'99.
The single scalar per vertex scheme of
heirarchical representation of meshes is used
here for compact encoding.
Pierre Alliez and Mathieu Desbrun
Progressive Encoding for Lossless Transmission of 3D Meshes
SIGGRAPH 2001.

This paper addresses progressive encoding
more than geometry compression. A fine-grained
fully progressive encoding is achieved for
lossless transmission of 3d polygonal models.
The encoding scheme has potential for compressed
representation.
Pierre Alliez and Mathieu Desbrun
Valence-Driven Connectivity Encoding for 3D Meshes
EUROGRAPHICS '2001

Based on valence-based approach pioneered by
Touma and Gotsman, this work is a new scheme for
valence-based conquest for arbitrary meshes that
always guarantees smaller compression rates than
the original method. Interesting discussion on
novel theoretical entropy study hinting the optimality
of the valence-driven approach.
Dinesh Shikhare, Sushil Bhakar and S. P. Mudur,Compression of Large 3D Engineering Models using Automatic Discovery of Repeating Geometric Features, 6th International Fall Workshop on Vision, Modeling and Visualization (VMV2001), November 21-23, 2001, Stuttgart, Germany.
(project page)
Captol buildingsome of the repeating shapes
Compression is achieved by automatically discovering
the shapes that repeat in the model and then by
avoiding the repeated description of those shapes.
This approach works exceedingly well on models of
engineering class where invariably one can find many
repeating shapes.
Mesh Compression and Its Hardware/Software Applications, Tulika Mitra, PhD Thesis, Department of Computer Science, State University of New York at Stony Brook, December 2000. This research reports (a) a breadth-first approach to
efficient mesh traversal, (b) on-the-fly rendering of
losslessly compressed irregular volume data, and
(c) compression domain triangle mesh rendering.

Sites