Compression of Large Engineering Models
Summary of work being done by Dinesh Shikhare
towards PhD Thesis under the supervision of Dr. S. P. Mudur at NCST's Graphics and CAD division.
(Completion of thesis is expected
in December 2001.)
Research
Direction:
Large 3D models like architectural designs,
heritage monuments, chemical plants and mechanical CAD designs (see examples) are increasingly
deployed in various applications involving interactive visualization and
Internet-based access. Very often the complexity of these 3D models far exceeds
the limits of what can be quickly downloaded at popular connection speeds
and what can be easily stored and rendered for interactive exploration on
personal desktops. To enable compact storage and fast transmission of such
models, compression of geometric description of these models has been an
area of
considerable interest. However, the earlier efforts do not specifically
address engineering models.
The polygonal models of the engineering class
have the following characteristics:
- the models consist of many small to medium
sized connected components,
- many geometric features repeat in various
positions, scales and orientations (see examples in the figure below),
- macro-structures consisting of multiple
features repeating across the models,
- these features repeat within and across
connected components in the models, and
- although the modeling tools and file
formats have facilties for specification and representation of repeating
features, in practice all the models we see have these features repeatedly
described -- this is more due to the difficulties in manual identification
of all the identical features at different scales.
Our research pioneers a new strategy for
compression of such models through the following new contributions:
- a new algorithm for automatic discovery
of repeating feature patterns in 3D polygonal mesh models,
- compression of geometry of 3D polygonal
mesh models by removing the redundancy in the representation of repeating
geometric feature patterns,
- a new scheme that can incorporate the
best results achieved in the area of connectivity compression of polygon
meshes, thereby always achieving compression ratios at least as good as
those reported in the literature.
In fact for engineering models, our technique
gives large benefits which cannot be realised by using the connectivity compression
algorithms alone. We demonstrate the efficacy of this new strategy through
examples of large engineering models -- including some that are freely available
on the net -- with excellent compression ratios. To the best of our knowledge,
ours is the first significant attempt to address compression of engineering
models.
Resources
for Research:
Our Papers/Reports:
- Dinesh Shikhare, State of the
Art in Geometry Compression, NCST Technical Report, December 2000
- Dinesh Shikhare and S. P. Mudur, Discovery of Repeating Feature Patterns
in Large 3D Mesh Models Workshop on Uncertainty in Geometric Computations
2001, July 5-6 2001, Sheffield, England.
- 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.
Unpublished writings
for local use:
These are some of the reports I wrote to
record the progress of my studies and research.
- Content acquisition techniques
for 3D visual computing, a short note, April 1998.
- Geometric data handling:
implementation issues, May 1999.
- Healing of 3D geometric databases,
May 1999.
- Use of 3D on the Web: Fundamental
issues and emerging technologies, November 1999.
- Feature-based techniques for
handling geometric models, a survey, November 1999.
- Signal processing of
triangle meshes, short note, December 2000.
- Complexity of geometric
models, a survey, May 2001.
- Compression of polygon meshes
of engineering models using discovery of repeating geometric features,
technical report, May 2001.
