Tessellation, Fairing, Shape Design and Trimming Techniques for Subdivision Surface based Modeling
Supported by NSF award DMI-0422126
Subdivision surfaces have become popular recently in graphical modeling, animation and CAD/CAM because of their flexibility, numerical stability, simplicity in coding and, most importantly, their capability in modeling and representing complex shapes of arbitrary topology. Actually, subdivision surfaces have already been used as primitives in several commercial systems such as Alias|Wavefronts's Maya, Pixar's Renderman, Nichimen's Mirai, and Newtek's Lightwave 3D. The objective of this project is to develop efficient techniques and algorithms for tessellating, rendering, fairing, shape designing and trimming of a Catmull-Clark Subdivision Surface (CCSS). Without a parametric representation, it is essentially impossible for subdivision surfaces to be included in a modern object modeling system because of problems with standard operations such as picking, rendering and texture mapping. In this project, I brought forth a new method to explicitly parameterize a CCSS based on Discrete Fourier transform techniques. With the availability of CCSS parameterization techniques, many standard 3D object operations have now become much easier. For example, based on our parameterization techniques, we developed a simple and efficient approach for interpolation of meshes with arbitrary topology, proposed a new method for adaptive tessellation and rendering of surfaces with accurate error control, and designed and implemented a simple yet effective texture mapping approach for objects with arbitrary topology. I have written several papers on this project.
My Contributions:
· Propose an explicit parameterization method for CCSSs;
·
Propose a new method for adaptive rendering and
tessellation of surfaces with accurate error control;
·
Design and implement a texture mapping technique
for shapes with arbitrary topology.
· Develop a new method for shadow generation of complex scenes represented by CCSSs.
Subdivision Surface based One-Piece Representation
Supported by NSF award DMS-0310645
The
objective of this project is to develop necessary mathematical theories and
geometric algorithms to support subdivision surface based one-piece
representation, i.e., representing any final object in a design process with
only one subdivision surface, no matter how complicated the object's topology
or shape. No decomposition of the object into simpler components is necessary.
Hence the number of parts in the final representation is always the minimum:
one. Performing accurate Boolean operations on different 3D objects and
representing the resulting object are very difficult problems in 3D object
design and modeling. Although many people have tried different approaches, no
perfect solution has been obtained yet. I proposed a novel method to join CCSSes while keeping the number of resulting CCSS to be
one. Our method also can be used for rendering CSG trees. We are still
improving this method so that it can be used for any complicated surface while
keeping the resulting joining surface as accurate as possible. My Ph.D.
dissertation is based mostly on this project.
My Contributions:
·
Develop new techniques for performing accurate
Boolean operations on 3D objects with arbitrary topology.
·
Propose an efficient method for voxelization of free-form solids.
·
Develop a robust and efficient approach for
interpolation of meshes with arbitrary topology;
·
Design novel algorithms for representing
resulting shapes after Boolean operations using only one mesh.
The Light Portal - 3D Reconstruction and Visualization over Space & Time
Supported by NSF award IIS-0448185
The focus of this project is on the problems associated with reconstruction and visualization of three-dimensional objects from images or video streams. The goal is to develop the fundamental underlying algorithms, and to implement an end-to-end light portal that can capture and "re-enact" real-world objects or events in a different time or space. A light portal captures a scene using a number of cameras with different viewpoints. From these input images, we either reconstruct a coherent parametric scene model that includes not only geometry but also lighting and surface reflectance properties (suitable for storage, editing, and playback over time), or directly interpolate the complete bundle of color rays to be displayed (suitable for live transmission over space). These results will then be visualized on a unique light-field display that utilizes an array of projectors and micro-lens screens to emit a large collection of controllable light rays, so that any number of viewers can simultaneously perceive view-dependent stereoscopic effects as if they were in the real scene. My part in this project was to implement a system designed to render 3D live scenes with stereo techniques using view dependent textured splatting approaches. I wrote a paper about this technique and had a poster exhibited at the SIGGRAPH 2004.
My Contribution:
· Develop and implement algorithms for rendering live scenes using view dependent textured splatting technique.
Constrained Design, Streamline Modeling, Automatic Fairing and Automatic Joining Techniques for NURB Surfaces
Supported by NSF award DMI-9912069
Results on this project will provide the design and
modeling community a shape generation technique that can be controlled not only
in shape ranges, but also in convexity and curvature distribution. This will
significantly shorten object modeling processes and, consequently, will help
reduce both the cost and cycle time of the product development process. I was involved
in this NSF funded project to implement constrainedly scaling of a CCSS. I
proposed a method to scale a CCSS while keeping the shape and size of specific
features (sub-structures) unchanged. The major contributions include CCSS
energy computation techniques and energy optimization techniques. The new
method is more powerful than previous methods in that it can handle more
complicated shapes and, can therefore be used for more challenging
applications. I wrote a paper about this technique.
My Contribution:
· Propose and implement a method for constrained scaling a CCSS while holding the shape and size of certain areas unchanged.
Uniform Message Center
Shanghai Holdfast Online
I
worked at Shanghai Holdfast Online Co., which is a subsidiary company of China TeleCom,
My Contribution:
·
Design and implement a gateway of a UMC (
Parallel Object-Relational Database Management System
Supported by
I took
part in a team to design and develop a parallel object-relational database
system named PORLES. My responsibility was to design and implement part of the
infrastructure of this system. I implemented the parallel data partition
storage system, the parallel index system, and the parallel large object
(multimedia data object) control system. The prototype runs in Linux and
Solaris operating systems and has a graphical user interface that was
implemented with java. My Master's thesis was about my work in this project. I
proposed some innovative algorithms in doing the project and published several
papers on this project.
My Contribution:
· Design and implement part of the infrastructure of a parallel object-relational database management system.
Portable Environment for Parallel Programming
Supported by
I
joined in a project team to design and implement a Portable Environment for
Parallel Programming (PEPP). Cooperating with a classmate, we designed and
developed a parallel program debugging system that is an important part of
PEPP. It is suitable for debugging BSP parallel programs, and has a graphical
user interface that was implemented with TCL/TK. We completed the project on
time and won a prize from the national education committee due to our
outstanding work. Currently, the PEPP is still used by students who write
parallel programs using BSP in
My Contribution:
· Design and implement a parallel programming debugger.
|
|
