CSE 612 - Advanced Visualization
Instructor: Prof. Klaus Mueller
Office hours: Tu 3-5 pm (send
email for other arrangements), CS 2428
Email: mueller AT cs DOT sunysb DOT edu
Meeting time and venue:
CS 2311 (new Auditorium), Tuesdays
and Thursdays 5:20-6:40 pm
Summary (from course bulletin):
This course discusses
advanced concepts in the area of volumetric data modeling and
visualization. Topics included are: Visual exploration of multi-variate
and multi-dimensional datasets on regular and irregular grids, modeling
of natural phenomena and simulation of realistic illumination, volumes
as magic clay for sculpting and deformation effects, non-photorealistic
rendering for illustration and artistic works, information-centric
exploration of large datasets, and exploitation of hardware for
acceleration. The course strives to provide a snapshot on the current
state of the art and will be supported mostly by recent research papers.
Students will expand on a topic of their choice by completing an
CSE 564: Scientific Visualization
Research papers and handouts
Pick a project, either from a list provided early in
the semester, or on your own. It should be a project of substance, in
accordance to the level of the class and grade-percentage
Please attend every session. The occasional
pop-quizzes are meant to make sure that you’ve read the paper when one
was assigned to read. The quizzes won’t go into much depth, but they
will be decisive in what they’re designed for. Your presentation should
be related to the topic you’ve chosen for your project. Mid-way through
the course, we will also have 2 sessions allocated for progress reports
on each project in form of 15-20 minute presentations. A lively
discussion among the class is highly encouraged. The project
presentations (and the presentations in general) are meant as a
mechanism to improve the project, clarify pertaining issues, and come
up with cool new ideas, enhancements, and applications. There have been
a number of published research papers that resulted from last year’s
This year, a special emphasis is laid on
implementions that will run, or at least have good prospects to run, on
graphics hardware, although this is not a requirement. It is understood
that not all algorithms are well suited to hardware acceleration, but
nevertheless, recognizing properties that make hardware acceleration
possible is a goal of this class.
Course topics (for actual presentation times see website):
Getting to know the data:
Application domains - computational science,
engineering, design, medical, data warehouses, sensors, statistics,
business, entertainment, the arts.
Data generation - simulation, acquisition,
Data representations - grids, points, multi-valued,
multi-modal, time-variant, high-dimensional, vectors, tensors.
Tracking the energy - the high-albedo volume
rendering integral, radiative transfer.
Signal processing theory- Fourier analysis,
wavelets, implicit functions.
Signal processing practice - reconstruction,
filtering, sampling, compression.
Perception - modeling the human visual system,
metrics to gauge image quality.
User interfaces - for the general user, developer,
Quality vs. speed - where to strike the compromise.
Beyond the O-notation - the importance of cache,
MMX, pipelines, GPUs.
Regular grids - recap from CSE564 (raycasting,
splatting, shear-warp, texture-mapping).
Irregular grids - curvilinear, unstructured.
Higher dimensional and multivariate - tensor
rendering, time-varying, N-dimensional
Alternative algorithms - Fourier domain rendering,
wavelet domain rendering.
Primitives - points, micro-textures, cells, CSG, 3D
Decompositions - octree, space-time,
On-the-fly computed promitives - images, super-rays,
Techniques - 3D image-processing on-the-fly.
Applications - Illustrations, Impressionist and
Pen-and-Ink rendering, extension to SciVis.
Segmentation and registration:
Representations - Contour graphs, Reeb graphs.
Interactive segmentation - transfer function-based,
analysis of moments, segmentation by example
Techniques for finding good transfer functions
Feature extraction in scientific datasets - vortex
cores, complex flows, singularities
Boundary finding - active contours, balloons.
Modeling of physical processes and phenomena and simulation of
realistic illumination and appearance:
Volumetric radiosity - optical models, photon maps,
volumetric backprojection, Monte-Carlo
Realistic appearance models - BRDFs, reflectance
fields, synthesized textures.
Amorphous phenomena - clouds, smoke, fire, water,
Physically-based simulation of physical processes -
melting, ablation, breaking, fracturing.
Solution mechanisms for physical processes -
Navier-Stokes, lattice methods.
Efficiency considerations - cumputation
management, GPU acceleration.
Volumes as magic clay:
Deformations - volumetric sculpting, warping,
Effects to exterior influences- thawing, melting,
Level-of-detail modeling of real materials.
Sensitive rendering - rendering with haptic
Consumer graphics boards - programmable GPUs
(GeForce FX, ATI Radeon).
Parallel architectures - distributed vs. shared
memory, network of workstations (NOW).
Specialized hardware - VolumePro, Vizard
Information representation - glyphs, graphs,
parallel coordinates, tree plots
Feature extraction - the art of finding and
displaying the most relevant data
Visual data mining and monitoring - explorative
visualization with the scientist in the loop
The inter-relationship between information
visualization and scientific visualization
Visualization systems and case studies:
SciRun - Computational steering package from the
University of Utah.
Collaborative visualization system with network
VTK - Visualization Toolkit from Kitware, Inc.
Data Explorer - open-source system, formerly IBM.
Surgical simulator with haptic feedback
TeraGrid - a new large effort to facilitate data
sharing and distributed computing over the web.