Skip to main content
notice

Seminar by Dr. Manfred Lau (Lancaster University)


Speaker: Dr. Manfred Lau
                Lancaster University (UK)

Title: Computer Animation and 3D Modeling for Virtual and Physical Worlds

Date: Monday, February 9th, 2015

Time: 10:30AM - 12PM

Place: EV 3.309

ABSTRACT

In this talk, I will describe my research work in a number of areas in 3D computer graphics,
user interfaces, and digital fabrication. One area of my research is in computer animation and
computer games. Generating animations of crowds of characters is an important problem for applications
such as games, films, and virtual reality simulations. This problem is challenging because of the need
to generate natural human-like motion interactively for a large number of characters. I will describe
our Behavior Planning approach that performs a global search on a small set of motion capture clips
to solve this problem. I then discuss our Precomputed Search Trees technique which provides a two
orders of magnitude improvement over Behavior Planning. Another important issue for crowd simulations
is that while variation in human motion exists because people do not perform actions in precisely the
same manner every time, animation systems do not realistically produce these subtle variations.
My novel approach learns a generative Dynamic Bayesian Network model from data and can synthesize
a family of spatial and temporal variants.

Another area of my work is in sketch-based 3D modeling. Despite years of research in this area,
it is still difficult for the layperson to create detailed 3D models intuitively. My Face Poser
system combines sketch-based interfaces and captured human face data to allow novice users to
interactively pose new 3D face expressions with natural appearances. Natural face expressions
can provide a sense of realism for virtual characters in games, films, and virtual simulations.
Building such an interface is challenging because there can be many unnatural facial expressions
that are consistent with the ambiguous user input. My system automatically learns a facial prior
from a prerecorded face expression database and uses it to remove the ambiguity.

Fabrication is an emerging research area, and my recent work has been in geometry modeling and
user interfaces for fabrication. 3D printers and laser cutters are becoming increasingly common
and affordable, and it has been predicted that such technologies will lead to a new industrial
revolution. My contribution in this area from the computer graphics perspective is in
fabrication-aware 3D modeling. I will describe our novel formal grammar framework for converting
virtual 3D furniture models into real-world furniture. In the past few years, I have contributed
to the fabrication area from the human-computer interaction perspective by building a number of
user interface tools for personal fabrication. I will discuss our photo-based interface
(Modeling-in-Context), 2D sketch-based interface (SketchChair), and 3D gesture based interface
(MixFab). These interfaces have the common goal of making it easier for the layperson to design,
visualize, and fabricate personalized products by themselves and in their own home. Finally,
I will end the talk by describing possible ideas for future work.

BIO

Manfred is an Assistant Professor (Lecturer in UK) in the School of Computing and Communications
at Lancaster University. He was a post-doc researcher working with Takeo Igarashi and Jun Mitani
in Tokyo, Japan at the Japan Science and Technology Agency - Igarashi Design Interface Project.
He received his B.Sc. degree in Computer Science from Yale University, and his Ph.D. degree in
Computer Science, supervised by James Kuffner, from Carnegie Mellon University. Manfred's research
interests are in computer graphics, geometric modeling, animation, user interfaces, and digital
fabrication. His Ph.D. thesis work explores a combination of motion planning techniques and
captured data to generate realistic crowd animation for games and films. His recent research in
3D modeling and fabrication focuses on building natural user interfaces for the layperson to model,
design, and fabricate their own products.




Back to top

© Concordia University