          |
GRASP Seminar Series: Fall 2005
November 4, 11:00 AM, 307 Levine Hall
Kristin Dana
Rutgers University
“Computational Skin Texture: Surface Detail in Object Models”
Abstract: Surfaces in natural scenes are detailed landscapes, with complex geometry and local optical properties. Consider the surface of human skin with pores, wrinkles, hairs and other variations. Surface detail is a necessary component of an object representation for high level tasks such as recognition and for low level tasks such as point correspondences. Modeling surface detail is fundamentally different from object modeling because a geometry + shading model is inadequate. Fine scale geometry, such as skin pores or fabric wrinkles are difficult to measure. Moreover, even with precise geometric models, pointwise shading models are insufficient because they do not account for local shadowing, occlusion and interreflections that occur due to fine scale geometry. Our approach to surface modeling uses the bidirectional texture function. We develop models of surface texture and demonstrate their use in recognition tasks. For measurement, we develop two novel methods for capturing fine-scale surface details: (1) a texture camera to obtain reflectance and surface height variation using curved mirrors and (2) an imaging method called polarization multiplexing where multiple unknown light sources illuminate the scene simultaneously, and the individual contributions to the overall surface reflectance are estimated.
Biography: Kristin Dana is an assistant professor in
the Department of Electrical and Computer Engineering at Rutgers, The
State University of New Jersey. She is currently conducting research on
computer vision and graphics, computational models for object appearance
and image texture with applications in pattern recognition and scene rendering,
and optical systems for measurements of surface appearance. Dr. Dana is
also a member of Rutgers Center for Advanced Information Processing, Rutgers
Center for Cognitive Science, a member of Graduate Faculty of the Computer
Science Department (2002-2005), and an adjunct assistant professor of
medicine at UMDNJ, Robert Wood Johnson Medical School. Dr. Dana received
the PhD from Columbia University (NY, NY) in 1999 and the MS degree from
Massachusetts Institute of Technology in 1992, and a BS degree in 1990
from the Cooper Union (NY, NY). From 1992-1995 she was on the research
staff at Sarnoff Corporation developing real-time motion estimation algorithms
for applications in defense, biomedicine and entertainment industries.
She is the recipient of the General Electric "Faculty of the Future"
fellowship in 1990, the Sarnoff Corporation Technical Achievement Award
in 1994 for the development of a practical algorithm for the real-time
alignment of visible and infrared video images, and the National Science
Foundation Career Award (2001) for a program investigating surface science
for vision and graphics.
GRASP
Laboratory
Site maintained by graspadm@grasp.cis.upenn.edu
Last update:
23 August, 2005