Wednesday, Oct 1st, 11:00, in BC 329
First steps in Goal-driven Appearance Fabrication
We are entering an era where fabrication of various objects can be done fully automatically. The fabrication devices are becoming affordable for home use, and are widely accessible through online services.
Traditionally the input for such devices is some geometry and material specification. On the other hand, users usually have in mind higher level appearance or mechanical goals. Many of these goals can be achieved in theory, but in practice designing the objects using traditional CAD tools can be infeasible.
My research focus is on the appearance aspect of fabrication, and my long term research goal is to expand the gamut of possible fabricated objects by designing novel methods that allow high quality Goal-driven Appearance Fabrication.
In this talk I will present a few examples of methods that we developed. One such example is an early method for automatic generation of lenses which create a desired caustic when illuminated. Another example is a method for automatically defining mixing ratios of pigments which allows fabrication with desired color and translucency.
Marios Papas is currently a 4th year Ph.D. student at the Computer Graphics Laboratory of ETH Zurich. He received his B.Sc. in 2007 from the University of Cyprus and his M.Sc. in 2009 from University of California San Diego, both in Computer Science. Marios masters was funded by Fulbright and his P.hD. by Disney Research.
He has published in the areas of Appearance Capture and Fabrication. During his M.Sc. he developed and validated a physically based analytic model for modeling the appearance of rough, highly scattering, thin paper-like materials under the guidance of Dr. Henrik Wann Jensen. During his Ph.D., under the guidance of Dr. Markus Gross and Dr. Wojciech Jarosz, he developed methods and apparatus for capturing and fabricating the appearance of homogeneous translucent materials. In addition he developed a method for controlling light interaction with lenses, which allows for automated generation of the lens geometry.