Imaging Spectroscopy and Spectral Unmixing

Antonio Robles-Kelly (NICTA)

COMPUTER VISION AND ROBOTICS SERIES

DATE: 2011-04-14
TIME: 16:00:00 - 17:00:00
LOCATION: NICTA - 7 London Circuit
CONTACT: JavaScript must be enabled to display this email address.

ABSTRACT:
In this seminar, I will introduce imaging spectroscopy, its capabilities and complexities. We will depart from the differences and similarities between imaging spectroscopy and trichromatic technologies. We will then look at the relationship between spectra and material identification and will explore the problem of unmixing. Along these lines, I will present how the generally ill-posed problem of recovering the constituent materials of objects in the scene can be addressed by re-stating it as that of clustering spectral radiance vectors at image pixels into dichromatic hyperplanes representing basis materials. I will explain how pixels can be decomposed into end-member materials in a manner similar to soft-clustering, where the features for clustering are the spectral radiance vectors and the abundance coefficients can be viewed as mixture weights in a maximum likelihood formulation.
BIO:
Antonio Robles-Kelly received the B.Eng. degree in electronics and telecommunications in 1998. In 2001 he received the William Gibbs/Plessey Award to the best research proposal and, in 2003, the Ph.D. degree in computer science from the University of York. He remained at the University of York until December 2004 as a Research Associate under the Mathematics for IT (MathFIT) initiative of the Engineering and Physical Sciences Research Council (EPSRC). In 2005, he took a research scientist appointment with NICTA.

He is also an adjunct Associate Professor at the ANU. He has been a Postdoctoral Research Fellow of the ARC and currently serves as a conjoint Senior Lecturer with the UNSW at the Australian Defense Force Academy. He is also an Associate Editor of the IET Computer Vision and Pattern Recognition journals. His research interests are in the areas of computer vision, statistical pattern recognition, computer vision and image processing. Dr. Robles-Kelly has been a Chair, co-chair, and technical committee member of mainstream computer-vision and pattern-recognition conferences.