Abstract:
Finding efficient algorithms to describe, measure and compare shapes is a central problem in numerous disciplines that generate extensive quantitative and visual information. Among these, biology occupies a central place. Registration of brain anatomy for example is essential to many studies in neurobiology; at a molecular level, comparison of protein shapes is a key step in understanding the relationships between their functions. In this talk I will introduce the idea of a globally optimal conformal mapping between two (discrete) surfaces of genus zero as one method to solve this problem. In this approach, the whole mesh representing the source surface is warped onto the target surface, using the mapping defined through the composition of discrete conformal mappings of the surfaces onto the sphere and the Mobius transformation between these mappings. The Mobius transformation is then optimized to lead to minimal distortion between the source mesh and its image, where distortion is measured as difference from isometry. I will show that this approach leads to the definition of a metric in the space of genus-zero surfaces. I will describe the implementation of this approach into a software and its applications on biological examples, from brain surface matching to 3D morphometrics on bones of primates.

Biography:
I was born and grew up in France. After graduation from the Ecole Centrale de Paris, a higher education establishment for engineers, I completed a PhD program in molecular biology and biophysics at the University Louis Pasteur of Strasbourg, France. The same year, I was appointed staff scientist of the CNRS, and joined the biological NMR laboratory at the University Louis Pasteur of Strasbourg, France. In 1997, I came for a sabbatical to Stanford University. I liked California so much that I decided to stay: in 2004, I joined the University of California, Davis, with a joint appointment as Professor in the department of Computer Science and member of its Genome Center.

Host:
Dr. Yiying Tong