AI Seminar: Diffeomorphic curve matching with finite elements
Andreas Bock, PhD student, Imperial College London.
We study closed curve matching where the goal is to find a suitable curve in the space of diffeomorphisms bringing into alignment two closed curves, called the template and the target. We treat the case where we have a full parameterisation of the template, but target is not fully observed and is only represented by a finite number of point observations. The matching problem must therefore be independent of the parameterisation. We present a finite element discretisation of Hamilton's equations (determining the velocity fields that induces the diffeomorphism) and apply a Kalman filter to determine suitable parameterisations. We show some preliminary numerics and highlight some open problems.
Andreas Bock is a PhD student at Imperial College London working with Professor Colin Cotter. His research interests include inverse problems in shape analysis and applications of finite element analysis. Before studying at Imperial College he completed an MSc and BSc from DIKU.
The seminar is free and open for everyone.
This seminar is a part of the AI Seminar Series organised by SCIENCE AI Centre. The series highlights advances and challenges in research within Machine Learning, Data Science, and AI. Like the AI Centre itself, the seminar series has a broad scope, covering both new methodological contributions, ground-breaking applications, and impacts on society.