||Confirmation of PhD Candidature: A new convexity prior for n-dimensional shapes and its applications.|
||LI Lingfeng (李凌丰), Department of Mathematics, Hong Kong Baptist University, HKSAR|
||10:30 - 12:00|
FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
| || |
In this work, we present a prior for representing convex shapes
using level-set methods. Compared to other methods in the literature,
the main advantage of this prior is that it works not only for
2 dimension but also higher dimensions. To test our new prior,
we apply it to two different types of models, which are convex
hull models and segmentation models, and design appropriate algorithms.
The algorithms are based on the Alternating direction methods
of multipliers and the convexity constraint can be handled easily
by a projection. We then conduct numerical experiments for these
two models on 2-D and 3-D cases. The results show that our convexity
prior can characterize the convex shapes very accurately.