Abstract. Image segmentation separates the image into different regions to simplify the image and identify the objects easily. Image segmentation has been extensively studied, and Mumford-Shah and Chan-Vese model are one of the most well-known variational model in the field. This talk will focus on the multiphase aspect of such problem, in particular unsupervised model and its extensions to various applications. The connection to data clustering will be considered, scale segmentation and application to 3D flash Lidar image will presented. This talk will also include a model segmenting piecewise constant images with irregular object boundaries.

15:45 - 16:15 Uhr: coffee and tea

Subsequently, M. Zuhair Nashed will give a talk on Conductivity Imaging from Interior Data and Related Nonsmooth Optimization Problems.

The official invitation for both talks can be found here.