46                                    

1.Registration-Based Segmentations with Shape Priors  

2.A Unifying Framework for n-Dimensional Quasi-Conformal Mappings

主讲人：张道平

摘要：1：As one of the most important tasks in image processing, Image registration aims tofind a plausible transformation to match the corresponding data, which has manyapplications in computer vision, remote sensing and medical imaging. In this talk, I willfirst introduce the registration model, especially the diffeomorphic registration model withquasi conformal theories. Then I will talk about the recent work, the registration-basedsegmentation method, including topology-preserving segmentation method,convexity-preserving segmentation method, star-shape-preserving segmentation methodand interactive segmentation.                                                           2：With the advancement of computer technology, there is a surge of interest ineffective mapping methods for objects in higher-dimensional spaces. To establish aone-to-one correspondence between objects, higher-dimensional quasi-conformaltheory can be utilized for ensuring the bijectivity of the mappings. In addition, it isoften desirable for the mappings to satisfy certain prescribed geometric constraintsand possess low distortion in conformality or volume. In this talk, we develop aunifying framework for computing n-dimensional quasi-conformal mappings. Morespecifically, we propose a variational model that integrates quasi-conformaldistortion, volumetric distortion, landmark correspondence, intensity mismatch, andvolume prior information to handle a large variety of deformation problems. Wefurther prove the existence of a minimizer for the proposed model and deviseefficient numerical methods to solve the optimization problem. We demonstrate theeffectiveness of the proposed framework using various experiments in two and threedimensions, with applications to medical image registration, adaptive remeshing, andshape modeling.

主讲人简介：张道平, 南开大学数学科学学院讲师, 分别于 2012 年和 2019年在大连理工大学和英国利物浦大学获得学士和博士学位。曾在香港中文大学从事博士后研究工作。2021 年加入南开大学数学科学学院。研究兴趣包括图像处理及其优化算法。近年来在国际图像科学及计算数学知名杂志比如 Siam Journal onImaging Sciences, Journal of Mathematical Imaging andVision, Numerical Algorithms, Journal of ScientificComputing, Applied Mathematical Modelling 等上发表多篇学术论文。

邀请人：刘海霞

时间：12月29日14:00-18：00

地点：腾讯会议室：398-763-196,388-707-016