王小强:Edge-Weighted Centroidal Voronoi Tessellation Based Algorithms for Image Segmentation
发布者: 崔琪
发布时间:2020-12-04
浏览次数:665

要:

Centroidal Voronoi tessellations (CVTs) are special Voronoi tessellations having the property that the generators of the Voronoi tessellation are also the centers of mass. The application of CVT to the Image Segmentation is the traditional k-means algorithm. Here we extend the CVT method to the Edge-Weithed CVT method. The new algorithm produces segmentations that minimize a total energy: a sum of the classic CVT energy and the weighted length of cluster boundaries. The EWCVT method is easy in implementation, fast in computation, and natural for any number of clusters.


报告人:

王小强,Florida State University科学计算系教授,1995年毕业于武汉大学,1998年毕业于中科院数学所,获得硕士学位,2005年毕业于Pennsylvania State University,获理学博士学位,主要从事数学生物、大数据与数据挖掘、图像处理和高性能科学计算等领域的研究,先后在《Journal of Computational Physics》、《Physica D》、《SIAM J. Math. Anal.》及《IEEE Transactions on Image Processing》等数学领域顶尖期刊发表科研论文30余篇。


间:2020年12月9日周三上9:00-10:00

腾讯会议:626 995 970     密码2020


欢迎广大师生参加!


数学科学学院

姓            名:

王小强:Edge-Weighted Centroidal Voronoi Tessellation Based Algorithms for Image Segmentation

职            称:

邮            箱:

办     公     室:

办 公 室 电 话:

研  究  方  向:

要:

Centroidal Voronoi tessellations (CVTs) are special Voronoi tessellations having the property that the generators of the Voronoi tessellation are also the centers of mass. The application of CVT to the Image Segmentation is the traditional k-means algorithm. Here we extend the CVT method to the Edge-Weithed CVT method. The new algorithm produces segmentations that minimize a total energy: a sum of the classic CVT energy and the weighted length of cluster boundaries. The EWCVT method is easy in implementation, fast in computation, and natural for any number of clusters.


报告人:

王小强,Florida State University科学计算系教授,1995年毕业于武汉大学,1998年毕业于中科院数学所,获得硕士学位,2005年毕业于Pennsylvania State University,获理学博士学位,主要从事数学生物、大数据与数据挖掘、图像处理和高性能科学计算等领域的研究,先后在《Journal of Computational Physics》、《Physica D》、《SIAM J. Math. Anal.》及《IEEE Transactions on Image Processing》等数学领域顶尖期刊发表科研论文30余篇。


间:2020年12月9日周三上9:00-10:00

腾讯会议:626 995 970     密码2020


欢迎广大师生参加!


地址:青岛市崂山区松岭路238号   
电话:0532-66787153
邮编:266100
版权所有©中国海洋大学     数学科学学院     鲁ICP备05002467号-1     鲁公网安备 37021202000030号

官方微信

Baidu
map
.