基本信息
浏览量:1744
职业迁徙
个人简介
My main interest is in sequential and parallel algorithm design. Of particular interest are problems that arise in scientific computation and image processing. We have been working on three classes of problems. Our work is both more theoretical yet more practical since we require two important properties of our algorithms: they should be both be fast and have strong guarantees of quality, size, and speed.
Mesh Generation The question of correctly and efficiently partitioning space into tetrahedra with good aspect ratio so that the features appear in the mesh is at least a fifty year old problem. The computer science community has been working on this problem for about twenty years. The most accurate simulations in the science, engineering, and graphic require small size quality meshes. CMU has been a leader on this problem for the last fifteen years.
Spectral Graph Theory The interplay between graph theory and linear algebra is possibly one of the most interesting and practical areas in modern algorithm design. Possibly the most famous example is Google PAGE-RANK which is the Perron-Frobenius eigenvector of the link graph. The second example is the interplay between fast linear solvers and graph theory. We have ongoing collaboration on fast solvers and eigen calculations. These new algorithm work in near linear time.
Image Processing We are interested in fast and reliable algorithm for image processing. Of special interest is 3D medical image processing. Our main tool that allows us to get new fast algorithms is spectral graph theory. This has allowed us to compute eigenvectors of our image as speed comparable to many simple filtering algorithms. We are also using interior point methods from convex optimization to de-noise images. Again these methods all use our fast solvers for speed.
Mesh Generation The question of correctly and efficiently partitioning space into tetrahedra with good aspect ratio so that the features appear in the mesh is at least a fifty year old problem. The computer science community has been working on this problem for about twenty years. The most accurate simulations in the science, engineering, and graphic require small size quality meshes. CMU has been a leader on this problem for the last fifteen years.
Spectral Graph Theory The interplay between graph theory and linear algebra is possibly one of the most interesting and practical areas in modern algorithm design. Possibly the most famous example is Google PAGE-RANK which is the Perron-Frobenius eigenvector of the link graph. The second example is the interplay between fast linear solvers and graph theory. We have ongoing collaboration on fast solvers and eigen calculations. These new algorithm work in near linear time.
Image Processing We are interested in fast and reliable algorithm for image processing. Of special interest is 3D medical image processing. Our main tool that allows us to get new fast algorithms is spectral graph theory. This has allowed us to compute eigenvectors of our image as speed comparable to many simple filtering algorithms. We are also using interior point methods from convex optimization to de-noise images. Again these methods all use our fast solvers for speed.
研究兴趣
论文共 320 篇作者统计合作学者相似作者
按年份排序按引用量排序主题筛选期刊级别筛选合作者筛选合作机构筛选
时间
引用量
主题
期刊级别
合作者
合作机构
arxiv(2021)
引用0浏览0EI引用
0
0
user-5e9d449e4c775e765d44d7c9(2020)
引用0浏览0引用
0
0
user-6144298de55422cecdaf68a5(2020)
引用0浏览0引用
0
0
user-5e9d449e4c775e765d44d7c9(2020)
引用0浏览0引用
0
0
arXiv (Cornell University) (2020)
引用0浏览0引用
0
0
加载更多
作者统计
合作学者
合作机构
D-Core
- 合作者
- 学生
- 导师
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn