2. 您的位置
3. 首页
4. 会议论文
5. Disparity and Optical Flow Partitioning Using Extended Potts Priors

Disparity and Optical Flow Partitioning Using Extended Potts Priors

来源 :第八届工业与应用数学国际大会 | 被引量 : 0次 | 上传用户：ellen719420908

【摘 要】

：

　　This talk addresses the problems of disparity and optical flow partitioning.We investigate new variational approaches to these problems using Potts priors a

【作 者】

：

Fitschen Jan Henrik 

【机 构】

：

Univ.of Kaiserslautern

【出 处】

：

第八届工业与应用数学国际大会

【发表日期】

：

2015年期

下载到本地 , 更方便阅读

下载此文 赞助VIP

声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架

论文部分内容阅读

　　This talk addresses the problems of disparity and optical flow partitioning.We investigate new variational approaches to these problems using Potts priors and propose a modified alternating direction method of multipliers.

其他文献

Iterative Minimization Formulation and Algorithm for Transition State Calculation

　　Calculation of transition states on energy landscape is very important in computational chemistry,material sciences,etc.

会议

Nonlinear Subdivision Schemes and Applications

　　Subdivision schemes and their associated multiscale algorithms have led,over the past 20 years,to important breakthroughs in scientific computing including

会议

Statistical Mechanics of Random Walks,with Application to Molecular Evolution

　　Understanding transport in complex media is crucial for many areas of science.

会议

Rare Events and Energy Landscape for the Chemical Reaction Kinetics

　　Rare event study attracts much attention in the chemical reaction kinetics recent years.

会议

Stochastic Surface Walking Method for Global Structure Search and Pathway Sampling

　　The algorithm and the extension of stochastic surface walking method will be overviewed.

会议

Rare Events in Complex Physical System

　　Many problems arising from chemistry,physics and materials science involve rare but significant exit events and/or transition events between stable states.

会议

Multiscale Adaptive Learning Algorithms for High-dimensional Data

　　Many data sets in image analysis are in a high-dimensional space but exhibit a low-dimensional structure.

会议

Edge-preserving Regularization for Electrical Impedance Tomography

　　The D-bar method provides a non-iterative solution to the full nonlinear problem of Electrical Impedance Tomography,based on a low-pass filter in the(nonlin

会议

Rates for Coordinate and Block-descent Algorithms

　　This work is based on very simple(old)observations relating Dykstra's-like alternating minimization algorithms and forward-backward splitting.

会议

Regularization of Inverse Problems in Imaging Sciences:Theoretical and Numerical Aspects

　　Inverse problems have become a central theme in various fields of sciences and engineering such as imaging sciences.

会议