Solving inverse problems with Iterative regularisation and convex penalties

使用迭代正则化和凸惩罚解决反问题

• 批准号: DP150102345
• 负责人: Dr Qinian Jin
• 金额: $ 15.59万
• 依托单位: The Australian National University
• 依托单位国家: 澳大利亚
• 项目类别: Discovery Projects
• 财政年份: 2015
• 资助国家: 澳大利亚
• 起止时间: 2015-01-01 至 2021-12-31
• 项目状态: 已结题
• 来源: https://dataportal.arc.gov.au/RGS/Web/Grants/DP150102345
• 关键词: Solving; inverse; problems; Iterative; regularisation

This project aims to develop and investigate new computational procedures for the solution of inverse problems which do not have the usual smoothness properties (or source conditions) required for the traditional regularisation methods. Examples of such inverse problems are very common and include image restoration, photo-acoustic tomography and spectroscopy. It is anticipated that this project will substantially extend the toolbox of methods for such problems utilising ideas from Banach spaces, convex analysis, parallel computing and optimisation. This project is expected to make a substantial contribution to a better understanding of inverse problems and their solution procedures.

该项目旨在开发和研究新的计算程序，用于解决不具有传统正则化方法所需的通常光滑特性（或源条件）的反问题。这样的逆问题的例子是非常常见的，包括图像恢复，光声层析成像和光谱。预计该项目将大大扩展工具箱的方法，利用这些问题的想法从Banach空间，凸分析，并行计算和优化。该项目有望为更好地理解反问题及其求解程序做出实质性贡献。