arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.09026eess.IVcs.ITmath.IT
收藏

基于算子的信息理论在成像中的应用:熵、容量与物理测量系统中的不可逆性

Operator-Based Information Theory for Imaging: Entropy, Capacity, and Irreversibility in Physical Measurement Systems

Future Technology Centre, School of Electrical and Mechanical Engineering, University of Portsmouth(朴茨茅斯大学电气与机械工程学院未来技术中心)

Charles Wood

AI总结:

本文提出基于算子的信息理论框架,用于分析成像系统中的熵、容量和不可逆性,为理解物理限制和后续算法发展提供基础。

摘要:

成像系统通常用分辨率、对比度和信噪比来描述,但这些量并不能全面说明物理转换如何影响信息的流动。本文介绍了用于成像的信息理论的算子方法。该方法将成像链建模为对函数作用的有界算子的组合,并利用这些算子的谱性质来表征信息的重新分布。三种度量被开发出来。算子熵量化算子如何将能量分布到其奇异谱上。算子信息容量描述在噪声依赖的阈值之上可恢复的模式数量。不可逆性指数衡量通过抑制或消除模式所丢失的信息,并捕捉在算子组合下的信息损失积累。该框架适用于线性、非线性和随机算子,并不依赖于特定的成像模式。分析示例展示了衰减、模糊和采样如何以不同方式影响熵、容量和不可逆性。结果提供了一个分析成像物理限制的一般结构,并为后续的信息几何、时空预算、非线性信道和重建算法的工作奠定了基础。

英文摘要:

Imaging systems are commonly described using resolution, contrast, and signal-to-noise ratio, but these quantities do not provide a general account of how physical transformations affect the flow of information. This paper introduces an operator-based formulation of information theory for imaging. The approach models the imaging chain as a composition of bounded operators acting on functions, and characterises information redistribution using the spectral properties of these operators. Three measures are developed. Operator entropy quantifies how an operator distributes energy across its singular spectrum. Operator information capacity describes the number of modes that remain recoverable above a noise-dependent threshold. An irreversibility index measures the information lost through suppression or elimination of modes and captures the accumulation of information loss under operator composition. The framework applies to linear, nonlinear, and stochastic operators and does not depend on the specific imaging modality. Analytical examples show how attenuation, blur, and sampling affect entropy, capacity, and irreversibility in different ways. The results provide a general structure for analysing the physical limits of imaging and form the basis for subsequent work on information geometry, spatiotemporal budgets, nonlinear channels, and reconstruction algorithms.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑