基于混沌压缩感知和深度学习网络的压缩感知新模型

打开文本图片集

摘要：文章提出了一种基于混沌压缩感知和深度学习网络的压缩感知新模型，称为混沌深度压缩感知模型。该模型将传统压缩感知中的迭代步骤转化为深度网络形式，并将相关混沌参数应用于测量矩阵生成和深度网络训练过程。混沌深度网络中的所有参数都将通过程序自动学习获取，不再需要人工设计。

关键词：压缩感知；深度学习；神经网络；混沌理论

doi：10.3969/J.ISSN.1672-7274.2023.02.032

中图分类号：TP 391.41               文献标示码：A               文章编码：1672-7274（2023）02-00-03

A New Compressed Sensing Model Based on Chaotic Compressed Sensing and Deep Learning Network

CHEN Yixin， MA Zeng

（Basic Department of Naval Submarine Academy， Qingdao 266000， China）

Abstract： A new compressed sensing model based on chaotic compressed sensing and deep learning network is proposed， which is called chaotic deep compressed sensing. The iterative steps in traditional compressed sensing are transformed into deep network form， and the relevant chaotic parameters are applied to the measurement matrix generation and deep network training process. All parameters in the chaotic depth network will be acquired automatically through program learning， and no manual design is required.

Key words： compressed sensing; deep learning; neural network; chaos theory

0  引言

压缩感知理论证明，当一个信号在某些变换域表现出稀疏性时，它能够以较高的概率使用比奈奎斯特抽样理论所确定的少得多的测量值重构[1]。（剩余3722字）