基于深度学习的图像压缩感知算法综述
首发时间:2021-01-22
摘要:近年来,压缩感知理论突破了Nyquist采样频率的限制,为信号处理领域带来了革命性的变化。随着各领域数据的海量增长,压缩感知理论可以极大程度上缓解大量图像数据的获取与传输带来的硬件压力,因此,探索压缩感知的概念及其在图像处理领域中的应用非常重要且必要。本文讨论了压缩感知的基本概念以及在压缩感知中结合深度学习方法算法的优势,介绍了典型模型的网络结构与特点,并介绍了提高方法硬件适用性的离散化采样矩阵压缩感知方法研究进展。最后分析与总结了基于深度学的图像压缩感知算法领域的研究趋势。
For information in English, please click here
Summary of Image Compressed Sensing Algorithms Based on Deep Learning
Abstract:In recent years, compressed sensing theory has broken through the limitation of Nyquist sampling frequency and brought revolutionary changes to the field of signal processing. With the massive growth of data in various fields, compressed sensing theory can greatly ease the hardware pressure caused by the acquisition and transmission of large amounts of image data. Therefore, it is very important to explore the concept of compressed sensing and its application in the field of image processing. necessary. This article discusses the basic concepts of compressed sensing and the advantages of combining deep learning algorithms in compressed sensing, introduces the network structure and characteristics of typical models, and introduces the research progress of discrete sampling matrix compressed sensing methods to improve the applicability of the method hardware. Finally, it analyzes and summarizes the research trends in the field of image compression sensing algorithms based on depth science.
Keywords: compressed sensing deep learning sampling matrix reconstruction algorithm
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
基于深度学习的图像压缩感知算法综述
评论
全部评论