期刊文献+

联合全局与局部特征的深度压缩感知图像重构

Image reconstruction based on deep compressive sensing combined with global and local features
下载PDF
导出
摘要 从极少量的测量值中有效且高概率高质量恢复出原始信号是压缩感知图像重建研究的核心问题,学者们相继提出了传统和基于深度学习的压缩感知图像重建算法,传统算法通常基于优化模型迭代求解,重建质量和重建速度都无法保证;基于深度学习的算法重建质量相对较高,但缺乏物理可解释性。受滤波流的启发,本文提出了联合全局与局部的深度压缩感知图像重建模型(G2LNet),其以卷积层执行压缩采样以及初始重建过程,利用快速傅里叶卷积与滤波流,同时考虑了图像全局上下文信息和图像像素局部邻域信息,联合学习优化测量矩阵与滤波流,建立了完整的端到端可训练的深度图像压缩感知重建网络。经实验验证,在压缩感知图像重建领域常用的Set5,Set11,BSD68测试集上取得了良好的重建效果,在采样率为20%的情况下,G2LNet的图像重建质量相比于经典的传统算法MH与基于深度学习的算法CS⁃Net的平均PSNR分别提高了2.29 dB,0.51 dB,有效提升了重建图像质量。 Effectively restoring the original signal with high probability and high quality from a very small number of measured values is the core issue of compressive sensing for image reconstruction.Researchers have successively proposed traditional and deep learning-based compressive sensing image reconstruction algorithms.The traditional algorithms are based on mathematical derivation.Although they are compre⁃hensible,their reconstruction quality is relatively poor.On the contrary,deep learning-based algorithms have relatively high reconstruction quality,but they cannot guarantee intelligibility.Inspired by filter flow,this study proposes a global-to-local compressive sensing image reconstruction model called G2LNet,which performs compressed sampling and initial reconstruction processes with convolutional layers using fast Fourier convolution and convolutional filter flow,taking into account the global contextual information of the image and local neighborhood information of the image pixel simultaneously.It learns to jointly opti⁃mize the measurement matrix and convolution filter flow and establishes a complete end-to-end trainable deep image reconstruction network.Verification experiments were performed on the Set5,Set11,and BSD68 test datasets commonly used in the field of compressive sensing image reconstruction at a 20%sampling rate.The image reconstruction quality of G2LNet was compared with that of the traditional algo⁃rithm MH and algorithm based on deep learning;the average peak signal-to-noise ratio of CSNet increased by 2.29 dB and 0.51 dB,respectively,effectively improving the quality of the reconstructed image.
作者 仲元红 徐乾锋 周宇杰 王姗姗 ZHONG Yuanhong;XU Qianfeng;ZHOU Yujie;WANG Shanshan(School of Microelectronics and Communication Engineering,Chongqing University,Chongqing 400044,China;Institute of Physical Science and Information Technology,Anhui University,Hefei 230039,China)
出处 《光学精密工程》 EI CAS CSCD 北大核心 2023年第14期2135-2146,共12页 Optics and Precision Engineering
基金 国家自然科学基金资助项目(No.61501069) 重庆市技术创新与应用发展面上项目资助(No.cstc2019jscx-msxmX0167)。
关键词 压缩感知 图像重建 快速傅里叶卷积 滤波流 深度神经网络 compressive sensing image reconstruction fast Fourier convolution filter flow deep neural network
  • 相关文献

参考文献2

二级参考文献11

共引文献2

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部