典型文献
Snapshot spectral compressive imaging reconstruction using convolution and contextual Transformer
文献摘要：
Spectral compressive imaging (SCI) is able to encode a high-dimensional hyperspectral image into a two-dimensional snapshot measurement, and then use algorithms to reconstruct the spatio-spectral data-cube. At present, the main bottleneck of SCI is the reconstruction algorithm, and state-of-the-art (SOTA) reconstruction methods generally face problems of long reconstruction times and/or poor detail recovery. In this paper, we propose a hybrid network module, namely, a convolution and contextual Transformer (CCoT) block, that can simultaneously acquire the inductive bias ability of convolution and the powerful modeling ability of Transformer, which is conducive to improving the quality of reconstruction to restore fine details. We integrate the proposed CCoT block into a physics-driven deep unfolding framework based on the generalized alternating projection (GAP) algorithm, and further propose the GAP-CCoT network. Finally, we apply the GAP-CCoT algorithm to SCI reconstruction. Through experiments on a large amount of synthetic data and real data, our proposed model achieves higher reconstruction quality (>2 dB in peak signal-to-noise ratio on simulated benchmark datasets) and a shorter running time than existing SOTA algorithms by a large margin. The code and models are publicly available at .
文献关键词：
中图分类号：
作者姓名：
Lishun Wang;Zongliang Wu;Yong Zhong;Xin Yuan
作者机构：
Chengdu Institute of Computer Application, Chinese Academy of Sciences, Chengdu 610041, China;University of Chinese Academy of Sciences, Beijing 100049, China;Research Center for Industries of the Future and School of Engineering, Westlake University, Hangzhou 310030, China;e-mail: zhongyong@casit.com.cn;e-mail: xyuan@westlake.edu.cn
文献出处：
引用格式：
[1]Lishun Wang;Zongliang Wu;Yong Zhong;Xin Yuan-.Snapshot spectral compressive imaging reconstruction using convolution and contextual Transformer)[J].光子学研究（英文）,2022(08):1848
A类：
CCoT
B类：
Snapshot,compressive,imaging,reconstruction,using,convolution,contextual,Transformer,Spectral,SCI,encode,dimensional,hyperspectral,image,into,snapshot,measurement,then,use,algorithms,spatio,cube,At,present,main,bottleneck,state,art,SOTA,methods,generally,face,problems,long,times,poor,recovery,In,this,paper,hybrid,network,module,namely,block,that,can,simultaneously,acquire,inductive,bias,ability,powerful,modeling,which,conducive,improving,quality,restore,fine,details,We,integrate,proposed,physics,driven,deep,unfolding,framework,generalized,alternating,projection,GAP,further,Finally,apply,Through,experiments,large,amount,synthetic,real,our,achieves,higher,dB,peak,signal,noise,ratio,simulated,benchmark,datasets,shorter,running,than,existing,by,margin,models,are,publicly,available
AB值：
0.581792
相似文献
机标中图分类号，由域田数据科技根据网络公开资料自动分析生成，仅供学习研究参考。
