摘要
We propose a deep learning denoising computational ghost imaging (CGI) method to obtain a clear object image with a sub-Nyquist sampling ratio. We develop an end-to-end deep neural network (DDANet) for CGI image reconstruction. DDANet uses a one-dimensional (1-D) bucket signals (BSs) and multiple tunable noise-level maps as input, and outputs a clear image. We train DDANet with simulated BSs and ground-truth pairs, and then retrieve the object image directly from an experimental obtained 1-D BSs. The effectiveness of the proposed method is experimentally investigated. The proposed method has practical applications in image denoising and enhancement of the CGI and single-pixel computational imaging.
学科与关键词
关键词
Computational ghost imagingDeep learningSub-Nyquist samplingImage denoising
- Sub-Nyquist computational ghost imaging with deep learningWu, Heng; Wang, Ruizhou*; Zhao, Genping; Xiao, Huapan; Wang, Daodang; Liang, Jian; Tian, Xiaobo; Cheng, Lianglun; Zhang, XianminOptics Express, 2020, 28(3): 3846-3853.被引 49
- Deep-learning based multi-scale computational ghost imaging for high-performance complex image recoveryZhu, Yiming; Zhang, Xizheng; Xu, Chenxiang; Liu, Bingyi; Guo, Kai; Yin, Zhiping; Guo, Zhongyi*Optics Communications, 2024, 554: 129916.
- Computational ghost imaging with 4-step iterative rank minimizationWu, Heng; Wu, Wenhao; Chen, Meiyun*; Luo, Shaojuan*; Zhao, Ruitong; Xu, Limin; Xiao, Huapan; Cheng, Lianglun; Zhang, Xianmin; Xu, YipingPhysics Letters, Section A: General, Atomic and Solid State Physics , 2021, 394: 127199.被引 4
