光学学报, 2020, 40 (1): 0111003, 网络出版: 2020-01-06
深度学习下的计算成像:现状、挑战与未来 下载: 11290次特邀综述
Deep Learning Based Computational Imaging: Status, Challenges, and Future
图 & 表
图 2. 基于“目的与动机”对典型深度学习计算成像技术所作的分类
Fig. 2. Classification of typical deep learning based computational imaging techniques according to their objectives and motivations
图 4. 使用深度学习技术进行少图快速FPM成像[10]
Fig. 4. Fast FPM imaging with few images using deep learning technology[10]
图 5. 基于深度学习的条纹分析方法原理与相位重构结果对比[37]。(a)基于深度学习的条纹分析方法原理图; (b)傅里叶变换法重构结果;(c)加窗傅里叶变换法重构结果;(d)深度学习法重构结果;(e) 12步相移法重构结果
Fig. 5. Principle of fringe analysis method based on deep learning and comparison of phase reconstruction results[37]. (a)Principle of fringe analysis method based on deep learning; (b) reconstruction result of FT; (c) reconstruction result of WFT; (d) reconstruction result of proposed deep-learning method; (e) reconstruction result of 12-step phase-shifting profilometry
图 7. 基于深度学习进行散射介质成像的网络原理图[28]
Fig. 7. Network of deep learning based imaging through scattering medium[28]
图 8. 基于深度学习进行三维衍射层析重建的基本框图[26]
Fig. 8. Basic framework of 3D diffraction tomography reconstruction based on deep learning[26]
图 9. 基于深度学习进行光学衍射层析的网络原理图[27]
Fig. 9. Schematic of network of optical diffraction tomography based on deep learning[27]
图 10. 使用深度神经网络的数字全息离焦距离计算框架[6]
Fig. 10. Framework of defocusing distance calculation in digital holography based on deep neural network[6]
图 11. 针对视网膜光学相干断层图像的边界自动分割原理图[23]
Fig. 11. Schematic of automatic boundary segmentation framework for retinal OCT image[23]
图 12. 基于深度学习进行超分辨率成像的网络框架示意图[20]
Fig. 12. Network framework of super-resolution imaging based on deep learning[20]
图 13. 基于深度学习进行STED超分辨率成像的实验结果[17]
Fig. 13. Experimental results of STED super-resolution imaging based on deep learning[17]
图 14. 基于深度学习进行极弱光成像的结果[34]。(a)摄像机输出(ISO 8000);(b)摄像机输出(ISO 409600);(c)由原始数据(a)恢复得到的结果
Fig. 14. Results of imaging using very weak light based on deep learning[34]. (a) Camera output with ISO 8000; (b) Camera output with ISO 409600; (c) recovered result from raw data of Fig. 14 (a)
图 15. 基于深度学习进行虚拟染色成像的网络框架示意图[35]
Fig. 15. Network framework of virtual staining imaging based on deep learning[35]
图 17. 物理(左)和图像分类(右)关联的因果层次结构[99]
Fig. 17. Causal hierarchy structure relevant to physics (left) and image classification (right)[99]
图 18. 在熊猫图片中加入轻微随机噪声,CNN模型将图片识别为长臂猿[104]
Fig. 18. After adding slight noise into Panda image, CNN model recognizes image as Gibbon[104]
图 19. 深度学习与经典理论算法之间的客观公证对比
Fig. 19. Comparison between deep learning and classical theoretical algorithm should be objective
左超, 冯世杰, 张翔宇, 韩静, 陈钱. 深度学习下的计算成像:现状、挑战与未来[J]. 光学学报, 2020, 40(1): 0111003. Chao Zuo, Shijie Feng, Xiangyu Zhang, Jing Han, Chen Qian. Deep Learning Based Computational Imaging: Status, Challenges, and Future[J]. Acta Optica Sinica, 2020, 40(1): 0111003.