南京大学学报(自然科学版) ›› 2023, Vol. 59 ›› Issue (5): 752–758.doi: 10.13232/j.cnki.jnju.2023.05.003

• • 上一篇    

基于cGAN的下采样LG谱图像优化重建

叶皓1, 王麓懿2, 吴雪炜1(), 张勇2   

  1. 1.固体微结构物理国家重点实验室，南京大学物理学院，南京，210093
    2.南京大学现代工程与应用科学学院，南京，210023
  • 收稿日期:2023-07-20 出版日期:2023-09-30 发布日期:2023-10-13
  • 通讯作者: 吴雪炜 E-mail:wuxuewei@nju.edu.cn
  • 作者简介:第一联系人：共同第一作者，
  • 基金资助:
    中央高校基本业务费(021314380220);南京大学技术创新基金(020414913416)

Optimal image reconstruction from down⁃sampled LG spectrum based on cGAN

Hao Ye1, Luyi Wang2, Xuewei Wu1(), Yong Zhang2   

  1. 1.National Laboratory of Solid State Microstructures，School of Physics，Nanjing University，Nanjing，210093，China
    2.College of Engineering and Applied Sciences，Nanjing University，Nanjing，210023，China
  • Received:2023-07-20 Online:2023-09-30 Published:2023-10-13
  • Contact: Xuewei Wu E-mail:wuxuewei@nju.edu.cn

RichHTML

4

PDF (PC)

184

摘要：

对于复杂图像的拉盖尔高斯（Laguerre?Gaussian，LG）谱成像，因为满足奈奎斯特采样率的高阶LG模式系数无法测得，重建图像的失真不可避免，而神经网络算法通过先验学习，可以对失真图像实现较为清晰的复原.提出基于条件生成对抗网络（Conditional Generative Adversarial Nets，cGAN）的图像优化重建方法，在处理下采样的LG谱单像素成像和旋转运动模糊图像中均取得了较好的效果.在1.87%的LG谱采样率下，该方法能将Kaggle数据集人像二值图像的结构相似性（Structural Similarity，SSIM）指数提升至0.8以上，和经典图像去噪算法相比有显著提升.

关键词: 频谱下采样, 拉盖尔高斯模式, cGAN, 图像优化重建, 单像素成像, 旋转运动模糊图像复原

Abstract:

For Laguerre?Gaussian (LG) spectral imaging under Nyquist sampling rate，the reconstructed images are generally distorted because it is difficult to measure the higher?order LG mode coefficients. The neural network algorithm can be used to restore these images through prior learning. In this paper，we propose an optimal image reconstruction method based on Conditional Generative Adversarial Nets (cGAN)，which works well in down?sampled LG spectral single?pixel imaging and rotational motion blur recovery in LG spectral domain. We use the portrait binary images from Kaggle dataset as an example. At a sampling rate of 1.87%，the structural similarity (SSIM) index by using our method reaches 0.8 and above，which is significantly improved comparing with classical image denoising algorithms.

Key words: spectrum down sampling, Laguerre?Gaussian mode, cGAN, optimal image reconstruction, single?pixel imaging, rotational motion blurred images restoration

中图分类号: 

  • O439

图1

LG谱采样的单像素成像系统（a）；不同的LG模式截止阶数及其对应的重建图像：(b~c) l25 p25，(d~e) l50 p50和(f~g) l75 p75"

图2

静止原图(a)和旋转运动模糊图像，模糊角分别为5° （b），15° (c)和25° (d)"

图3

cGAN模型的结构图"

图4

人像图像LG谱下采样条件下使用MF，NLM，DAE和cGAN算法的重建效果"

图5

15°旋转模糊人像图像在LG谱下采样条件下使用MF，NLM，DAE，cGAN算法的重建效果"

表1

人像图像LG谱下采样的MF，NLM，DAE，cGAN方法重建图像的SSIM"

图号LG变换MFNLMDAEcGAN
4a0.4340.4540.5480.6540.841
4b0.6360.6530.7650.8500.932
4c0.6130.6310.7670.7260.886
4d0.5360.5540.6740.7380.868
4e0.5440.5610.6730.7590.884

表2

15°旋转模糊人像图像LG谱下采样的MF，NLM，DAE，cGAN方法重建图像的SSIM"

图号LG变换MFNLMDAEcGAN
5a0.3010.3070.3710.5110.825
5b0.2630.2690.3370.4790.858
5c0.3070.3130.3660.5080.822
5d0.2720.2790.3490.5300.834
5e0.3310.3380.4100.5850.866
1 Gonzalez R C， Woods R E. 数字图像处理. 阮秋琦，阮宇智译. 北京：电子工业出版社，2017：169-186.
2 Donoho D L. Compressed sensing. IEEE Transactions on Information Theory，2006，52(4)：1289-1306.
3 井花花，晏涛，刘渊. 融合全局和局部特征的光场图像空间超分辨率算法. 南京大学学报(自然科学)，2022，58(2)：298-308.
Jing H H， Yan T， Liu Y. A spatial super?resolution method for light filed images by fusing global and local features. Journal of Nanjing University (Natural Science)，2022，58(2)：298-308.
4 Dong C， Loy C C， He K M，et al. Image super?resolution using deep convolutional networks. IEEE Transactions on Pattern Analysis and Machine Intelligence，2016，38(2)：295-307.
5 Lyu M， Wang W， Wang H，et al. Deep?learning?based ghost imaging. Scientific Reports，2017(7)：17865.
6 Molina?Terriza G， Torres J P， Torner L. Manage?ment of the angular momentum of light：Preparation of photons in multidimensional vector states of angular momentum. Physical Review Letters，2001，88(1)：013601.
7 Torner L， Torres J P， Carrasco S. Digital spiral imaging. Optics Express，2005，13(3)：873-881.
8 Li F S， Xu T Z， Zhang W H，et al. Optical images rotation and reflection with engineered orbital angular momentum spectrum. Applied Physics Letters，2018，113(16)：161109.
9 Wang L Y， Ma J T， Xiao M，et al. Application of optical orbital angular momentum to rotation measurements. Results in Optics，2021(5)：100158.
10 Du J， Zhang R Z， Zhao Z，et al. Single?pixel identification of 2?dimensional objects by using complex Laguerre?Gaussian spectrum containing both azimuthal and radial modal indices. Optics Communications，2021(481)：126557.
11 Vu H Q. 105_classes_pins_dataset. https:∥www.kaggle.com/datasets/quangv/105?classes?pins?dataset，2022.
12 Ko S J， Lee Y H. Center weighted median filters and their applications to image enhancement. IEEE Transactions on Circuits and Systems，1991，38(9)：984-993.
13 Buades A， Coll B， Morel J M. A non?local algorithm for image denoising∥2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. San Diego，CA，USA：IEEE，2005：60-65，DOI：10.1109/CVPR.2005.38 .
14 Vincent P， Larochelle H， Bengio Y. Extracting and composing robust features with denoising autoencoders∥Proceedings of the 25th International Conference on Machine Learning. Helsinki，Finland：ACM，2008：1096-1103.
15 Goodfellow I J， Pouget?Abadie J， Mirza M，et al. Generative adversarial nets∥Proceedings of the 27th International Conference on Neural Information Processing Systems. Montreal，Canada：MIT Press，2014：2672-2680.
16 王坤峰，苟超，段艳杰，等. 生成式对抗网络GAN的研究进展与展望. 自动化学报，2017，43(3)：321-332.
Wang K F， Gou C， Duan Y J，et al. Generative adversarial networks：The state of the art and beyond. Acta Automatica Sinica，2017，43(3)：321-332.
17 Goodfellow I， Pouget?Abadie J， Mirza M，et al. Generative adversarial networks. Communications of the ACM，2020，63(11)：139-144.
18 Arjovsky M， Chintala S， Bottou L. Wasserstein generative adversarial networks∥Proceedings of the 34th International Conference on Machine Learning. Sydney，Australia：JMLR.org，2017：214-223.
19 陈文兵，管正雄，陈允杰. 基于条件生成式对抗网络的数据增强方法. 计算机应用，2018，38(11)：3305-3311.
Chen W B， Guan Z X， Chen Y J. Data augmentation method based on conditional generative adversarial net model. Journal of Computer Applications，2018，38(11)：3305-3311.
20 Mirza M， Osindero S. Conditional generative adversarial nets. 2014，arXiv:.
21 Zhang H， Sindagi V， Patel V M. Image de?raining using a conditional generative adversarial network. IEEE Transactions on Circuits and Systems for Video Technology，2020，30(11)：3943-3956.
22 Kingma D P， Ba J. Adam：A method for stochastic optimization. 2014，arXiv:.
23 曹志威，樊志杰，王青杨，等. 一种降噪自编码器的复杂网络链路预测算法. 小型微型计算机系统，2023，44(3)：665-672.
Cao Z W， Fan Z J， Wang Q Y，et al. Link prediction algorithm based on denoising autoencoder in complex networks. Journal of Chinese Computer Systems，2023，44(3)：665-672.
24 敦瑞静，鲁淑霞，张琦，等. 基于行列式点过程的变分拉普拉斯自编码器. 南京大学学报(自然科学)，2022，58(4)：629-639.
Dun R J， Lu S X， Zhang Q，et al. Variational Laplace autoencoders based on determinantal point process. Journal of Nanjing University (Natural Science)，2022，58(4)：629-639.
No related articles found!
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!

版权所有 © 2021 《南京大学学报(自然科学)》

地址：江苏省南京市鼓楼区汉口路22号，南京大学学报编辑部，210093

电话：025-83592704 , 传真：025-83592704        技术支持：北京玛格泰克科技发展有限公司