基于生成对抗网络的低光照图像增强算法

doi:10.12052/gdutxb.220039

摘要/Abstract

摘要： 传统的基于深度学习的方法在低照度图像增强中已经有比较好的发挥，但是这些方法通常需要成对的数据集进行训练，而相对应的数据集正是目前难以收集的。目前的增强方法在真实的低照度图像增强中也会产生增强效果不完美和出现图像噪声等问题。针对这些问题，设计了无监督生成对抗网络，使其可以不用配对训练数据集进行训练，并且把网络分解为注意力机制网络和增强网络2个子网络。通过注意力机制网络把低照度图像中的低光区域和亮光区域区分开，并使用残差增强网络结合全局局部判别器，对图像进行增强。实验结果表明，本文的方法在低光照图像增强方面优于Enlighten-GAN、Cycle-GAN等方法。

关键词: 生成对抗网络, 低照度图像增强, 注意力机制

Abstract: Traditional deep learning-based methods have achieved promising performance for low-illumination image enhancement. However, these methods usually need to be trained on the pair-wise datasets, which are difficult to collect. Moreover, most existing enhancement methods have the problems of imperfect enhancement effect and image noise in real low illumination image enhancement. To address this, a unsupervised generative adversarial network is designed for low-illumination image enhancement, which has no requirement of training on the pair-wise datasets. The proposed network consists of two subnetworks: attentional mechanism network and enhancement network. The attentional mechanism network is used to distinguish the low-light region from the bright region of the low-illumination image, and the residual enhancement network is used to enhance the image by combining with the global-local discriminator. By doing this, a low-illumination image can be well enhanced. Extensive experimental results show that the proposed method outperforms the baseline Enlighten-GAN and Cycle-GAN for low-light image enhancement.

Key words: generative adversarial network, low-light image enhancement, attentional mechanism

中图分类号:

TP391

杨镇雄, 谭台哲. 基于生成对抗网络的低光照图像增强算法[J]. 广东工业大学学报, 2024, 41(01): 55-62.

Yang Zhen-xiong, Tan Tai-zhe. Low Illumination Image Enhancement Algorithm Based on Generative Adversarial Network[J]. Journal of Guangdong University of Technology, 2024, 41(01): 55-62.

参考文献

[1] ZHOU Z, FENG Z, LIU J, et al. Single-image low-light enhancement via generating and fusing multiple sources [J]. Neural Computing and Applications, 2020, 32: 6455-6465.
[2] GHARBI M, CHEN J, BARRON J T, et al. Deep bilateral learning for real-time image enhancement [J]. ACM Transactions on Graphics (TOG), 2017, 36(4): 118.
[3] GUO X, LI Y, LING H. LIME: Low-light image enhancement via illumination map estimation [J]. IEEE Transactions on Image Processing, 2016, 26(2): 982-993.
[4] LORE K G, AKINTAYO A, SARKAR S. LLNet: a deep autoencoder approach to natural low-light image enhancement [J]. Pattern Recognition, 2017, 61: 650-662.
[5] SHEN L , YUE Z , FENG F , et al. MSR-net: low-light image enhancement using deep convolutional network[EB/OL]. arXiv: 1711.02488 (2017-11-07) [2022-03-25]. https://doi.org/10.48550/arXiv.1711.02488.
[6] WEI C , WANG W , YANG W, et al. Deep retinex decomposition for low-light enhancement[EB/OL]. arXiv: 1808.04560 (2018-08-14) [2022-03-25]. https://doi.org/10.48550/arXiv.1808.04560.
[7] CHEN C , CHEN Q , XU J , et al. Learning to see in the dark[EB/OL]. arXiv: 1805.01934 (2018-05-04) [2022-03-25]. https://doi.org/10.48550/arXiv.1805.01934.
[8] KALANTARI N K, RAMAMOORTHI R. Deep high dynamic range imaging of dynamic scenes[J]. ACM Transactions on Graphics, 2017, 36(4): 1-12.
[9] WU X , SHAO J , GAO L , et al. Unpaired image-to-image translation from shared deep space[C]//2018 25th IEEE International Conference on Image Processing (ICIP) . Athens: IEEE, 2018: 2127-2131.
[10] LIU M Y , BREUEL T , KAUTZ J . Unsupervised image-to-image translation networks[EB/OL]. arXiv: 1703.00848 (2018-07-23) [2022-03-25]. https://doi.org/10.48550/arXiv.1703.00848.
[11] MADAM N T, KUMAR S, RAJAGOPALAN A N. Unsupervised class-specific deblurring[C]//Computer Vision–ECCV 2018: 15th European Conference. Munich: Springer International Publishing, 2018: 358-374.
[12] HUANG X, LIU M Y, BELONGIE S, et al. Multimodal unsupervised image-to-image translation[EB/OL]. arXiv: 1804.04732 (2018-08-14) [2022-03-25]. https://doi.org/10.48550/arXiv.1804.04732.
[13] CHOI Y, CHOI M, KIM M, et al. Stargan: unified generative adversarial networks for multi-domain image-to-image translation[EB/OL]. arXiv: 1711.09020 (2018-09-21) [2022-03-25]: https://doi.org/10.48550/arXiv.1711.09020.
[14] HOFFMAN J, TZENG E, PARK T, et al. CyCADA: cycle-consistent adversarial domain adaptation[EB/OL].arXiv: 1711.03213 (2017-12-29) [2022-03-25]. https://doi.org/10.48550/arXiv.1711.03213.
[15] ZHU J Y, PARK T, ISOLA P, et al. Unpaired image-to-image translation using cycle-consistent adversarial networks[EB/OL].arXiv: 1703.10593 (2020-08-24) [2022-03-25]. https://doi.org/10.48550/arXiv.1703.10593.
[16] GOODFELLOW I, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial networks [J]. Communications of the ACM, 2020, 63(11): 139-144.
[17] GAWADE A, PANDHARKAR R, DEOLEKAR S. Gantoon: creative cartoons using generative adversarial network[C]//Information, Communication and Computing Technology: 5th International Conference, ICICCT 2020. Singapore: Springer, 2020: 222-230.
[18] HU D. An introductory survey on attention mechanisms in NLP problems[EB/OL]. arXiv: 1811.05544 (2018-11-12) [2022-03-25]. https://doi.org/10.48550/arXiv.1811.05544.
[19] WANG F, TAX D M J. Survey on the attention based RNN model and its applications in computer vision[EB/OL]. arXiv: 1601.06823 (2016-01-25) [2022-03-25]. https://doi.org/10.48550/arXiv.1601.06823.
[20] JIANG X, ZHANG L, XU M, et al. Attention scaling for crowd counting[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, WA: IEEE. 2020: 4706-4715.
[21] HAN Y S, YOO J, YE J C. Deep residual learning for compressed sensing CT reconstruction via persistent homology analysis[EB/OL]. arXiv: 1611.06391 (2016-11-25) [2022-03-25]. https://doi.org/10.48550/arXiv.1611.06391.
[22] SHIBATA N, TANITO M, MITSUHASHI K, et al. Development of a deep residual learning algorithm to screen for glaucoma from fundus photography [J]. Scientific Reports, 2018, 8(1): 14665.
[23] KUANG D Y. On reducing negative jacobian determinant of the deformation predicted by deep registration networks[EB/OL]. arXiv: 1907.00068 (2019-06-28) [2022-03-25]. https://doi.org/10.48550/arXiv.1907.00068.
[24] XIONG W, LIU D, SHEN X, et al. Unsupervised real-world low-light image enhancement with decoupled networks. [EB/OL]. arXiv: 2005.02818 (2020-05-06) [2022-03-25]. https://doi.org/10.48550/arXiv.2005.02818.
[25] RONNEBERGER O, FISCHER P, BROX T. U-net: convolutional networks for biomedical image segmentation[EB/OL]. arXiv: 1505.04597 (2015-05-18) [2022-03-25]. https://doi.org/10.48550/arXiv.1505.04597
[26] JIANG Y, GONG X, LIU D, et al. Enlightengan: deep light enhancement without paired supervision [J]. IEEE Transactions on Image Processing, 2021, 30: 2340-2349.
[27] IIZUKA S, SIMO-SERRA E, ISHIKAWA H. Globally and locally consistent image completion [J]. ACM Transactions on Graphics (TOG), 2017, 36(4): 107.
[28] MAO X, LI Q, XIE H, et al. Least squares generative adversarial networks[EB/OL]. arXiv: 1611.04076 (2017-04-05) [2022-03-25]. https://doi.org/10.48550/arXiv.1611.04076.
[29] HUYNH-THU Q, GHANBARI M. Scope of validity of PSNR in image/video quality assessment [J]. Electronics letters, 2008, 44(13): 800-801.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed