基于暗区域引导的低照度图像增强

doi:10.3969/j.issn.0255-8297.2025.02.005

摘要/Abstract

摘要： 针对现有增强方法在图像照度分布不均匀时出现的局部过度增强、颜色失真以及细节丢失问题，提出了一种结合暗区域引导与注意力机制的低照度图像增强方法。首先，采用简单线性迭代聚类方法生成暗区域引导图，指导网络在保障正常曝光区域不过度增强的情况下，重点增强图像曝光不足区域；其次，设计通道注意力模块，提高网络对颜色信息的提取能力，更好地恢复图像颜色，保证颜色自然度；再次，设计全局上下文模块，增加网络全局感知能力，丰富图像细节信息；最后，增强网络融合输入特征和暗区域注意力网络输出特征，实现图像对比度再增强。在6个公共数据集上进行多组对比实验，分别从主观与客观两方面进行性能对比，结果表明所提方法能够有效解决低照度图像存在的颜色失真、细节丢失和曝光不均匀问题，具有较好的视觉增强效果与泛化性。

关键词: 低照度图像增强, 暗区域引导, 通道注意力模块, 全局上下文模块, 深度学习

Abstract: To address the issues of overexposure, color distortion and detail loss in existing enhancement methods when image illumination distribution is uneven, a low-light image enhancement method combining dark region guidance and attention mechanism is proposed. Firstly, the simple linear iterative clustering (SLIC) method is used to generate a dark region guidance map, which guides the network to enhance the underexposed regions of the image while ensuring that the normally exposed regions are not overexposed. Secondly, a channel attention module is designed to improve the extraction of color information, effectively restoring the image color while maintaining natural color fidelity. Subsequently, a global context module is established to enhance the network’s global perception capability, enriching image details. Finally, an enhancement network is designed to fuse the input features with the output features of the dark area attention network,achieving contrast re-enhancement. Multiple comparative experiments are conducted on six public datasets to compare the performance from both subjective and objective aspects. It is shown that the proposed method effectively solves the problems of color distortion, detail loss and uneven exposure in low-light images, delivering superior visual enhancement effect and generalizability.

Key words: low-light image enhancement, dark region guidance, channel attention module, global context module, deep learning

中图分类号:

TP751.1

汪婉灵, 熊邦书, 欧巧凤, 余磊, 饶智博. 基于暗区域引导的低照度图像增强[J]. 应用科学学报, 2025, 43(2): 245-256.

WANG Wanling, XIONG Bangshu, OU Qiaofeng, YU Lei, RAO Zhibo. Low-Light Image Enhancement Based on Dark Region Guidance[J]. Journal of Applied Sciences, 2025, 43(2): 245-256.

参考文献

[1] Li G, Yang Y, Qu X, et al. A deep learning based image enhancement approach for autonomous driving at night [J]. Knowledge-Based Systems, 2021, 213: 106617.
[2] Tan X, Xu K, Cao Y, et al. Night-time scene parsing with a large real dataset [J]. IEEE Transactions on Image Processing, 2021, 30: 9085-9098.
[3] Jebadass J R, Balasubramaniam P. Low light enhancement algorithm for color images using intuitionistic fuzzy sets with histogram equalization [J]. Multimedia Tools and Applications, 2022, 81(6): 8093-8106.
[4] Shang X, An N, Zhang S, et al. Toward robust and efficient low-light image enhancement: progressive attentive retinex architecture search [J]. Tsinghua Science and Technology, 2022, 28(3): 580-594.
[5] Huang Z, Wang Z, Zhang J, et al. Image enhancement with the preservation of brightness and structures by employing contrast limited dynamic quadri-histogram equalization [J]. Optik, 2021, 226: 165877.
[6] Li X, Gao M, Shang J, et al. A complexity reduction based retinex model for low luminance retinal fundus image enhancement [J]. Network Modeling Analysis in Health Informatics and Bioinformatics, 2022, 11(1): 30.
[7] Ren X, Yang W, Cheng W H, et al. LR3M: robust low-light enhancement via low-rank regularized retinex model [J]. IEEE Transactions on Image Processing, 2020, 29: 5862-5876.
[8] Zhang Y H, Zhang J W, Guo X J. Kindling the darkness: a practical low-light image enhancer [C]//27th ACM International Conference on Multimedia, 2019: 1632-1640.
[9] Ma L, Liu R, Zhang J, et al. Learning deep context-sensitive decomposition for low-light image enhancement [J]. IEEE Transactions on Neural Networks and Learning Systems, 2021, 33(10): 5666-5680.
[10] Liu R, Ma L, Zhang J, et al. Retinex-inspired unrolling with cooperative prior architecture search for low-light image enhancement [C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021: 10561-10570.
[11] Zhang F, Shao Y, Sun Y, et al. Unsupervised low-light image enhancement via histogram equalization prior [DB/OL]. 2021[2024-01-28]. https://arxiv.org/abs/2112.01766.
[12] Fan M, Wang W, Yang W, et al. Integrating semantic segmentation and retinex model for low-light image enhancement [C]//28th ACM International Conference on Multimedia, 2020: 2317-2325.
[13] Wang H, Xu K, Lau R W H. Local color distributions prior for image enhancement [C]//European Conference on Computer Vision, 2022: 343-359.
[14] 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.
[15] Chen L, Chu X, Zhang X, et al. Simple baselines for image restoration [C]//European Conference on Computer Vision, 2022: 17-33.
[16] Achanta R, Susstrunk S. Superpixels and polygons using simple non-iterative clustering [C]//IEEE Conference on Computer Vision and Pattern Recognition, 2017: 4651-4660.
[17] Wei C, Wang W J, Yang W H, et al. Deep retinex decomposition for low-light enhancement [C]//British Machine Vision Conference (BMVC 2018), 2018: 1-10.
[18] Lee C, Kim C S. Contrast enhancement based on layered difference representation of 2D histograms [J]. IEEE Transactions on Image Processing, 2013, 22(12): 5372-5384.
[19] Fu X Y, Zeng D L, Huang Y, et al. A fusion-based enhancing method for weakly illuminated images [J]. Signal Processing, 2016, 129: 82-96.
[20] Guo X J, Li Y, Ling H B. LIME: low-light image enhancement via illumination map estimation [J]. IEEE Transactions on Image Processing, 2017, 26(2): 982-993.
[21] Wang S, Zheng J, Hu H M, et al. Naturalness preserved enhancement algorithm for nonuniform illumination images [J]. IEEE Transactions on Image Processing, 2013, 22(9): 3538- 3548.
[22] Vonikakis V, Andreadis I, Gasteratos A. Fast centre-surround contrast modification [J]. IET Image Processing, 2008, 2(1): 19-34.
[23] Mittal A, Soundararajan R, Bovik A C. Making a “completely blind” image quality analyzer [J]. IEEE Signal Processing Letters, 2013, 20(3): 209-212.
[24] Gu K, Lin W, Zhai G, et al. No-reference quality metric of contrast-distorted images based on information maximization [J]. IEEE Transactions on Cybernetics, 2016, 47(12): 4559-4565.
[25] Mittal A, Moorthy A K, Bovik A C. No-reference image quality assessment in the spatial domain [J]. IEEE Transactions on Image Processing, 2012, 21(12): 4695-4708.
[26] Qu J, Liu R W, Gao Y, et al. Double domain guided real-time low-light image enhancement for ultra-high-definition transportation surveillance [DB/OL]. 2023[2024-01-28]. https://arxiv.org/abs/2309.08382.
[27] Fan G D, Fan B, Gan M, et al. Multiscale low-light image enhancement network with illumination constraint [J]. IEEE Transactions on Circuits and Systems for Video Technology, 2022, 32(11): 7403-7417.
[28] Lu Y, Guo Y, Liu R W, et al. MTRBNet: multi-branch topology residual block-based network for low-light enhancement [J]. IEEE Signal Processing Letters, 2022, 29: 1127-1131.
[29] Wu W, Weng J, Zhang P, et al. Uretinex-net: retinex-based deep unfolding network for low-light image enhancement [C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022: 5901-5910.
[30] Fu Z, Yang Y, Tu X, et al. Learning a simple low-light image enhancer from paired lowlight instances [C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023: 22252-22261.