生成对抗式双重解耦的分阶段阴影去除算法

doi:10.3778/j.issn.1002-8331.2406-0307

摘要/Abstract

摘要： 针对当前阴影去除算法中普遍存在的阴影边缘处有伪影以及阴影与非阴影区域颜色恢复不一致问题，提出了一种生成对抗式双重解耦的分阶段阴影去除算法。该算法构建了一个生成对抗式主体模型框架，采用分阶段不同图像特征替代传统的单一彩色图像作为监督信号，实现了图像在通道和空间维度的解耦处理。在第一阶段，将图像在通道维度解耦为亮度通道（L）和颜色通道（AB）。针对亮度通道，提出了多尺度注意增强的空洞网络，通过逐级增加扩张率的空洞卷积捕获多尺度上下文信息，结合注意力机制精确提取局部特征，有效恢复阴影区域亮度并避免边缘伪影。针对颜色通道，设计了结合通道和空间注意力的颜色调整网络，专注于颜色校正。经过亮度恢复和颜色调整后的L通道和AB通道重新拼接，作为第二阶段的输入。在第二阶段，将图像在空间维度解耦为阴影和非阴影区域，设计了动态残差网络，采用标准卷积和深度可分离卷积分别提取特征，进一步解决边缘伪影和颜色不一致问题。实验结果表明，所提算法相较于其他先进算法表现优异，在避免边缘伪影、消除颜色不一致和提升图像质量方面具有显著优势。

关键词: 阴影去除, 生成对抗网络（GAN）, 通道解耦, 空洞卷积, 动态卷积

Abstract: Aiming at the problems of artifacts at shadow edges and inconsistent color recovery between shadow and non-shadow regions, which are commonly found in current shadow removal algorithms, this paper proposes a generative adversarial double decoupling staged shadow removal algorithm. The algorithm constructs a generative adversarial subject model framework, adopts different image features in stages instead of the traditional single color image as the supervisory signal, and realizes the decoupling processing of the image in the channel and spatial dimensions. Firstly, the image is decoupled into luminance channel (L) and color channel (AB) in the channel dimension. For the luminance channel, a multi-scale attention-enhanced dilated network is proposed, which captures multi-scale contextual information through dilated convolution with stepwise increasing dilation rate, and accurately extracts local features by combining with the attention mechanism to effectively restore the brightness of shadow regions and avoid edge artifacts. For the color channel, a color adjustment network combining channel and spatial attention is designed to focus on color correction. The L-channel and AB-channel after luminance recovery and color adjustment are re-spliced and used as inputs in the second stage. Secondly, the image is decoupled into shadow and non-shadow regions in the spatial dimension, and a dynamic residual network is designed to extract features using standard convolution and depth-separable convolution, respectively, to further solve the problems of edge artifacts and color inconsistency. The experimental results show that the algorithm in this paper performs well compared with other state-of-the-art algorithms, and has significant advantages in avoiding edge artifacts, eliminating color inconsistency and improving image quality.

Key words: shadow removal, generative adversarial network(GAN), channel decoupling, dilated convolution, dynamic convolution

曾文献, 张曼钰, 孙磊. 生成对抗式双重解耦的分阶段阴影去除算法[J]. 计算机工程与应用, 2025, 61(20): 260-269.

ZENG Wenxian, ZHANG Manyu, SUN Lei. Generative Adversarial Double Decoupling Staged Shadow Removal Algorithm[J]. Computer Engineering and Applications, 2025, 61(20): 260-269.

参考文献

[1] KHAN S H, BENNAMOUN M, SOHEL F, et al. Automatic shadow detection and removal from a single image[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(3): 431-446.
[2] FAN H, HAN M, LI J J. Image shadow removal using end-to-end deep convolutional neural networks[J]. Applied Sciences, 2019, 9(5): 1009.
[3] GOODFELLOW I, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial nets[C]//Proceedings of the 28th International Conference on Neural Information Processing Systems, 2014: 2672-2680.
[4] WANG J F, LI X, YANG J. Stacked conditional generative adversarial networks for jointly learning shadow detection and shadow removal[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 1788-1797.
[5] HU X W, JIANG Y T, FU C W, et al. Mask-ShadowGAN: learning to remove shadows from unpaired data[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2019: 2472-2481.
[6] JIN Y Y, SHARMA A, TAN R T. DC-ShadowNet: single-image hard and soft shadow removal using unsupervised domain-classifier guided network[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2021: 5007-5016.
[7] AKENINE-M?LLER T, HAINES E, HOFFMAN N. Real-time rendering[M]. Boca Raton: CRC Press, 2018.
[8] LIU Z, YIN H, MI Y, et al. Shadow removal by a lightness-guided network with training on unpaired data[J]. IEEE Transactions on Image Processing, 2021, 30: 1853-1865.
[9] YANG H, NAN G, LIN M, et al. LAB-Net: LAB color-space oriented lightweight network for shadow removal[J]. arXiv: 2208.13039, 2022.
[10] WANG Q L, WU B G, ZHU P F, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 11531-11539.
[11] WANG C, SHAO M W, ZHANG W T, et al. A context-based multi-scale discriminant model for natural image inpainting[C]//Proceedings of the European Conference on Computer Vision, 2021: 1-14.
[12] CUN X D, PUN C M, SHI C. Towards ghost-free shadow removal via dual hierarchical aggregation network and shadow matting GAN[C]//Proceedings of the AAAI Conference on Artificial Intelligence, 2020: 10680-10687.
[13] YU F, KOLTUN V. Multi-scale context aggregation by dilated convolutions[J]. arXiv:1511.07122, 2025.
[14] HE K M, ZHANG X Y, REN S Q, et al. Deep residual lear-ning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 770-778.
[15] VASWANI A. Attention is all you need[C]//Advances in Neural Information Processing Systems, 2017: 6000-6010.
[16] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 7132-7141.
[17] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]///Proceedings of the European Conference on Computer Vision, 2018: 3-19.
[18] VAN MOLLE P, VERBELEN T, VANKEIRSBILCK B, et al. Leveraging the Bhattacharyya coefficient for uncertainty quantification in deep neural networks[J]. Neural Computing and Applications, 2021, 33(16): 10259-10275.
[19] HE K M, ZHANG X Y, REN S Q, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[C]//Proceedings of the European Conference on Computer Vision, 2014: 346-361.
[20] XU Y M, LIN M B, YANG H, et al. Shadow-aware dynamic convolution for shadow removal[J]. Pattern Recognition, 2024, 146: 109969.
[21] LE H, SAMARAS D. Shadow removal via shadow image decomposition[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2019: 8577-8586.
[22] QU L Q, TIAN J D, HE S F, et al. DeshadowNet: a multi-context embedding deep network for shadow removal[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 2308-2316.
[23] KINGMA D P. Adam: a method for stochastic optimization[J]. arXiv:1412.6980, 2014.
[24] PEDREGOSA F, VAROQUAUX G, GRAMFORT A, et al. Scikit-learn: machine learning in python journal of machine learning research[J]. Journal of Machine Learning Research, 2011, 12: 2825-2830.
[25] FU L, ZHOU C Q, GUO Q, et al. Auto-exposure fusion for single-image shadow removal[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021: 10566-10575.
[26] LIU J, WANG Q, FAN H, et al. A shadow imaging bilinear model and three-branch residual network for shadow removal[J]. IEEE Transactions on Neural Networks and Learning Systems 2024, 35(11): 15857-15871.