自适应Transformer网络下的单幅图像去雾方法

doi:10.3778/j.issn.1002-8331.2209-0112

摘要/Abstract

摘要： 针对现有图像去雾方法对不同雾浓度的雾处理不完善、颜色失真、细节恢复不佳等问题，提出一种基于自适应Transformer的单图像去雾方法。该方法包括编码器、自适应Transformer、解码器三个模块。编码器将输入的图像进行局部特征提取，得到细节特征。为更好地对不同雾浓度进行提取，提出自适应Transformer，自适应Transformer可以在提取全局特征的基础上，自适应提取不同尺度颗粒雾度特征。解码器用于图像进行恢复与重建。该方法在SOTS室外测试集进行实验，其PSNR为27.45?dB，SSIM为0.954?6，与对比方法中最高的相比分别提高0.43?dB和0.005?3，从而验证出该方法取得了较好的客观结果。此外主观方面去雾图像有效改善了颜色失真，以及不同雾浓度处理不理想的问题。

关键词: 图像去雾, Transformer, 特征提取, 注意力机制

Abstract: Aiming at the problems of imperfect haze processing, color distortion, and poor detail recovery of existing image dehazing methods with different haze concentrations, a single image dehazing method based on adaptive Transformer is proposed. The method includes three modules：encoder, adaptive Transformer, and decoder. The encoder performs local feature extraction on the input image to obtain detailed features. In order to better extract different haze concentrations, an adaptive Transformer is proposed. The adaptive Transformer can adaptively extract the haze features of different scale particles on the basis of extracting global features. The decoder is used for image restoration and reconstruction. The method is tested on the SOTS outdoor test set, and its PSNR is 27.45 dB and SSIM is 0.954 6, which are respectively 0.43?dB and 0.005?3 higher than that in the comparison method, which verifies that the method achieves good objective results. In addition, the subjective aspect of the dehazing image can effectively improve the color distortion and the problem of unsatisfactory handling of different fog concentrations.

Key words: image dehazing, Transformer, feature extraction, attention mechanism

金海波, 马琳琳, 田桂源. 自适应Transformer网络下的单幅图像去雾方法[J]. 计算机工程与应用, 2024, 60(3): 237-245.

JIN Haibo, MA Linlin, TIAN Guiyuan. Single Image Defogging Method Under Adaptive Transformer Network[J]. Computer Engineering and Applications, 2024, 60(3): 237-245.

参考文献

[1] SHARMA N, KUMAR V, SINGLA S K. Single image defogging using deep learning techniques: past, present and future[J]. Archives of Computational Methods in Engineering, 2021, 28(7): 4449-4469.
[2] ENGIN D, GEN A, EKENEL H K. Cycle-dehaze: enhanced cycleGAN for single image dehazing[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, June 18-22, 2018. Washington: IEEE Computer Society, 2018: 937-938.
[3] 张晓东, 秦娟娟, 贾仲仲. 多尺度Retinex在图像去雾算法中的应用研究[J]. 西昌学院学报(自然科学版), 2021, 35(3): 60-65.
ZHANG X D, QIN J J, JIA Z Z. Study on the application of multi-scale Retinex to image defogging algorithm[J]. Journal of Xichang Unversity (Natural Science Edition), 2021, 35(3): 60-65.
[4] KIM T K, PAIK J K, KANG B S. Contrast enhancement system using spatially adaptive histogram equalization with temporal filtering[J]. IEEE Transactions on Consumer Electronics, 1998, 44(1): 82-87.
[5] HSE W Y, CHEN Y S. Single image dehazing using wavelet-based haze-lines and denoising[J]. IEEE Access, 2021, 9: 104547-104559.
[6] NARASIMHAN S G, NAYAR S K. Vision and the atmosphere[J]. International Journal of Computer Vision, 2022, 48(3): 233-254.
[7] HE K M, SUN J, TANG X O. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2341-2353.
[8] MENG G F, WANG Y, DUAN J Y, et al. Efficient image dehazing with boundary constraint and contextual regularization[C]//Proceedings of the IEEE International Conference on Computer Vision. Australia: IEEE, 2013: 617-624.
[9] 黄文君, 李杰, 齐春. 低秩与字典表达分解的浓雾霾场景图像去雾算法[J]. 西安交通大学学报, 2020, 54(4): 118-125.
HUANG W J, LI J, QI C. A defogging algorithm for dense fog images via low rank and dictionary expression decomposition[J]. Journal of Xi’an Jiaotong University, 2020, 54(4): 118-125.
[10] LI B, PENG X, WANG Z, et al. AOD-Net: all-in-one dehazing network[C]//2017 IEEE International Conference on Computer Vision. Venice, Italy: IEEE, 2017.
[11] ZHENG Z, REN W, CAO X, et al. Ultra-high-definition image dehazing via multi-guided bilateral learning[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway, NJ: IEEE, 2021: 16180-16189.
[12] 刘广洲, 李金宝, 任东东, 等. 密集连接扩张卷积神经网络的单幅图像去雾[J]. 计算机科学与探索, 2021, 15(1): 185-194.
LIU G Z, LI J B, REN D D, et al. Single image dehazing method based on densely connected dilated convolutional neural network[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(1): 185-194.
[13] WANG C, LI Z, WU J, et al. Deep residual haze network for image dehazing and deraining[J]. IEEE Access, 2020, 8: 9488-9500.
[14] 王德兴, 高凯, 袁红春, 等. 基于色彩校正和TransFormer细节锐化的水下图像增强[J/OL]. 吉林大学学报(工学版): 1-13[2022-08-22]. http://kns.cnki.net/kcms/detail/22.1341. T.20220815.1339.003.html.
WANG D X, GAO K, YUAN H C, et al. Underwater image enhancement based on color correction and TransFormer detail sharpening[J/OL]. Journal of Jilin University (Engineering and Technology Edition): 1-13[2022-08-22]. http://kns.cnki.net/kcms/detail/22.1341.T.20220815.1339.003.html.
[15] 梁礼明, 周珑颂, 尹江, 等. 融合多尺度Transformer的皮肤病变分割算法[J/OL]. 吉林大学学报(工学版): 1-13. DOI:10.13229/j.cnki.jdxbgxb20220692.
LIANG L M, ZHOU L S, YIN J, et al. Fusion multi-scale transformer skin lesion segmentation algorithm[J/OL]. Journal of Jilin University (Engineering and Technology Edition): 1-13. DOI:10.13229/j.cnki.jdxbgxb20220692.
[16] 高辉, 邓淼磊, 赵文君, 等. 基于弱监督的改进Transformer在人群定位中的应用[J]. 计算机工程与应用, 2023, 59(19): 92-98.
GAO H, DENG M L, ZHAO W J, et al. Application of improved Transformer based on weakly supervised in crowd localization[J]. Computer Engineering and Applications, 2023, 59(19): 92-98.
[17] 王玉萍, 曾毅, 李胜辉, 等. 一种基于Transformer的三维人体姿态估计方法[J]. 图学学报, 2023, 44(1): 139-145.
WANG Y P, ZENG Y, LI S H, et al. A Transformer-based 3D human pose estimation method[J]. Journal of Graphics, 2023, 44(1): 139-145.
[18] 田战胜, 刘立波. 基于改进Transformer的细粒度图像分类模型[J]. 激光与光电子学进展, 2023, 60(2): 161-168.
TIAN Z S, LIU L B. Fine-grained image classification model based on improved Transformer[J]. Laser & Optoelectronics Progress, 2023, 60(2): 161-168.
[19] LI X, HUA Z, LI J. Two stage single image dehazing network using swin transformerr[J]. IET Image Processing, 2022, 16(9): 2518-2534.
[20] GAO G, CAO J, BAO C, et al. A novel Transformer-based attention network for image dehazing[J]. Sensors, 2022, 22(9): 3428.
[21] LI S, YUAN Q, ZHANG Y, et al. Image dehazing algorithm based on deep learning coupled local and global features[J]. Applied Sciences, 2022, 12(17): 8552.
[22] WANG Z D, CUN X D, BAO J M, et al. Uformer: a general u-shaped Transformer for image restoration[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). USA: IEEE, 2022: 17683-17693.
[23] JIAO Q, LIU M, NING B, et al. Image dehazing based on local and non-local features[J]. Fractal and Fractional, 2022, 6(5): 262.
[24] RONNEBERGER O, FISCHER P, BROX T. U-net: convolutional networks for biomedical image segmentation[C]//International Conference on Medical image Computing and Computer-Assisted Intervention. Cham: Springer, 2015: 234-241.
[25] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C//Advances in Neural Information Processing Systems, 2017: 6000-6010.
[26] LI B Y, REN W Q, FU D P, et al. Benchmarking single-image dehazing and beyond[J]. IEEE Transactions on Image Processing, 2019, 28(1): 492-505.
[27] WANG Z, BOVIK A C, SHEIKH H R, et al. Image quality assessment: from error visibility to structural similarity[J]. IEEE Transactions on Image Processing, 2004, 13(4): 600-612.