AWTV和高斯注意力引导的LDCT图像去噪网络

doi:10.3778/j.issn.1002-8331.2402-0037

摘要/Abstract

摘要： 低剂量CT（low-dose CT，LDCT）图像的去噪任务是一个高度复杂且不确定的逆问题。现有的基于CNN的方法虽然有效，但提升空间有限且计算成本高。相比之下，将图像先验知识与模型相结合来辅助图像去噪是一种更有效的方法。提出了一种名为AWTV_GANet的LDCT图像去噪框架。该框架利用自适应加权总变分（adaptive weighted total variation，AWTV）展开和高斯注意力引导的方法，通过端到端的CNN模型，将噪声优化模型、边缘检测模型和图像重建模型集成在一起。实验证明，AWTV_GANet能够准确地去除伪影噪声，并恢复出更精细的结构细节，与其他方法相比具有优异的性能。

关键词: 低剂量CT, 图像去噪, 深度学习, 自适应加权总变分

Abstract: The task of denoising low-dose CT (LDCT) images is a highly complex and uncertain inverse problem. Although the existing CNN-based methods are effective, the improvement space is limited and the calculation cost is high. In contrast, it is a more effective method to combine image prior knowledge with model to assist image denoising. In this paper, an LDCT image denoising framework named AWTV_GANet is proposed. The framework uses adaptive weighted total variation (AWTV) expansion and Gaussian attention guidance to integrate the noise optimization model, edge detection model and image reconstruction model through the end-to-end CNN model. Experiments show that AWTV_GANet can accurately remove artifact noise and recover finer structural details, which has excellent performance compared with other methods.

Key words: low-dose CT, image denoising, deep learning, adaptive weighted total variation

李志媛, 刘祎, 张鹏程, 张丽媛, 任时磊, 芦婧, 桂志国. AWTV和高斯注意力引导的LDCT图像去噪网络[J]. 计算机工程与应用, 2025, 61(3): 253-263.

LI Zhiyuan, LIU Yi, ZHANG Pengcheng, ZHANG Liyuan, REN Shilei, LU Jing, GUI Zhiguo. Adaptive Weighted Total Variational and Gaussian Attention-Guided LDCT Image Denoising Networks[J]. Computer Engineering and Applications, 2025, 61(3): 253-263.

参考文献

[1] BOAS F E, FLEISCHMANN D. CT artifacts: causes and reduction techniques[J]. Imaging in Medicine, 2012, 4(2): 229-240.
[2] WANG J, LI T, LU H, et al. Penalized weighted least-squares approach to sinogram noise reduction and image reconstruction for low-dose X-ray computed tomography[J]. IEEE Transactions on Medical Imaging, 2006, 25: 1272-1283.
[3] MANDUCA A, YU L, TRZASKO J D, et al. Projection space denoising with bilateral filtering and CT noise modeling for dose reduction in CT[J]. Medical Physics, 2009, 36(11): 4911-4919.
[4] BALDA M, HORNEGGER J, HEISMANN B. Ray contribution masks for structure adaptive sinogram filtering[J]. IEEE Transactions on Medical Imaging, 2012, 31(6): 1228-1239.
[5] LIU L, LI X, XIANG K, et al. Low-dose CBCT reconstruction using hessian schatten penalties[J]. IEEE Transactions on Medical Imaging, 2017, 36(12): 2588-2599.
[6] LIU J, HU Y, YANG J, et al. 3D feature constrained reconstruction for low dose CT imaging[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2016, 28(99): 1232-1247.
[7] 刘祎，高净植，桂志国. 基于NSST域卷积神经网络的低剂量CT图像恢复[J]. 计算机工程与应用, 2019, 55(23): 209-215.
LIU Y, GAO J Z, GUI Z G. Low-dose CT restoration based on CNN in NSST domain[J]. Computer Engineering and Applications, 2019, 55(23): 209-215.
[8] DABOV K, FOI A, KATKOVNIK V, et al. Image denoising by sparse 3?D transform-domain collaborative filtering[J]. IEEE Transactions on Image Processing, 2007, 16(8): 2080-2095.
[9] YIN Y, SUN W, LU J, et al. A non-local means approach for PET image denoising[J]. ?Journal of Biomedical Engineering, 2010, 27(2): 274-281 .
[10] CHEN H, ZHANG Y, KALRA M K, et al. Low-dose CT with a residual encoder-decoder convolutional neural network (RED-CNN)[J]. IEEE Transactions on Medical Imaging, 2017, 36(99): 2524-2535.
[11] FAN F, SHAN H, KALRA M, et al. Quadratic autoencoder (Q-AE) for low-dose CT denoising[J]. IEEE Transactions on Medical Imaging, 2020, 39(6): 2035-2050.
[12] 刘帅, 安冬, 须颖, 等. 基于RN-CNN模型的低剂量CT图像去噪方法[J]. 计算机应用与软件, 2023, 40(1): 241-247.
LIU S, AN D, XU Y, et al. Low dose CT image denoising method based on RN-CNN model[J]. Computer Applications and Software, 2023, 40(1): 241-247.
[13] LIANG T, JIN Y, LI Y, et al. EDCNN: edge enhancement-based densely connected network with compound loss for Low-Dose CT denoising[C]//Proceedings of the 2020 15th IEEE International Conference on Signal Processing (ICSP), Beijing, China, 2020: 193-198.
[14] JIANG X, WANG L, HE Z, et al. Learning a frequency separation network with hybrid convolution and adaptive aggregation for low-dose CT denoising[C]//Proceedings of the 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2021: 919-925.
[15] LI Z, LIU Y, SHU H, et al. Multi-scale feature fusion network for low-dose CT denoising[J]. Journal of Digital Imaging, 2023, 36(4): 1808-1825.
[16] 郭志涛, 周峰, 赵琳琳, 等. 边缘保护与多阶段网络相结合的LDCT图像去噪[J]. 计算机工程与应用, 2023, 59(1): 252-258.
GUO Z T, ZHOU F, ZHAO L L, et al. LDCT image denoising based on edge protection and multi-stage network[J]. Computer Engineering and Applications, 2023, 59(1): 252-258.
[17] 欧阳婉卿, 张剑, 彭辉, 等. 基于改进生成对抗网络的低剂量CT去噪算法[J]. 光电子·激光, 2022, 33(2): 171-180.
OUYANG W Q, ZHANG J, PENG H, et al. Based on improved generated against network of low dose CT denoising algorithm[J]. Journal of Optoelectronics·Laser, 2022, 33(2): 171-180.
[18] YANG Q, YAN P, ZHANG Y, et al. Low-dose CT image denoising using a generative adversarial network with wasserstein distance and perceptual loss[J]. IEEE Transactions on Medical Imaging, 2018, 37(6): 1348-1357.
[19] GENG M, MENG X, YU J, et al. Content-noise complementary learning for medical image denoising[J]. IEEE Transactions on Medical Imaging, 2022, 41(2): 407-419.
[20] 冯婉妍, 蒙祖强, 陈冲. 融合Transformer与UNet的低剂量CT图像去噪方法[J]. 广西大学学报 (自然科学版), 2023, 48(4): 972-984.
FENG W Y, MENG Z Q, CHEN C. Low dose CT image denoising method with Transformer and UNet[J]. Journal of Guangxi University (Natural Science Edition), 2023, 13(4): 972-984.
[21] WANG D, FAN F, WU Z, et al. CTformer: convolution-free Token2Token dilated vision transformer for low-dose CT denoising[J]. Physics in Medicine & Biology, 2023, 68(6): 065012.
[22] 颜溶標, 刘文婷, 谷亚男, 等. 基于多尺度边缘提取和加权卷积稀疏编码的低剂量CT去噪算法[J]. 国外电子测量技术, 2022, 41(9): 9-15.
YAN R B, LIU W T, GU Y N, et al. Low dose CT denoising algorithm based on multi-scale edge extraction and weighted convolutional sparse coding[J]. Foreign Electronic Measurement Technology, 2022, 41(9): 9-15.
[23] WANG H, LI Y, MENG D, et al. Adaptive convolutional dictionary network for CT metal artifact reduction[C]//Proceedings of the 2022 31st International Joint Conference on Artificial Intelligence, 2022: 1401-1407.
[24] YAN R, LIU Y, LIU Y H, et al. Image denoising for low-dose CT via convolutional dictionary learning and neural network[J]. IEEE Transactions on Computational Imaging, 2023, 9: 83-93.
[25] LI Z, LIU Y, CHEN Y, et al. Dual-domain fusion deep convolutional neural network for low-dose CT denoising[J]. Journal of X-Ray Science and Technology, 2023, 31(4): 757-775.
[26] LIU Y, MA J, FAN Y, et al. Adaptive-weighted total variation minimization for sparse data toward low-dose X-ray computed tomography image reconstruction[J]. Physics in Medicine & Biology, 2012, 57(23): 7923.
[27] LIU Y H, LIU Y, ZHANG P, et al. Cone-beam computed tomography based on truncated adaptive-weight total variation[J]. NDT & E International, 2023, 133: 102755.
[28] QI H, TAN S, LI Z. Anisotropic weighted total variation feature fusion network for remote sensing image denoising[J]. Remote Sensing, 2022, 14(24): 6300.
[29] GONG Y, SBALZARINI I F. Curvature filters efficiently reduce certain variational energies[J]. IEEE Transactions on Image Processing, 2017, 26(4): 1786-1798.
[30] SU Q, BI B, ZHANG P, et al. GPR image clutter suppression using gaussian curvature decomposition in the PCA domain[J]. Remote Sensing, 2022, 14(19): 4879.
[31] TANG W, LIN Z, GONG Y. GC-Net: an unsupervised network for gaussian curvature optimization on images[J]. Journal of Signal Processing Systems, 2023, 95(1): 77-88.
[32] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018: 7132-7141.
[33] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]//Proceedings of the European Conference on Computer Vision (ECCV), 2018: 3-19.
[34] FANG F, LI J, YUAN Y, et al. Multilevel edge features guided network for image denoising[J]. IEEE Transactions on Neural Networks and Learning Systems, 2021, 32(9): 3956-3970.
[35] 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.
[36] 王宇庆, 刘维亚, 王勇. 一种基于局部方差和结构相似度的图像质量评价方法[J]. 光电子·激光, 2008, 19(11): 1546-1553.
WANG Y Q, LIU W Y, WANG Y. Image quality assessment based on local variance and structure similarity[J]. Journal of Optoelectronics·Laser, 2008, 19(11): 1546-1553.
[37] XUE W, ZHANG L, MOU X, et al. Gradient magnitude similarity deviation: a highly efficient perceptual image quality index[J]. IEEE Transactions on Image Processing, 2013, 23(2): 684-695.
[38] ZHANG Lin ZHANG Lei, MOU X, et al. FSIM: a feature similarity index for image quality assessment[J]. IEEE Transactions on Image Processing, 2011, 20(8): 2378-2386.
[39] SHEIKH H R, BOVIK A C. Image information and visual quality[J]. IEEE Transactions on Image Processing, 2006, 15(2): 430-444.