《计算机应用》唯一官方网站 ›› 2024, Vol. 44 ›› Issue (8): 2571-2579.DOI: 10.11772/j.issn.1001-9081.2023081131
收稿日期:
2023-08-23
修回日期:
2023-11-01
接受日期:
2023-11-14
发布日期:
2024-08-22
出版日期:
2024-08-10
通讯作者:
石洪波
作者简介:
丁宇伟(1998—),男,山西忻州人,硕士研究生,CCF会员,主要研究方向:机器学习、图像处理基金资助:
Yuwei DING1, Hongbo SHI1(), Jie LI1,2, Min LIANG1
Received:
2023-08-23
Revised:
2023-11-01
Accepted:
2023-11-14
Online:
2024-08-22
Published:
2024-08-10
Contact:
Hongbo SHI
About author:
DING Yuwei, born in 1998, M. S. candidate. His research interests include machine learning, image processing.Supported by:
摘要:
针对当前基于Transformer的图像去噪算法侧重于捕获图像的全局特征,而忽视局部特征对于恢复图像细节关键作用的问题,提出一种基于局部和全局特征解耦的图像去噪网络。该网络包含2个基于混合Transformer模块(HTB)的多尺度分支和1个基于卷积神经网络(CNN)的单尺度分支,旨在将HTB强大的全局建模能力与CNN的局部建模优势有机结合,生成上下文信息丰富且空间细节准确的输出。HTB采用自注意力机制自适应地对空间和通道维度的依赖关系建模,以激活范围更广的输入像素进行重建。鉴于不同分支间可能存在的信息冲突,设计特征传递模块,通过跨分支传递全局特征并抑制低频信息,从而确保各分支间的协同作用。实验结果表明,在真实世界图像数据集SIDD上,与基于Transformer的去噪网络Uformer相比,所提网络的峰值信噪比(PSNR)提高了0.09 dB,结构相似度(SSIM)提高了0.001;在合成图像数据集Urban100上,与多阶段去噪网络MSPNet(Multi-Stage Progressive denoising Network)相比,所提网络的平均PSNR提高了0.41 dB。可见,所提网络能有效去除图像噪声,并重建出更精细的纹理细节。
中图分类号:
丁宇伟, 石洪波, 李杰, 梁敏. 基于局部和全局特征解耦的图像去噪网络[J]. 计算机应用, 2024, 44(8): 2571-2579.
Yuwei DING, Hongbo SHI, Jie LI, Min LIANG. Image denoising network based on local and global feature decoupling[J]. Journal of Computer Applications, 2024, 44(8): 2571-2579.
方法 | DND | SIDD | ||
---|---|---|---|---|
PSNR/dB | SSIM | PSNR/dB | SSIM | |
CBM3D[ | 34.51 | 0.851 | 25.65 | 0.685 |
DnCNN[ | 32.43 | 0.790 | 23.66 | 0.583 |
CBDNet[ | 38.06 | 0.942 | 30.78 | 0.801 |
RIDNet[ | 39.26 | 0.953 | 38.71 | 0.951 |
MIRNet[ | 39.88 | 0.956 | 39.72 | 0.959 |
MPRNet[ | 39.80 | 0.954 | 39.71 | 0.958 |
MSANet[ | 39.65 | 0.955 | 39.56 | 0.912 |
MSPNet[ | 39.75 | 0.954 | 39.78 | 0.959 |
MIRNet-v2[ | 39.86 | 0.955 | 39.84 | 0.959 |
Uformer[ | 39.96 | 0.956 | 39.77 | 0.959 |
ADFNet[ | 39.87 | 0.955 | 39.79 | 0.960 |
LGDNet | 40.10 | 0.956 | 39.86 | 0.960 |
表1 不同方法在真实世界图像数据集上的去噪结果
Tab. 1 Denoising results of different methods on real-world image datasets
方法 | DND | SIDD | ||
---|---|---|---|---|
PSNR/dB | SSIM | PSNR/dB | SSIM | |
CBM3D[ | 34.51 | 0.851 | 25.65 | 0.685 |
DnCNN[ | 32.43 | 0.790 | 23.66 | 0.583 |
CBDNet[ | 38.06 | 0.942 | 30.78 | 0.801 |
RIDNet[ | 39.26 | 0.953 | 38.71 | 0.951 |
MIRNet[ | 39.88 | 0.956 | 39.72 | 0.959 |
MPRNet[ | 39.80 | 0.954 | 39.71 | 0.958 |
MSANet[ | 39.65 | 0.955 | 39.56 | 0.912 |
MSPNet[ | 39.75 | 0.954 | 39.78 | 0.959 |
MIRNet-v2[ | 39.86 | 0.955 | 39.84 | 0.959 |
Uformer[ | 39.96 | 0.956 | 39.77 | 0.959 |
ADFNet[ | 39.87 | 0.955 | 39.79 | 0.960 |
LGDNet | 40.10 | 0.956 | 39.86 | 0.960 |
方法 | CBSD68 | Urban100 | Kodak24 | ||||||
---|---|---|---|---|---|---|---|---|---|
CBM3D[ | 29.73 | 27.38 | 26.00 | 30.36 | 27.95 | 26.31 | 30.89 | 28.63 | 27.27 |
DnCNN[ | 30.40 | 28.01 | 26.56 | 30.28 | 28.16 | 26.17 | 31.39 | 29.16 | 27.64 |
RDN[ | 30.67 | 28.31 | 26.85 | 31.69 | 29.29 | 27.63 | 31.94 | 29.66 | 28.20 |
MSPNet[ | 30.76 | 28.47 | 27.03 | 31.64 | 29.40 | 27.66 | 31.99 | 29.74 | 28.34 |
MDRN[ | 30.61 | 28.27 | 26.82 | 31.41 | 29.00 | 27.37 | 31.73 | 29.44 | 28.01 |
MSANet[ | 30.67 | 28.36 | 29.96 | N/A | N/A | N/A | 31.78 | 29.57 | 28.17 |
PANet[ | 30.70 | 28.33 | 26.89 | 31.87 | 29.47 | 27.87 | 31.96 | 29.65 | 28.20 |
LGDNet | 30.79 | 28.50 | 27.04 | 31.98 | 29.78 | 28.05 | 32.02 | 29.76 | 28.37 |
表2 不同方法在合成彩色图像数据集上的去噪结果
Tab. 2 Denoising results of different methods on synthetic color image datasets
方法 | CBSD68 | Urban100 | Kodak24 | ||||||
---|---|---|---|---|---|---|---|---|---|
CBM3D[ | 29.73 | 27.38 | 26.00 | 30.36 | 27.95 | 26.31 | 30.89 | 28.63 | 27.27 |
DnCNN[ | 30.40 | 28.01 | 26.56 | 30.28 | 28.16 | 26.17 | 31.39 | 29.16 | 27.64 |
RDN[ | 30.67 | 28.31 | 26.85 | 31.69 | 29.29 | 27.63 | 31.94 | 29.66 | 28.20 |
MSPNet[ | 30.76 | 28.47 | 27.03 | 31.64 | 29.40 | 27.66 | 31.99 | 29.74 | 28.34 |
MDRN[ | 30.61 | 28.27 | 26.82 | 31.41 | 29.00 | 27.37 | 31.73 | 29.44 | 28.01 |
MSANet[ | 30.67 | 28.36 | 29.96 | N/A | N/A | N/A | 31.78 | 29.57 | 28.17 |
PANet[ | 30.70 | 28.33 | 26.89 | 31.87 | 29.47 | 27.87 | 31.96 | 29.65 | 28.20 |
LGDNet | 30.79 | 28.50 | 27.04 | 31.98 | 29.78 | 28.05 | 32.02 | 29.76 | 28.37 |
方法 | BSD68 | Urban100 | Kodak24 | ||||||
---|---|---|---|---|---|---|---|---|---|
BM3D[ | 27.76 | 25.62 | 24.44 | 28.75 | 25.95 | 24.27 | 29.13 | 26.99 | 25.73 |
DnCNN[ | 28.36 | 26.23 | 24.90 | 28.88 | 26.28 | 24.36 | 29.62 | 27.51 | 26.08 |
RNAN[ | 28.61 | 26.48 | 25.18 | 30.20 | 27.65 | 25.89 | 30.04 | 27.93 | 26.60 |
RDN[ | 28.58 | 26.43 | 25.12 | 30.08 | 27.47 | 25.71 | 30.02 | 27.88 | 26.57 |
MDRN[ | N/A | 26.44 | N/A | N/A | 27.31 | N/A | N/A | N/A | N/A |
MSPNet[ | 28.64 | 26.55 | 25.31 | 30.09 | 27.64 | 25.98 | 30.06 | 28.01 | 26.59 |
MSANet[ | 28.61 | 26.51 | 25.25 | N/A | N/A | N/A | 29.91 | 27.81 | 26.54 |
LGDNet | 28.68 | 26.58 | 25.31 | 30.45 | 28.14 | 26.44 | 30.08 | 28.04 | 26.77 |
表3 不同方法在合成灰度图像数据集上的去噪结果
Tab. 3 Denoising results of different methods on synthetic grayscale image datasets
方法 | BSD68 | Urban100 | Kodak24 | ||||||
---|---|---|---|---|---|---|---|---|---|
BM3D[ | 27.76 | 25.62 | 24.44 | 28.75 | 25.95 | 24.27 | 29.13 | 26.99 | 25.73 |
DnCNN[ | 28.36 | 26.23 | 24.90 | 28.88 | 26.28 | 24.36 | 29.62 | 27.51 | 26.08 |
RNAN[ | 28.61 | 26.48 | 25.18 | 30.20 | 27.65 | 25.89 | 30.04 | 27.93 | 26.60 |
RDN[ | 28.58 | 26.43 | 25.12 | 30.08 | 27.47 | 25.71 | 30.02 | 27.88 | 26.57 |
MDRN[ | N/A | 26.44 | N/A | N/A | 27.31 | N/A | N/A | N/A | N/A |
MSPNet[ | 28.64 | 26.55 | 25.31 | 30.09 | 27.64 | 25.98 | 30.06 | 28.01 | 26.59 |
MSANet[ | 28.61 | 26.51 | 25.25 | N/A | N/A | N/A | 29.91 | 27.81 | 26.54 |
LGDNet | 28.68 | 26.58 | 25.31 | 30.45 | 28.14 | 26.44 | 30.08 | 28.04 | 26.77 |
模型 | S-A | C-A | 多阶段 | Local | FTB | PSNR/dB |
---|---|---|---|---|---|---|
29.26 | ||||||
29.30 | ||||||
26.36 | ||||||
26.38 | ||||||
26.42 | ||||||
LGDNet | 26.44 |
表4 不同模块对去噪性能的影响
Tab. 4 Effects of different modules on denoising performance
模型 | S-A | C-A | 多阶段 | Local | FTB | PSNR/dB |
---|---|---|---|---|---|---|
29.26 | ||||||
29.30 | ||||||
26.36 | ||||||
26.38 | ||||||
26.42 | ||||||
LGDNet | 26.44 |
模型 | 分支数 | 网络类型 | HTB | CAB | PSNR/dB |
---|---|---|---|---|---|
1 | U-Net | 26.12 | |||
1 | Local | 26.05 | |||
2 | U-Nets | 26.27 | |||
2 | U-Net+Local | 26.30 | |||
3 | U-Net+Locals | 26.39 | |||
3 | U-Nets+Local | 26.44 |
表5 不同分支和模块组合的性能比较
Tab. 5 Performance comparison of different branches and module combinations
模型 | 分支数 | 网络类型 | HTB | CAB | PSNR/dB |
---|---|---|---|---|---|
1 | U-Net | 26.12 | |||
1 | Local | 26.05 | |||
2 | U-Nets | 26.27 | |||
2 | U-Net+Local | 26.30 | |||
3 | U-Net+Locals | 26.39 | |||
3 | U-Nets+Local | 26.44 |
方法 | 参数量/106 | FLOPs/109 | PSNR/dB |
---|---|---|---|
MSPNet[ | 54.6 | 298.0 | 39.75 |
MIRNet[ | 31.8 | 1 572.0 | 39.88 |
MPRNet[ | 20.1 | 1 176.0 | 39.80 |
SwinIR[ | 11.8 | 788.6 | 40.01 |
Restormer[ | 26.1 | 155.0 | 40.03 |
MAXIM[ | 22.2 | 339.2 | 39.84 |
LGDNet | 18.3 | 340.2 | 40.10 |
表6 不同去噪方法的综合性能对比
Tab. 6 Comprehensive performance comparison of different denoising methods
方法 | 参数量/106 | FLOPs/109 | PSNR/dB |
---|---|---|---|
MSPNet[ | 54.6 | 298.0 | 39.75 |
MIRNet[ | 31.8 | 1 572.0 | 39.88 |
MPRNet[ | 20.1 | 1 176.0 | 39.80 |
SwinIR[ | 11.8 | 788.6 | 40.01 |
Restormer[ | 26.1 | 155.0 | 40.03 |
MAXIM[ | 22.2 | 339.2 | 39.84 |
LGDNet | 18.3 | 340.2 | 40.10 |
1 | 张雯雯, 韩裕生. 基于非局部自相似性的低秩稀疏图像去噪[J]. 计算机应用, 2018, 38(9): 2696-2700, 2746. |
ZHANG W W, HAN Y S. Nonlocal self-similarity based low-rank sparse image denoising[J]. Journal of Computer Applications, 2018, 38(9): 2696-2700, 2746. | |
2 | DABOV K, FOI A, KATKOVNIK V, et al. Color image denoising via sparse 3D collaborative filtering with grouping constraint in luminance-chrominance space [C]// Proceedings of the 2007 IEEE International Conference on Image Processing. Piscataway: IEEE, 2007: 313-316. |
3 | 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. |
4 | ZHANG K, ZUO W, CHEN Y, et al. Beyond a Gaussian denoiser: residual learning of deep CNN for image denoising[J].IEEE Transactions on Image Processing, 2017, 26(7): 3142-3155. |
5 | ANWAR S, BARNES N. Real image denoising with feature attention[C]// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway:IEEE, 2019: 3155-3164. |
6 | ZAMIR S W, ARORA A, KHAN S, et al. Learning enriched features for real image restoration and enhancement[C]// Proceedings of the 16th European Conference on Computer Vision. Cham:Springer, 2020: 492-511. |
7 | 靳华中, 张修洋, 叶志伟, 等. 基于近似U型网络结构的图像去噪模型[J]. 计算机应用, 2022, 42(8): 2571-2577. |
JIN H Z, ZHANG X Y, YE Z W, et al. Image denoising model based on approximate U-shaped network structure[J]. Journal of Computer Applications, 2022, 42(8): 2571-2577. | |
8 | 盖杉,鲍中运.基于深度学习的高噪声图像去噪算法[J].自动化学报, 2020, 46(12): 2672-2680. |
GAI S, BAO Z Y. High noise image denoising algorithm based on deep learning[J]. Acta Automatica Sinica, 2020, 46(12): 2672-2680. | |
9 | BAI Y, LIU M, YAO C, et al. MSPNet: multi-stage progressive network for image denoising[J]. Neurocomputing, 2023, 517: 71-80. |
10 | ZAMIR S W, ARORA A, KHAN S, et al. Learning enriched features for fast image restoration and enhancement[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 45(2): 1934-1948. |
11 | CHEN H, WANG Y, GUO T, et al. Pre-trained image processing transformer[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2021: 12294-12305. |
12 | LIANG J, CAO J, SUN G, et al. SwinIR: image restoration using swin Transformer[C]// Proceedings of the 2021 IEEE International Conference on Computer Vision Workshops. Piscataway:IEEE, 2021: 1833-1844. |
13 | ZAMIR S W, ARORA A, KHAN S, et al. Restormer: efficient transformer for high-resolution image restoration[C]// Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2022: 5718-5729. |
14 | WANG Z, CUN X, BAO J, et al. Uformer: a general U-shaped Transformer for image restoration[C]// Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2022: 17662-17672. |
15 | YIN H, MA S. CSformer: cross-scale features fusion based transformer for image denoising[J]. IEEE Signal Processing Letters, 2022, 29: 1809-1813. |
16 | HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2018: 7132-7141. |
17 | CHEN X, WANG X, ZHOU J, et al. Activating more pixels in image super-resolution transformer[C]// Proceedings of the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2023: 22367-22377. |
18 | ZHANG Y, TIAN Y, KONG Y, et al. Residual dense network for image restoration[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43(7): 2480-2495. |
19 | ZAMIR S W, ARORA A, KHAN S, et al. Multi-stage progressive image restoration[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway: IEEE, 2021: 14816-14826. |
20 | LIU Z, LIN Y, CAO Y, et al. Swin Transformer: hierarchical vision Transformer using shifted windows[C]// Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2021: 9992-10002. |
21 | SHAW P, USZKOREIT J, VASWANI A. Self-attention with relative position representations[EB/OL]. (2018-04-12)[2023-07-20]. . |
22 | ABDELHAMED A, LIN S, BROWN M S. A high-quality denoising dataset for smartphone cameras[C]// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 1692-1700. |
23 | PLÖTZ T, ROTH S. Benchmarking denoising algorithms with real photographs[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2017: 2750-2759. |
24 | AGUSTSSON E, TIMOFTE R. NTIRE 2017 challenge on single image super-resolution: dataset and study[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway: IEEE, 2017: 1122-1131. |
25 | LIM B, SON S, KIM H, et al. Enhanced deep residual networks for single image super-resolution[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway:IEEE, 2017: 1132-1140. |
26 | MA K, DUANMU Z, WU Q, et al. Waterloo exploration database: new challenges for image quality assessment models[J]. IEEE Transactions on Image Processing, 2017, 26(2): 1004-1016. |
27 | ARBELÁEZ P, MAIRE M, FOWLKES C, et al. Contour detection and hierarchical image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(5): 898-916. |
28 | MARTIN D, FOWLKES C, TAL D, et al. A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics[C]// Proceedings of the 8th IEEE International Conference on Computer Vision. Piscataway: IEEE, 2001: 416-423. |
29 | HUANG J-B, SINGH A, AHUJA N. Single image super-resolution from transformed self-exemplars[C]// Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE,2015: 5197-5206. |
30 | FRANZEN R. Kodak lossless true color image suite[EB/OL]. (2013-01-27)[2023-08-01]. . |
31 | GUO S, YAN Z, ZHANG K, et al. Toward convolutional blind denoising of real photographs[C]// Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway: IEEE, 2019: 1712-1722. |
32 | GOU Y, HU P, LV J, et al. Multi-scale adaptive network for single image denoising[EB/OL]. (2022-03-08)[2023-07-01]. . |
33 | SHEN H, ZHAO Z-Q, ZHANG W. Adaptive dynamic filtering network for image denoising[C]// Proceedings of the 37th AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2023: 2227-2235. |
34 | LI J, YANG H, YI Q, et al. Multiple degradation and reconstruction network for single image denoising via knowledge distillation[C]// Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2022: 555-566. |
35 | MEI Y, FAN Y, ZHANG Y, et al. Pyramid attention network for image restoration[J]. International Journal of Computer Vision, 2023, 131: 3207-3225. |
36 | ZHANG Y, LI K, LI K, et al. Residual non-local attention networks for image restoration[EB/OL]. (2019-03-24)[2023-07-01]. . |
37 | TU Z, TALEBI H, ZHANG H, et al. MAXIM: multi-axis MLP for image processing[C]// Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2022: 5759-5770. |
[1] | 贾洁茹, 杨建超, 张硕蕊, 闫涛, 陈斌. 基于自蒸馏视觉Transformer的无监督行人重识别[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2893-2902. |
[2] | 徐志刚, 张创. 基于门控位置编码的壁画图像多级色彩还原[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2931-2937. |
[3] | 黄云川, 江永全, 黄骏涛, 杨燕. 基于元图同构网络的分子毒性预测[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2964-2969. |
[4] | 杨鑫, 陈雪妮, 吴春江, 周世杰. 结合变种残差模型和Transformer的城市公路短时交通流预测[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2947-2951. |
[5] | 任烈弘, 黄铝文, 田旭, 段飞. 基于DFT的频率敏感双分支Transformer多变量长时间序列预测方法[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2739-2746. |
[6] | 方介泼, 陶重犇. 应对零日攻击的混合车联网入侵检测系统[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2763-2769. |
[7] | 李金金, 桑国明, 张益嘉. APK-CNN和Transformer增强的多域虚假新闻检测模型[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2674-2682. |
[8] | 邓凯丽, 魏伟波, 潘振宽. 改进掩码自编码器的工业缺陷检测方法[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2595-2603. |
[9] | 杨帆, 邹窈, 朱明志, 马振伟, 程大伟, 蒋昌俊. 基于图注意力Transformer神经网络的信用卡欺诈检测模型[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2634-2642. |
[10] | 李大海, 王忠华, 王振东. 结合空间域和频域信息的双分支低光照图像增强网络[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2175-2182. |
[11] | 吕锡婷, 赵敬华, 荣海迎, 赵嘉乐. 基于Transformer和关系图卷积网络的信息传播预测模型[J]. 《计算机应用》唯一官方网站, 2024, 44(6): 1760-1766. |
[12] | 黎施彬, 龚俊, 汤圣君. 基于Graph Transformer的半监督异配图表示学习模型[J]. 《计算机应用》唯一官方网站, 2024, 44(6): 1816-1823. |
[13] | 黄梦源, 常侃, 凌铭阳, 韦新杰, 覃团发. 基于层间引导的低光照图像渐进增强算法[J]. 《计算机应用》唯一官方网站, 2024, 44(6): 1911-1919. |
[14] | 孙子文, 钱立志, 杨传栋, 高一博, 陆庆阳, 袁广林. 基于Transformer的视觉目标跟踪方法综述[J]. 《计算机应用》唯一官方网站, 2024, 44(5): 1644-1654. |
[15] | 席治远, 唐超, 童安炀, 王文剑. 基于双路时空网络的驾驶员行为识别[J]. 《计算机应用》唯一官方网站, 2024, 44(5): 1511-1519. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||