面向低质量裂缝图像的多知识蒸馏分类

doi:10.19678/j.issn.1000-3428.0065323

计算机工程 ›› 2023, Vol. 49 ›› Issue (7): 204-213. doi: 10.19678/j.issn.1000-3428.0065323

面向低质量裂缝图像的多知识蒸馏分类

曹坪¹, 杨怀志², 薄一军², 尤嘉³, 张淳杰¹^,*, 李丹勇⁴

1. 北京交通大学现代信息科学与网络技术北京市重点实验室, 北京 100044
2. 京沪高速铁路股份有限公司, 北京 100038
3. 中国铁道科学研究院集团有限公司电子计算技术研究所, 北京 100081
4. 北京交通大学电子信息工程学院, 北京 100044

收稿日期:2022-07-22 出版日期:2023-07-15 发布日期:2022-10-11
通讯作者: 张淳杰
作者简介:
曹坪（1995—），女，博士研究生，主研方向为计算机视觉
杨怀志，正高级工程师
薄一军，工程师
尤嘉，高级工程师、博士
李丹勇，副教授、博士
基金资助:
国家自然科学基金面上项目(62072026); 北京市自然科学基金(JQ20022); 京沪高速铁路股份有限公司科技研究项目-重大课题(京沪科研-2020-16)

Low-quality Crack Image Classification with Multi-Knowledge Distillation

Ping CAO¹, Huaizhi YANG², Yijun BO², Jia YOU³, Chunjie ZHANG¹^,*, Danyong LI⁴

1. Beijing Key Laboratory of Modern Information Science and Network Technology, Beijing Jiaotong University, Beijing 100044, China
2. Beijing-Shanghai High Speed Railway Co., Ltd., Beijing 100038, China
3. Electronic Computing Technology Research Institute of China Academy of Railway Sciences Group Co., Ltd., Beijing 100081, China
4. School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing 100044, China

Received:2022-07-22 Online:2023-07-15 Published:2022-10-11
Contact: Chunjie ZHANG

摘要/Abstract

摘要：

裂缝受损程度分类是混凝土安全检测中重要的环节之一，在真实场景下，由于裂缝图像分辨率的降低和模糊噪声的增多，导致现有的分类方法难以识别低质量裂缝图像的受损程度。提出一种面向低质量裂缝图像的多知识蒸馏分类方法，基于退化重建网络的教师模型通过模拟高分辨率裂缝图像的退化和重建过程，缓解与学生模型之间的域差异，利用高质量模型引导的学生模型从具有高分辨率信息的教师模型中学习类别知识和重建知识。通过将教师模型的类别知识迁移到学生模型中，使教师模型的重建网络监督学生模型生成更有利于分类的超分辨率图像，确保图像恢复和图像分类之间的动态交互。经过教师模型的重建知识和类别知识的多重知识引导，使缺乏高分辨率信息的学生模型能有效识别低质量裂缝图像的受损程度。实验结果表明，与单一知识蒸馏的方法相比，该方法在最低分辨率的情况下，在裂缝数据集上的分类准确率提高了4.02个百分点，能有效提高低质量裂缝图像受损程度的分类准确率。

关键词: 低质量图像, 知识蒸馏, 裂缝图像分类, 卷积神经网络, 图像重建

Abstract:

Classification of crack damage degree is a very important step in concrete safety detection. In real scenes, it is difficult for existing classification methods to classify low-quality crack images due to the reduction of resolution and increase of blurred noise in low-resolution crack images. Therefore, a low-quality crack image classification method guided by Multi-Knowledge Distillation(MKD) is proposed. The teacher model, which is based on the degenerate reconstruction network, alleviates the domain difference with the student model by simulating the degradation and reconstruction process of high-resolution crack images. The student model, which is guided by a high-quality model, learns classification and reconstruction knowledge from the teacher model with high-resolution information. This method not only transfers the classification knowledge of the teacher model to the student model but also supervises the student model through the reconstruction network of the teacher model. In this way, the student model is guided to generate super-resolution images that are more conducive to classification, ensuring a dynamic interaction between image restoration and image classification. Through the multiple knowledge guidance of reconstruction and category knowledge of the teacher model, the student model, which lacks high-resolution information, can also effectively identify the damage degree of low-quality crack images. The experimental results show that compared with the Single Knowledge Distillation (SKD) method, this method improves the classification accuracy on the crack dataset by 4.02 percentage points at the lowest resolution and can effectively improve the classification accuracy of the damage degree of low-quality crack images.

Key words: low-quality image, knowledge distillation, crack image classification, convolutional neural network, image reconstruction

曹坪, 杨怀志, 薄一军, 尤嘉, 张淳杰, 李丹勇. 面向低质量裂缝图像的多知识蒸馏分类[J]. 计算机工程, 2023, 49(7): 204-213.

Ping CAO, Huaizhi YANG, Yijun BO, Jia YOU, Chunjie ZHANG, Danyong LI. Low-quality Crack Image Classification with Multi-Knowledge Distillation[J]. Computer Engineering, 2023, 49(7): 204-213.

https://www.ecice06.com/CN/Y2023/V49/I7/204

图/表 12

图1 多知识蒸馏引导的裂缝图像分类网络

Fig.1 Crack image classification network guided by multi-knowledge distillation

图2 不同受损程度的裂缝数据退化过程

Fig.2 Degradation process of crack data with different damage degrees

表1 不同方法在CRACK数据集上的分类准确率对比

Table 1 Comparison of classification accuracy of different methods on CRACK data set %

方法	s1尺度	s2尺度	s3尺度
Baseline方法	85.90	78.55	70.04
Baseline+SR方法	86.21	79.57	73.24
MKD方法	88.21	82.11	74.06

表2 不同蒸馏模式在CRACK数据集上的分类准确率对比

Table 2 Comparison of classification accuracy of different distillation models on CRACK data set %

蒸馏模式	s1尺度	s2尺度	s3尺度
单知识蒸馏	86.60	81.02	73.75
多知识蒸馏	88.21	82.11	74.06

图3 低质量CRACK数据集上的分类结果对比

Fig.3 Comparison of classification results on low-quality CRACK data set

表3 不同温度系数T下的分类准确率

Table 3 Classification accuracy under different temperature coefficients T

温度系数T	s1尺度	s2尺度	s3尺度
10	87.58	81.91	74.80
5	88.55	81.02	73.91
2	88.21	82.11	74.06

表4 不同重建蒸馏损失参数

$ {\mathit{\lambda }}_{2}^{\bf{S}} $

下的分类性能

Table 4 Classification performance under different reconstruction distillation loss parameters

$ {\mathit{\lambda }}_{2}^{\bf{S}} $

参数$ {\lambda }_{2}^{\mathrm{S}} $	s1尺度	s2尺度	s3尺度
0.01	85.62	80.55	73.36
0.001	86.52	81.52	74.96
0.000 1	88.21	82.11	74.06

图4 裂缝图像的重建效果

Fig.4 Reconstruction effect of crack images

表5 不同低质量图像分类方法的图像PSNR值对比

Table 5 Image PSNR values comparison of different low-quality image classification methods 单位：dB

方法	s1尺度	s2尺度	s3尺度
SKD方法	25.35	23.17	18.13
MKD方法	26.52	23.26	17.32

表6 不同重建目标引导下的分类准确率

Table 6 Classification accuracy guided by different reconstruction targets %

重建目标	s1尺度	s2尺度	s3尺度
HR图像	87.77	81.99	73.05
SR图像	88.21	82.11	74.06

表7 不同知识蒸馏权重下的分类准确率

Table 7 Classification accuracy under different knowledge distillation weights %

权重$ {\lambda }_{4}^{\mathrm{S}} $	尺度	$ {\lambda }_{2}^{\mathrm{S}}=0.01 $	$ {\lambda }_{2}^{\mathrm{S}}=0.001 $	$ {\lambda }_{2}^{\mathrm{S}}=0.000\;1 $
$ 0.35 $	s1尺度	88.16	87.50	86.87
	s2尺度	80.63	80.43	79.58
	s3尺度	72.46	72.85	73.91
$ 0.30 $	s1尺度	86.02	86.52	88.21
	s2尺度	80.55	81.52	82.11
	s3尺度	73.36	74.96	74.06
$ 0.25 $	s1尺度	86.29	87.38	87.58
	s2尺度	80.82	81.09	81.95
	s3尺度	72.93	71.80	72.19

表8 不同分类方法的分类准确率对比

Table 8 Comparison of classification accuracy with different classification methods %

方法	s1尺度	s2尺度	s3尺度
文献[25]方法	74.60	66.40	60.00
文献[49]方法	51.48	51.29	51.09
文献[50]方法	73.80	65.60	58.80
文献[51]方法	71.40	68.60	65.20
文献[52]方法	87.07	81.25	73.13
文献[53]方法	82.00	79.40	72.91
文献[54]方法	82.38	80.21	72.08
本文方法	88.21	82.11	74.06

参考文献 54

1	DAIS D, BAL I E, SMYROU E, et al. Automatic crack classification and segmentation on masonry surfaces using convolutional neural networks and transfer learning. Automation in Construction, 2021, 125 (3): 606- 615. URL
2	SAFAEI N, SMADI O, MASOUD A, et al. An automatic image processing algorithm based on crack pixel density for pavement crack detection and classification. International Journal of Pavement Research and Technology, 2022, 15 (1): 159- 172. doi: 10.1007/s42947-021-00006-4
3	LI B X, WANG K C P, ZHANG A, et al. Automatic classification of pavement crack using deep convolutional neural network. International Journal of Pavement Engineering, 2020, 21 (4): 457- 463. doi: 10.1080/10298436.2018.1485917
4	PEI Y T, HUANG Y P, ZOU Q, et al. Effects of image degradation and degradation removal to CNN-based image classification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43 (4): 1239- 1253. doi: 10.1109/TPAMI.2019.2950923
5	WU Y W, ZHANG Z M, WANG G H. Unsupervised deep feature transfer for low resolution image classification[C]//Proceedings of IEEE/CVF International Conference on Computer Vision Workshop. Washington D. C., USA: IEEE Press, 2020: 1065-1069.
6	ZHU M J, HAN K, ZHANG C, et al. Low-resolution visual recognition via deep feature distillation[C]//Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing. Washington D. C., USA: IEEE Press, 2019: 3762-3766.
7	MA T, TIAN W, XIE Y. Multi-level knowledge distillation for low-resolution object detection and facial expression recognition. Knowledge-Based Systems, 2022, 240 (4): 36- 43. doi: 10.1016/j.knosys.2022.108136
8	QI L, KUEN J, GU J X, et al. Multi-scale aligned distillation for low-resolution detection[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2021: 14438-14448.
9	TAN W M, YAN B, BARE B. Feature super-resolution: make machine see more clearly[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 3994-4002.
10	HAN K, HUANG Y, SONG C, et al. Adaptive super-resolution for person re-identification with low-resolution images. Pattern Recognition, 2021, 114 (7): 82- 93. doi: 10.1016/j.patcog.2020.107682
11	LU Z, JIANG X D, KOT A. Deep coupled ResNet for low-resolution face recognition. IEEE Signal Processing Letters, 2018, 25 (4): 526- 530. doi: 10.1109/LSP.2018.2810121
12	CHEN H, PEI Y, ZHAO H, et al. Super-resolution guided knowledge distillation for low-resolution image classification. Pattern Recognition Letters, 2022, 155 (12): 62- 68.
13	HARALICK R M, SHANMUGAM K, DINSTEIN I. Textural features for image classification. IEEE Transactions on Systems, Man, and Cybernetics, 1973, 3 (6): 610- 621. doi: 10.1109/TSMC.1973.4309314
14	BANERJI S, SINHA A, LIU C. New image descriptors based on color, texture, shape, and wavelets for object and scene image classification. Neurocomputing, 2013, 117 (7): 173- 185. doi: 10.1016/j.neucom.2013.02.014
15	LOWE D G. Distinctive image features from scale-invariant keypoints. International Journal of Computer Vision, 2004, 60 (2): 91- 110. doi: 10.1023/B:VISI.0000029664.99615.94
16	DALAL N, TRIGGS B. Histograms of oriented gradients for human detection[C]//Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2005: 886-893.
17	BAY H, TUYTELAARS T, VAN GOOL L. SURF: speeded up robust features[M]//LEONARDIS A, BISCHOF H, PINZ A. Computer Vision-ECCV 2006. Berlin, Germany: Springer, 2006: 404-417.
18	KHAN A, SOHAIL A, ZAHOORA U, et al. A survey of the recent architectures of deep convolutional neural networks. Artificial Intelligence Review, 2020, 53 (8): 5455- 5516. doi: 10.1007/s10462-020-09825-6
19	KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks. Communications of the ACM, 2017, 60 (6): 84- 90. doi: 10.1145/3065386
20	HOWARD A G, ZHU M, CHEN B, et al. MobileNets: efficient convolutional neural networks for mobile vision applications[EB/OL]. [2022-06-20]. https://arxiv.org/abs/1704.04861.
21	SANDLER M, HOWARD A, ZHU M L, et al. MobileNetV2: inverted residuals and linear bottlenecks[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 4510-4520.
22	ZHANG X Y, ZHOU X Y, LIN M X, et al. ShuffleNet: an extremely efficient convolutional neural network for mobile devices[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 6848-6856.
23	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2016: 770-778.
24	FU J, LIU J, TIAN H J, et al. Dual attention network for scene segmentation[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2020: 3141-3149.
25	LIU Z, LIN Y T, CAO Y, et al. Swin transformer: hierarchical vision transformer using shifted windows[C]//Proceedings of IEEE/CVF International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2022: 9992-10002.
26	DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16×16 words: transformers for image recognition at scale[EB/OL]. [2022-06-20]. https://arxiv.org/abs/2010.11929.
27	SRINIVAS A, LIN T Y, PARMAR N, et al. Bottleneck transformers for visual recognition[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2021: 16514-16524.
28	HARRIS J L. Diffraction and resolving power. Journal of the Optical Society of America, 1964, 54 (7): 931- 942. doi: 10.1364/JOSA.54.000931
29	TORALDO DI FRANCIA G. Resolving power and information. Journal of the Optical Society of America, 1955, 45 (7): 497- 503. doi: 10.1364/JOSA.45.000497
30	VANDEWALLE P, SBAIZ L, VANDEWALLE J, et al. Super-resolution from unregistered and totally aliased signals using subspace methods. IEEE Transactions on Signal Processing, 2007, 55 (7): 3687- 3703. doi: 10.1109/TSP.2007.894257
31	REN C, HE X H, NGUYEN T Q. Single image super-resolution via adaptive high-dimensional non-local total variation and adaptive geometric feature. IEEE Transactions on Image Processing, 2017, 26 (1): 90- 106. doi: 10.1109/TIP.2016.2619265
32	LIU C, SUN D Q. On Bayesian adaptive video super resolution. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36 (2): 346- 360. doi: 10.1109/TPAMI.2013.127
33	YANG D Q, LI Z M, XIA Y T, et al. Remote sensing image super-resolution: challenges and approaches[C]//Proceedings of 2015 IEEE International Conference on Digital Signal Processing. Washington D. C., USA: IEEE Press, 2015: 196-200.
34	DONG C, LOY C C, HE K M, et al. Learning a deep convolutional network for image super-resolution[M]//FLEET D, PAJDLA T, SCHIELE B, et al. Computer Vision-ECCV 2014. Berlin, Germany: Springer, 2014: 184-199.
35	WANG Z H, CHEN J, HOI S C H. Deep learning for image super-resolution: a survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43 (10): 3365- 3387. doi: 10.1109/TPAMI.2020.2982166
36	KIM J, LEE J K, LEE K M. Accurate image super-resolution using very deep convolutional networks[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA. IEEE Press, 2016: 1646-1654.
37	CHAN K C K, ZHOU S C, XU X Y, et al. BasicVSR: improving video super-resolution with enhanced propagation and alignment[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA. IEEE Press, 2022: 5962-5971.
38	LYU Q, SHAN H M, STEBER C, et al. Multi-contrast super-resolution MRI through a progressive network. IEEE Transactions on Medical Imaging, 2020, 39 (9): 2738- 2749. doi: 10.1109/TMI.2020.2974858
39	GU J J, LU H N, ZUO W M, et al. Blind super-resolution with iterative kernel correction[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA. IEEE Press, 2020: 1604-1613.
40	HINTON G, VINYALS O, DEAN J. Distilling the knowledge in a neural network[EB/OL]. [2022-06-20]. https://arxiv.org/abs/1503.02531.
41	黄震华, 杨顺志, 林威, 等. 知识蒸馏研究综述. 计算机学报, 2022, 45 (3): 624- 653. URL
	HUANG Z H, YANG S Z, LIN W, et al. Knowledge distillation: a survey. Chinese Journal of Computers, 2022, 45 (3): 624- 653. URL
42	AFOURAS T, CHUNG J S, ZISSERMAN A. ASR is all You need: cross-modal distillation for lip reading[C]//Proceedings of ICASSP 2020 IEEE International Conference on Acoustics, Speech and Signal Processing. Washington D. C., USA: IEEE Press, 2020: 2143-2147.
43	YE H J, LU S, ZHAN D C. Distilling cross-task knowledge via relationship matching[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2020: 12393-12402.
44	GAO Q Q, ZHAO Y, LI G, et al. Image super-resolution using knowledge distillation[M]//JAWAHAR C V, LI H D, MORI G, et al. Computer Vision-ACCV 2018. Berlin, Germany: Springer, 2019: 527-541.
45	LIU Y S, ZHANG L Q, HAN Z Z, et al. Integrating knowledge distillation with learning to rank for few-shot scene classification. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60 (7): 1- 12.
46	HUI Z, GAO X B, YANG Y C, et al. Lightweight image super-resolution with information multi-distillation network[C]//Proceedings of the 27th ACM International Conference on Multimedia. New York, USA: ACM Press, 2019: 2024-2032.
47	LIU J, TANG J, WU G S. Residual feature distillation network for lightweight image super-resolution[M]//BARTOLI A, FUSIELLO A. Computer Vision-ECCV 2020 Workshops. Berlin, Germany: Springer, 2020: 41-55.
48	LEE W, LEE J, KIM D, et al. Learning with privileged information for efficient image super-resolution[M]//WONKYUNG L, JUNGHYUP, DOHYUNG K, et al. Computer Vision-ECCV 2020. Berlin, Germany: Springer, 2020: 465-482.
49	TOUVRON H, BOJANOWSKI P, CARON M, et al. ResMLP: feedforward networks for image classification with data-efficient training[EB/OL]. [2022-06-20]. https://arxiv.org/abs/2105.03404.
50	ZHANG J N, PENG H W, WU K, et al. MiniViT: compressing vision transformers with weight multiplexing[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2022: 12135-12144.
51	GRAHAM B, EL-NOUBY A, TOUVRON H, et al. LeViT: a vision transformer in ConvNet's clothing for faster inference[C]//Proceedings of IEEE/CVF International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2022: 12239-12249.
52	LENG Z Q, TAN M X, LIU C X, et al. PolyLoss: a polynomial expansion perspective of classification loss functions[EB/OL]. [2022-06-20]. https://arxiv.org/abs/2204.12511.
53	LIU Z, MAO H Z, WU C Y, et al. A ConvNet for the 2020s[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA. IEEE Press, 2022: 11966-11976.
54	DAI Z H, LIU H X, LE Q V, et al. CoAtNet: marrying convolution and attention for all data sizes[EB/OL]. [2022-06-20]. https://arxiv.org/abs/2106.04803.

[1]	耿丽丽, 牛保宁. 基于通道相似度熵的卷积神经网络裁剪[J]. 计算机工程, 2024, 50(7): 133-143.
[2]	张洋, 刘畅, 李少青. 基于可控制性度量的图神经网络门级硬件木马检测方法[J]. 计算机工程, 2024, 50(7): 164-173.
[3]	牛瑞婷, 严天峰, 高锐, 王映植. 低信噪比下基于深度学习TCNN-MobileNet的调制识别[J]. 计算机工程, 2024, 50(7): 204-215.
[4]	屠乃威, 焦猛, 阎馨. 复杂环境下输电线路鸟巢目标图像检测模型[J]. 计算机工程, 2024, 50(7): 216-226.
[5]	张溢文, 蔡满春, 陈咏豪, 朱懿, 姚利峰. 融合空间特征的多尺度深度伪造检测方法[J]. 计算机工程, 2024, 50(7): 240-250.
[6]	逯焕宇, 张永宏, 马光义, 谢东林, 田伟. 基于半监督对抗学习的遥感图像水体提取[J]. 计算机工程, 2024, 50(7): 251-263.
[7]	于洋, 孙芳芳, 吕华, 李扬, 王晓民. 基于多尺度时空注意力网络的微表情检测方法[J]. 计算机工程, 2024, 50(6): 228-235.
[8]	张雷, 沈国琛, 欧冬秀. 用于热成像数据的卷积神经网络特征图筛选方法[J]. 计算机工程, 2024, 50(4): 31-40.
[9]	张雷, 沈国琛, 欧冬秀. 用于热成像数据的卷积神经网络特征图筛选方法[J]. 计算机工程, 2024, 50(4): 31-40.
[10]	李政学, 李枝名, 彭德中, 陈杰. 基于特征对比学习和图卷积的社交网络用户分类[J]. 计算机工程, 2024, 50(4): 258-266.
[11]	姜百浩, 刘静, 仇大伟, 姜良. 深度学习在脊柱图像分割中的应用综述[J]. 计算机工程, 2024, 50(3): 1-15.
[12]	王柏涵, 姜晓燕, 范柳伊. 基于深度监督隐空间构建的语义分割改进方法[J]. 计算机工程, 2024, 50(3): 191-199.
[13]	谢新林, 尹东旭, 张涛源, 谢刚. 基于注意力机制的多尺度融合人群计数算法[J]. 计算机工程, 2024, 50(3): 290-297.
[14]	赵佳圆, 张玉茹, 苏晓东, 徐红岩, 李世洲, 张玉荣. 基于注意力机制的人体关键点隐式建模网络[J]. 计算机工程, 2024, 50(3): 317-325.
[15]	徐浩宸, 刘满华. 基于多层次自注意力网络的人脸特征点检测[J]. 计算机工程, 2024, 50(2): 239-246.

选择文件类型/文献管理软件名称

选择包含的内容

面向低质量裂缝图像的多知识蒸馏分类

Low-quality Crack Image Classification with Multi-Knowledge Distillation

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 54

相关文章 15

编辑推荐

Metrics

本文评价

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

面向低质量裂缝图像的多知识蒸馏分类

Low-quality Crack Image Classification with Multi-Knowledge Distillation

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 54

相关文章 15

编辑推荐

Metrics

本文评价