Estimation of the convolutional neural network with attention mechanism and transfer learning on wood knot defect classification

doi:10.1063/5.0087060

	Estimation of the convolutional neural network with attention mechanism and transfer learning on wood knot defect classification
	Gao, Mingyu 2,3; Wang, Fei 2,3; Liu, Junyan 2,3; Song, Peng 3,4; Chen, Jianfeng 1,5,6; Yang, Hong 7; Mu, Hongbo 7; Qi, Dawei 7; Chen, Mingjun 2; Wang, Yang 2,3
刊名	JOURNAL OF APPLIED PHYSICS
	2022-06-21
卷号	131
ISSN号	0021-8979
DOI	10.1063/5.0087060
通讯作者	Wang, Fei(wangfeipublic@hit.edu.cn) ; Liu, Junyan(ljywlj@hit.edu.cn)
英文摘要	In the intelligent production process of wood products, the classification system of wood knot defects is a very practical solution. However, traditional image processing methods cannot handle it well due to the uncertainty of manually extracted features. Therefore, a lightweight and reliable artificial neural network model is proposed to classify and identify our objective. To solve this problem, a wood knot defect recognition model named SE-ResNet18 combining convolutional neural network, attention mechanism, and transfer learning is proposed in this paper. First, the Sequence-and-Exception (SE) module is combined with Basicblock and is constructed as two modules called RBBSE-1 and RBBSE-2. These modules learn to enhance features that are useful for the current task, suppress useless features, and fuse the output features with the original features. Then, the fully connected layer is replaced with a global average pooling layer, which can effectively reduce the parameters of the fully connected layer in the model. Finally, a SE-ResNet18 was constructed by one convolutional layer, five RBBSE-1 modules, and three RBBSE-2 modules of different channels. The SE-ResNet18 has a higher accuracy (98.85%) in the test set compared to the unimproved model ResNet-18. Compared with the previously proposed ReSENet-18, more SE modules are used in SE-ResNet18 to provide a basis for future training on a larger-scale dataset. Based on the same test set, a comparison with other classical models (such as LeNet-5, AlexNet, etc.) was conducted, and the results validated the superiority of the proposed model. The proposed model achieves the expected objective and provides a new way of thinking for non-destructive testing of wood. Published under an exclusive license by AIP Publishing.
资助项目	China Postdoctoral Science Foundation[2020M670902] ; China Postdoctoral Science Foundation[2021M690841] ; National Postdoctoral Program for Innovative Talents (Postdoctoral Innovation Talent Support Program of China)[BX2021092] ; Heilongjiang Postdoctoral Fund[LBHZ20156] ; Heilongjiang Postdoctoral Fund[LBH-Z20019] ; Aeronautical Science Foundation of China[2020Z057077001] ; Strategic Cooperation Program of the World Top Universities - Harbin Institute of Technology ; National Natural Science Foundation of China (NSFC)[61571153] ; National Natural Science Foundation of China (NSFC)[51173034] ; Self-planned Task of State Key Laboratory of Robotics and System (HIT) ; Program of Introducing Talents of Discipline of Universities[B07108] ; HIT Wuhu Robot Technology Research Institute
WOS关键词	DATA AUGMENTATION
WOS研究方向	Physics
语种	英语
出版者	AIP Publishing
WOS记录号	WOS:000811856900006
资助机构	China Postdoctoral Science Foundation ; National Postdoctoral Program for Innovative Talents (Postdoctoral Innovation Talent Support Program of China) ; Heilongjiang Postdoctoral Fund ; Aeronautical Science Foundation of China ; Strategic Cooperation Program of the World Top Universities - Harbin Institute of Technology ; National Natural Science Foundation of China (NSFC) ; Self-planned Task of State Key Laboratory of Robotics and System (HIT) ; Program of Introducing Talents of Discipline of Universities ; HIT Wuhu Robot Technology Research Institute
内容类型	期刊论文
源URL	[http://ir.hfcas.ac.cn:8080/handle/334002/131313]
专题	中国科学院合肥物质科学研究院
通讯作者	Wang, Fei; Liu, Junyan
作者单位	1.Chinese Acad 11 Sci, Anhui Inst Opt & Fine Mech, Key Lab Atmospher Opt, Hefei 230031, Peoples R China 2.Harbin Inst Technol, Sch Mechatron Engn, Harbin 150001, Peoples R China 3.HIT Wuhu Robot Technol Res Inst, Wuhu 241000, Peoples R China 4.Harbin Inst Technol, Sch Instrument Sci & Engn, Harbin 150001, Peoples R China 5.Univ Sci & Technol China, Sci Isl Branch Grad Sch, Hefei 230026, Peoples R China 6.Adv Laser Technol Lab Anhui Prov, Hefei 230037, Peoples R China 7.Northeast Forestry Univ, Coll Sci, Harbin 150040, Peoples R China
推荐引用方式 GB/T 7714	Gao, Mingyu,Wang, Fei,Liu, Junyan,et al. Estimation of the convolutional neural network with attention mechanism and transfer learning on wood knot defect classification[J]. JOURNAL OF APPLIED PHYSICS,2022,131.
APA	Gao, Mingyu.,Wang, Fei.,Liu, Junyan.,Song, Peng.,Chen, Jianfeng.,...&Yue, Honghao.(2022).Estimation of the convolutional neural network with attention mechanism and transfer learning on wood knot defect classification.JOURNAL OF APPLIED PHYSICS,131.
MLA	Gao, Mingyu,et al."Estimation of the convolutional neural network with attention mechanism and transfer learning on wood knot defect classification".JOURNAL OF APPLIED PHYSICS 131(2022).