基于卷积神经网络的短切毡缺陷分类
|
摘要
基于卷积神经网络,提出了短切毡缺陷分类的方法。通过旋转、平移和翻转对数据集进行扩充,解决了小数据样本在深度卷积神经网络中的过拟合问题;利用迁移学习的思想加速网络收敛,提高了网络的泛化能力;对比了不同网络结构并选择较好的网络进行数据集验证。结果表明,所提方法能够实现短切毡缺陷的有效分类,准确率为93%。
In this study,a classification method of chopped strand mat defects based on convolutional neural network is proposed.In the proposed method,the rotation,translation,and inversion are employed to expand the dataset for solving the overfitting problem caused by the small data samples in the deep convolutional neural networks.Transfer learning is employed to improve the convergence speed and generalization ability of the network.Further,the different network structures are compared,and the most optimal network structure is used to verify the database.The experimental results demonstrate that the proposed method can effectively classify the chopped strand mat defects with an accuracy rate of 93%.
In this study,a classification method of chopped strand mat defects based on convolutional neural network is proposed.In the proposed method,the rotation,translation,and inversion are employed to expand the dataset for solving the overfitting problem caused by the small data samples in the deep convolutional neural networks.Transfer learning is employed to improve the convergence speed and generalization ability of the network.Further,the different network structures are compared,and the most optimal network structure is used to verify the database.The experimental results demonstrate that the proposed method can effectively classify the chopped strand mat defects with an accuracy rate of 93%.
引文
[1] Deng H P,Bo S Q.Correlation offiberglass chopped strand mat strength with its binder[J].Fiber Glass,2015(2):18-21.邓和平,薄圣泉.玻璃纤维短切原丝毡中粘结剂和毡的强度关系[J].玻璃纤维,2015(2):18-21.
[2] Zhang J Y,Jing J F.Defect classification of chopped strand mats based on deep learning and piecewise linear interpolation[J].Journal of Xi′an Polytechnic University,2018,32(5):553-559.张君扬,景军锋.基于深度学习和分段线性插值的短切毡缺陷分类[J].西安工程大学学报,2018,32(5):553-559.
[3] Szegedy C,Liu W,Jia Y Q,et al.Going deeper with convolutions[C]∥2015IEEE Conference on Computer Vision and Pattern Recognition(CVPR),7-12June 2015,Boston,MA,USA.New York:IEEE,15523970.
[4] Du J, Hu B L, Zhang Z F. Gastriccarcinoma classification based on convolutional neural network and micro-hyperspectral imaging[J].Acta Optica Sinica,2018,38(6):0617001.杜剑,胡炳樑,张周锋.基于卷积神经网络与显微高光谱的胃癌组织分类方法研究[J].光学学报,2018,38(6):0617001.
[5] Chu J H, Wu Z R,LüW,et al.Breast cancer diagnosis system based on transfer learning and deep convolutional neural networks[J]. Laser&Optoelectronics Progress,2018,55(8):081001.褚晶辉,吴泽蕤,吕卫,等.基于迁移学习和深度卷积神经网络的乳腺肿瘤诊断系统[J].激光与光电子学进展,2018,55(8):081001.
[6] Jing J F,Fan X T,Li P F.Patterned fabric defect detection via convolutional matching pursuit dualdictionary[J].Optical Engineering,2016,55(5):053109.
[7] Fang X, Wang G H,Yang H C,et al. High resolutionremotesensingimageclassification combining with mean-shift segmentation and fully convolutionneuralnetwork[J]. Laser&Optoelectronics Progress,2018,55(2):022802.方旭,王光辉,杨化超,等.结合均值漂移分割与全卷积神经网络的高分辨遥感影像分类[J].激光与光电子学进展,2018,55(2):022802.
[8] Lai W H,Zhou M R,Wang Y,et al.Application of counterfeit liquor recognition based on deep learning and laser induced fluorescence[J]. Laser&Optoelectronics Progress,2018,55(4):043001.来文豪,周孟然,王亚,等.深度学习与激光诱导荧光在假酒识别中的应用[J].激光与光电子学进展,2018,55(4):043001.
[9] LeCun Y, Boser B, Denker JS,et al.Backpropagation applied to handwritten zip code recognition[J].Neural Computation,1989,1(4):541-551.
[10] Yang W,Wanli O Y,Li H S,et al.End-to-end learning of deformable mixture of parts and deep convolutional neural networks for human pose estimation[C]∥2016IEEE Conference on Computer Vision and Pattern Recognition(CVPR),27-30June2016,Las Vegas,NV,USA.New York:IEEE,16541427.
[11] Zeiler M D,Fergus R.Visualizing and understanding convolutional networks[M]∥Zeiler M D,Fergus R.eds.Computer Vision-ECCV 2014.Cham:Springer International Publishing,2014:818-833.
[12] Krizhevsky A,Sutskever I,Hinton G E.ImageNet classification with deep convolutional neural networks[C]∥International Conference on Neural Information Processing Systems.Dickinson:Curran Associates Inc.2012:1097-1105.
[13] Hearst M A,Dumais S T,Osuna E,et al.Support vector machines[J].IEEE Intelligent Systems and their Applications,1998,13(4):18-28.
[14] Oquab M,Bottou L,Laptev I,et al.Learning and transferring mid-level image representations using convolutional neural networks[C]∥2014 IEEE ConferenceonComputerVisionandPattern Recognition,23-28 June 2014, Columbus, OH,USA.New York:IEEE,14632209.
[15] Donahue J,Hendricks L A,Rohrbach M,et al.Long-term recurrent convolutional networks for visual recognition and description[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(4):677-691.
[16] Wang Z,Crammer K,Vucetic S.Breaking the curse of kernelization:budgeted stochastic gradient descent for large-scale SVM training[J].Journal of Machine Learning Research,2012,13(1):3103-3131.
[2] Zhang J Y,Jing J F.Defect classification of chopped strand mats based on deep learning and piecewise linear interpolation[J].Journal of Xi′an Polytechnic University,2018,32(5):553-559.张君扬,景军锋.基于深度学习和分段线性插值的短切毡缺陷分类[J].西安工程大学学报,2018,32(5):553-559.
[3] Szegedy C,Liu W,Jia Y Q,et al.Going deeper with convolutions[C]∥2015IEEE Conference on Computer Vision and Pattern Recognition(CVPR),7-12June 2015,Boston,MA,USA.New York:IEEE,15523970.
[4] Du J, Hu B L, Zhang Z F. Gastriccarcinoma classification based on convolutional neural network and micro-hyperspectral imaging[J].Acta Optica Sinica,2018,38(6):0617001.杜剑,胡炳樑,张周锋.基于卷积神经网络与显微高光谱的胃癌组织分类方法研究[J].光学学报,2018,38(6):0617001.
[5] Chu J H, Wu Z R,LüW,et al.Breast cancer diagnosis system based on transfer learning and deep convolutional neural networks[J]. Laser&Optoelectronics Progress,2018,55(8):081001.褚晶辉,吴泽蕤,吕卫,等.基于迁移学习和深度卷积神经网络的乳腺肿瘤诊断系统[J].激光与光电子学进展,2018,55(8):081001.
[6] Jing J F,Fan X T,Li P F.Patterned fabric defect detection via convolutional matching pursuit dualdictionary[J].Optical Engineering,2016,55(5):053109.
[7] Fang X, Wang G H,Yang H C,et al. High resolutionremotesensingimageclassification combining with mean-shift segmentation and fully convolutionneuralnetwork[J]. Laser&Optoelectronics Progress,2018,55(2):022802.方旭,王光辉,杨化超,等.结合均值漂移分割与全卷积神经网络的高分辨遥感影像分类[J].激光与光电子学进展,2018,55(2):022802.
[8] Lai W H,Zhou M R,Wang Y,et al.Application of counterfeit liquor recognition based on deep learning and laser induced fluorescence[J]. Laser&Optoelectronics Progress,2018,55(4):043001.来文豪,周孟然,王亚,等.深度学习与激光诱导荧光在假酒识别中的应用[J].激光与光电子学进展,2018,55(4):043001.
[9] LeCun Y, Boser B, Denker JS,et al.Backpropagation applied to handwritten zip code recognition[J].Neural Computation,1989,1(4):541-551.
[10] Yang W,Wanli O Y,Li H S,et al.End-to-end learning of deformable mixture of parts and deep convolutional neural networks for human pose estimation[C]∥2016IEEE Conference on Computer Vision and Pattern Recognition(CVPR),27-30June2016,Las Vegas,NV,USA.New York:IEEE,16541427.
[11] Zeiler M D,Fergus R.Visualizing and understanding convolutional networks[M]∥Zeiler M D,Fergus R.eds.Computer Vision-ECCV 2014.Cham:Springer International Publishing,2014:818-833.
[12] Krizhevsky A,Sutskever I,Hinton G E.ImageNet classification with deep convolutional neural networks[C]∥International Conference on Neural Information Processing Systems.Dickinson:Curran Associates Inc.2012:1097-1105.
[13] Hearst M A,Dumais S T,Osuna E,et al.Support vector machines[J].IEEE Intelligent Systems and their Applications,1998,13(4):18-28.
[14] Oquab M,Bottou L,Laptev I,et al.Learning and transferring mid-level image representations using convolutional neural networks[C]∥2014 IEEE ConferenceonComputerVisionandPattern Recognition,23-28 June 2014, Columbus, OH,USA.New York:IEEE,14632209.
[15] Donahue J,Hendricks L A,Rohrbach M,et al.Long-term recurrent convolutional networks for visual recognition and description[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(4):677-691.
[16] Wang Z,Crammer K,Vucetic S.Breaking the curse of kernelization:budgeted stochastic gradient descent for large-scale SVM training[J].Journal of Machine Learning Research,2012,13(1):3103-3131.