作者：刘桂雄1, 何彬媛1, 吴俊芳2, 林镇秋3
作者单位：1. 华南理工大学机械与汽车工程学院, 广东 广州 510640;
2. 华南理工大学物理与光电学院, 广东 广州 510640;
3. 广州市华颉电子科技有限公司, 广东 广州 510663
关键词：图像分类;深度学习;视觉检测;零部件质量检测
摘要：
基于深度学习图像分类是视觉检测应用的基本任务。该文系统总结基于模型深度化图像分类网络、基于模型轻量化图像分类网络及其他优化网络主要思想、网络结构、实现技术、技术指标、应用场景，指出网络模型深度化、轻量化分别有助于提高图像分类准确性、实时性。最后，面向零部件质量检测需求，应根据其类型多少、结构复杂程度、特征异同等特点，结合实时性要求，选择合适的图像分类网络构建零部件质量智能检测系统。
Deep learning image classification network for visual inspection and its application in components quality test
LIU Guixiong1, HE Binyuan1, WU Junfang2, LIN Zhenqiu3
1. School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, China;
2. School of Physics and Optoelectronics, South China University of Technology, Guangzhou 510640, China;
3. Guangzhou Hua Jie Electronic Technology Co., Ltd., Guangzhou 510663, China
Abstract: Deep learning-based image classification is a precondition of visual inspection and its application. Image classification networks could be divide into deeper model-based, smaller model-based and other optimization-based three classes. The differences among major approach, network structures, deployment, technical indicators, application scenarios of these method are compared. When network becoming deeper and smaller, the accuracy and real-time performance of image classification have improved, respectively. The complexity of components, number of classes, similarities or differences of visual features, and real-time requirement should be taken into account for building up a deep learning visual inspection-based components quality test system.
Keywords: image classification;deep learning;visual inspection;components quality test
2019, 45(7):1-10 收稿日期: 2019-05-04;收到修改稿日期: 2019-06-02
基金项目: 广州市产学研重大项目（201802030006）；广东省现代几何与力学计量技术重点实验室开放课题（SCMKF201801）
作者简介: 刘桂雄(1968-),男,广东揭阳市人,教授,博导,主要从事先进传感与仪器研究
参考文献
[1] FREEMAN W T, ADELSON E H. The design and use of steerable filters[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 1991(9):891-906
[2] ZHANG D, LIU B, SUN C, et al. Learning the classifier combination for image classification[J]. Journal of Computers, 2011, 6(8):1756-1763
[3] BAY H, TUYTELAARS T, VAN G L. Surf:Speeded up robust features[C]//Springer. European Conference on Computer Vision, 2006:404-417.
[4] ITTI L, KOCH C, NIEBUR E. A model of saliency-based visual attention for rapid scene analysis[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 1998(11):1254-1259
[5] GUO Y, LIU Y, OERLEMANS A, et al. Deep learning for visual understanding:A review[J]. Neurocomputing, 2016, 187(C):27-48
[6] MENARD S. Applied logistic regression analysis[M]. London:Sage Publications, 2002:61-80.
[7] AMATO G, FALCHI F.On kNN classification and local feature based image similarity functions[C]//International Conference on Agents and Artificial Intelligence, 2011:224-239.
[8] BREIMAN L. Random forests[J]. Machine Learning, 2001, 45(1):5-32
[9] BAY H, ESS A, TUYTELAARS T, et al. Speeded-up robust features (SURF)[J]. Computer Vision and Image Understanding, 2008, 110(3):346-359
[10] DÉNIZ O, BUENO G, SALIDO J, et al. Face recognition using histograms of oriented gradients[J]. Pattern Recognition Letters, 2011, 32(12):1598-1603
[11] MARON M E, KUHNS J L. On relevance, probabilistic indexing and information retrieval[J]. Journal of the ACM (JACM), 1960, 7(3):216-244
[12] JOACHIMS T. Making large-scale SVM learning practical[R]. Universität Dortmund, 1998.
[13] LU D, WENG Q. A survey of image classification methods and techniques for improving classification performance[J]. International Journal of Remote Sensing, 2007, 28(5):823-870
[14] 罗建豪, 吴建鑫. 基于深度卷积特征的细粒度图像分类研究综述[J]. 自动化学报, 2017, 43(8):1306-1318
[15] RAWAT W, WANG Z. Deep convolutional neural networks for image classification:A comprehensive review[J]. Neural Computation, 2017, 29(9):2352-2449
[16] 杨真真, 匡楠, 范露, 等. 基于卷积神经网络的图像分类算法综述[J]. 信号处理, 2018, 34(12):1474-1489
[17] YANG L P, MACEACHREN A, MITRA P, et al. Visually-enabled active deep learning for (geo) text and image classification:a review[J]. ISPRS International Journal of Geo-Information, 2018, 7(2):65-103
[18] 田萱, 王亮, 丁琪. 基于深度学习的图像语义分割方法综述[J]. 软件学报, 2019(2):440-468
[19] LIU X, HOU F, QIN H, et al. Multi-view multi-scale CNNs for lung nodule type classification from CT images[J]. Pattern Recognition, 2018, 77:262-275
[20] TONUTTI M, RUFFALDI E, CATTANEO A, et al. Robust and subject-independent driving manoeuvre anticipation through Domain-Adversarial Recurrent Neural Networks[J]. Robotics and Autonomous Systems, 2019, 115:162-173
[21] 李云鹏, 侯凌燕, 王超. 基于YOLOv3的自动驾驶中运动目标检测[J]. 计算机工程与设计, 2019(4):1139-1144
[22] 马晓云, 朱丹, 金晨,等. 基于改进Faster R-CNN的子弹外观缺陷检测[J/OL]. 激光与光电子学进展:1-14[2019-05-04].http://kns.cnki.net/kcms/detail/31.1690.TN.20190308.1705.004.html.
[23] 常海涛, 苟军年, 李晓梅. Faster R-CNN在工业CT图像缺陷检测中的应用[J]. 中国图象图形学报, 2018, 23(7):1061-1071
[24] IWAHORI Y, TAKADA Y, SHⅡNA T, et al. Defect Classification of Electronic Board Using Dense SIFT and CNN[J]. Procedia Computer Science, 2018, 126:1673-1682
[25] 赵浩如, 张永, 刘国柱. 基于RPN与B-CNN的细粒度图像分类算法研究[J]. 计算机应用与软件, 2019, 36(3):210-213,264
[26] 王陈光, 王晋疆, 赵显庭. 低对比度图像特征点提取与匹配[J]. 半导体光电, 2017, 38(6):888-892,897
[27] RUSSAKOVSKY O, DENG J, SU H, et al. Imagenet large scale visual recognition challenge[J]. International Journal of Computer Vision, 2015, 115(3):211-252
[28] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once:Unified, real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2016:779-788.
[29] LAW H, DENG J. Cornernet:Detecting objects as paired keypoints[C]//Proceedings of the European Conference on Computer Vision (ECCV), 2018:734-750.
[30] CHOLLET F. Xception:Deep learning with depthwise separable convolutions[C]//Proceedings of the IEEE conference on computer vision and pattern recognition, 2017:1251-1258.
[31] JIA Y, SHELHAMER E, DONAHUE J, et al. Caffe:Convolutional architecture for fast feature embedding[C]//Proceedings of the 22nd ACM International Conference on Multimedia, 2014:675-678.
[32] HE K, SUN J. Convolutional neural networks at constrained time cost[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015:5353-5360.
[33] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. Imagenet classification with deep convolutional neural networks[C]//Advances in Neural Information Processing Systems, 2012:1097-1105.
[34] ZEILER M D, FERGUS R. Visualizing and understanding convolutional networks[C]//European Conference on Computer Vision, 2013:818-833.
[35] LIN T Y, MAJI S. Visualizing and understanding deep texture representations[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016:2791-2799.
[36] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[C]//2015 International Conference on Learning Representations, 2015:1-14.
[37] BENGIO Y, SIMARD P, FRASCONI P. Learning long-term dependencies with gradient descent is difficult[J]. IEEE Transactions on Neural Networks, 1994, 5(2):157-166
[38] SZEGEDY C, LIU W, JIA Y, et al. Going deeper with convolutions[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015.
[39] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016:770-778.
[40] ZHANG X, ZHOU X, LIN M, et al. Shufflenet:An extremely efficient convolutional neural network for mobile devices[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018:6848-6856.
[41] HUANG G, LIU Z, VAN DER MAATEN L, et al. Densely connected convolutional networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017.
[42] SANDLER M, HOWARD A, ZHU M, et al. Mobilenetv2:Inverted residuals and linear bottlenecks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018:4510-4520.
[43] IOFFE S, SZEGEDY C. Batch normalization:Accelerating deep network training by reducing internal covariate shift[C]//International Conference on Machine Learning, 2015:448-456.
[44] SZEGEDY C, IOFFE S, VANHOUCKE V, et al. Inception-v4, inception-resnet and the impact of residual connections on learning[C]//Thirty-First AAAI Conference on Artificial Intelligence, 2017.
[45] CHEN Y, LI J, XIAO H, et al. Dual path networks[C]//Advances in Neural Information Processing Systems, 2017:4467-4475.
[46] TAO X, ZHANG D, MA W, et al. Automatic metallic surface defect detection and recognition with convolutional neural networks[J]. Applied Sciences, 2018, 8(9):1575
[47] SHIPWAY N J, BARDEN T J, HUTHWAITE P, et al. Automated defect detection for fluorescent penetrant inspection using random forest[J]. NDT & E International, 2019, 101:113-123
[48] 郭雪梅, 刘桂雄, 黄坚, 等. 面向标准件装配质量的PI-SURF检测区域划分技术[J]. 中国测试, 2017, 43(8):101-105
[49] 黄坚, 刘桂雄, 林镇秋. 基于多角点结合的机箱标准件图像特征提取方法[J]. 中国测试, 2017, 43(9):123-127
[50] 李宜汀, 谢庆生, 黄海松,等. 基于Faster R-CNN的表面缺陷检测方法研究[J/OL]. 计算机集成制造系统:1-19[2019-05-04].http://kns.cnki.net/kcms/detail/11.5946.tp.20190110.1415.002.html.
[51] DENG Y S, LUO A C, DAI M J. Building an Automatic Defect Verification System Using Deep Neural Network for PCB Defect Classification[C]//20184th International Conference on Frontiers of Signal Processing (ICFSP), 2018.
[52] 何彬媛, 黄坚, 刘桂雄, 等. 面向机箱标准件装配质量局部特征的智能检测技术[J]. 中国测试, 2019, 45(3):18-23
