Advanced cover glass defect detection and classification based on multi-DNN model

Abstract Demands for display panels and relevant technologies are rapidly increasing with the recent advances in smart mobile devices. Many manufacturers have begun slowly investing in fully automated inspection systems that enable consistent and objective inspections. Carrying out such undertaking aims to satisfy high user requirements concerning quality while coping with high volumes. Cover glass is one of the important items for inspection because users directly interact with it. Despite the extensive use of typical machine vision-based solutions in this field, many manufacturers continue relying on human-based judgment because of a deficient understanding of defects or poor confidence in algorithms. To overcome these problems, this study proposes a deep-learning neural network (DLNN)-based defect inspection system. The DLNN has advantages over traditional computer vision- or human-based inspection in terms of flexibility and performance. We introduce a weighted multi-DLNN inspection system capable of efficiently utilizing multi-channel measurement data, with a detection rate of up to 99% and a false pass rate below 1%.

[1]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[2]  Paul W. Fieguth,et al.  A review on computer vision based defect detection and condition assessment of concrete and asphalt civil infrastructure , 2015, Adv. Eng. Informatics.

[3]  Nitesh V. Chawla,et al.  SPECIAL ISSUE ON LEARNING FROM IMBALANCED DATA SETS , 2004 .

[4]  Di Li,et al.  Defect inspection and extraction of the mobile phone cover glass based on the principal components analysis , 2014 .

[5]  Sven Behnke,et al.  Accelerating Large-Scale Convolutional Neural Networks with Parallel Graphics Multiprocessors , 2010, ICANN.

[6]  Dong-Joong Kang,et al.  Machine learning-based imaging system for surface defect inspection , 2016, International Journal of Precision Engineering and Manufacturing-Green Technology.

[7]  Jian Gao,et al.  Automatic surface defect detection for mobile phone screen glass based on machine vision , 2017, Appl. Soft Comput..

[8]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[9]  Ali Farhadi,et al.  YOLO9000: Better, Faster, Stronger , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[10]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[11]  Stephen Marshall,et al.  Activation Functions: Comparison of trends in Practice and Research for Deep Learning , 2018, ArXiv.

[12]  Carlo Alberto Avizzano,et al.  A Smart Monitoring System for Automatic Welding Defect Detection , 2019, IEEE Transactions on Industrial Electronics.