TY - GEN
T1 - The Classification of Wafer Defects
T2 - 10th International Conference on Robot Intelligence Technology and Applications, RiTA 2022
AU - Xuen, Lim Shi
AU - Mohd Khairuddin, Ismail
AU - Mohd Razman, Mohd Azraai
AU - Mat Jizat, Jessnor Arif
AU - Yuen, Edmund
AU - Jiang, Haochuan
AU - Yap, Eng Hwa
AU - P. P. Abdul Majeed, Anwar
N1 - Funding Information:
Acknowledgement. The authors would like to thank IdealVision Sdn Bhd for providing the image dataset as well as Universiti Malaysia Pahang for funding the study via UIC200815 and RDU202404.
Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2023
Y1 - 2023
N2 - Wafer defect detection is a non-trivial issue in the semiconductor industry. Conventional means of defect detection is often labor-intensive based that is prone to error owing to a myriad of issue. Hence, there is push toward automatic defect detection in the industry. This work shall investigate the efficacy of a transfer learning pipeline that consists of different pre-trained DenseNet convolutional neural network models in which its fully connected layer is swapped with different support vector machine (SVM) models in classifying the defect state of a wafer whether it passes or fail. The optimal hyperparameters are identified via the grid search technique. It was shown from the present investigation that the features extracted via the DenseNet121 transfer learning model with a linear-based SVM model with a C and gamma parameter of 0.01, respectively, could yield a validation and test classification accuracy of 93% and 86%, respectively on a stratified 60:20:20 data split ratio. The result from the present study demonstrates that the proposed pipeline is able to classify the defect level of the wafer well.
AB - Wafer defect detection is a non-trivial issue in the semiconductor industry. Conventional means of defect detection is often labor-intensive based that is prone to error owing to a myriad of issue. Hence, there is push toward automatic defect detection in the industry. This work shall investigate the efficacy of a transfer learning pipeline that consists of different pre-trained DenseNet convolutional neural network models in which its fully connected layer is swapped with different support vector machine (SVM) models in classifying the defect state of a wafer whether it passes or fail. The optimal hyperparameters are identified via the grid search technique. It was shown from the present investigation that the features extracted via the DenseNet121 transfer learning model with a linear-based SVM model with a C and gamma parameter of 0.01, respectively, could yield a validation and test classification accuracy of 93% and 86%, respectively on a stratified 60:20:20 data split ratio. The result from the present study demonstrates that the proposed pipeline is able to classify the defect level of the wafer well.
KW - DenseNet
KW - Transfer learning
KW - Wafer inspection
UR - http://www.scopus.com/inward/record.url?scp=85151066474&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-26889-2_27
DO - 10.1007/978-3-031-26889-2_27
M3 - Conference Proceeding
AN - SCOPUS:85151066474
SN - 9783031268885
T3 - Lecture Notes in Networks and Systems
SP - 304
EP - 309
BT - Robot Intelligence Technology and Applications 7 - Results from the 10th International Conference on Robot Intelligence Technology and Applications
A2 - Jo, Jun
A2 - Choi, Han-Lim
A2 - Helbig, Marde
A2 - Oh, Hyondong
A2 - Hwangbo, Jemin
A2 - Lee, Chang-Hun
A2 - Stantic, Bela
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 7 December 2022 through 9 December 2022
ER -