TY - GEN
T1 - Feature-Based Transfer Learning for IoT-Enabled Defect Detection for Quality Control in Industrial Manufacturing Processes
T2 - International Conference on Intelligent Manufacturing and Robotics, ICIMR 2023
AU - P. P. Abdul Majeed, Anwar
AU - Ateeq, Muhammad
AU - Hu, Bintao
AU - Isa, Wan Hasbullah Mohd
AU - Omar, Zaid
AU - Chen, Wei
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.
PY - 2024
Y1 - 2024
N2 - Surface defects in industrial products can lead to performance issues or early failure. Automated optical inspection using computer vision and deep learning provides an efficient approach for quality control. This study investigated a feature-based transfer learning technique for classifying surface defects in images. The resampled KolektorSDD dataset containing 50 defect and 50 non-defect images of electrical commutator surfaces was utilised. Pre-trained DenseNet convolutional neural network (CNN) models were used to extract discriminative features from the images. A support vector machine (SVM) classifier was then trained on these features to classify images as either defective or non-defective. Three DenseNet architectures were tested, namely DenseNet121, DenseNet169, and DenseNet201. All three pipelines achieved 100% accuracy on the test and validation sets when used with the SVM classifier. Owing to the size and the overall performance of the pipeline, it was established that the DenseNet169-SVM was suitable for the given dataset. The high accuracy demonstrates that transfer learning can provide an effective approach for classifying surface defects compared to training a deep neural network from scratch. This allows for the development of automated quality control and surface inspection systems powered by computer vision and deep learning.
AB - Surface defects in industrial products can lead to performance issues or early failure. Automated optical inspection using computer vision and deep learning provides an efficient approach for quality control. This study investigated a feature-based transfer learning technique for classifying surface defects in images. The resampled KolektorSDD dataset containing 50 defect and 50 non-defect images of electrical commutator surfaces was utilised. Pre-trained DenseNet convolutional neural network (CNN) models were used to extract discriminative features from the images. A support vector machine (SVM) classifier was then trained on these features to classify images as either defective or non-defective. Three DenseNet architectures were tested, namely DenseNet121, DenseNet169, and DenseNet201. All three pipelines achieved 100% accuracy on the test and validation sets when used with the SVM classifier. Owing to the size and the overall performance of the pipeline, it was established that the DenseNet169-SVM was suitable for the given dataset. The high accuracy demonstrates that transfer learning can provide an effective approach for classifying surface defects compared to training a deep neural network from scratch. This allows for the development of automated quality control and surface inspection systems powered by computer vision and deep learning.
KW - Deep learning
KW - Feature-based transfer learning
KW - Industrial IoT
KW - Industry 4.0
KW - IoT-enabled defect detection
KW - Machine learning
KW - Manufacturing processes
KW - Smart manufacturing
UR - http://www.scopus.com/inward/record.url?scp=85187783075&partnerID=8YFLogxK
U2 - 10.1007/978-981-99-8498-5_36
DO - 10.1007/978-981-99-8498-5_36
M3 - Conference Proceeding
AN - SCOPUS:85187783075
SN - 9789819984978
T3 - Lecture Notes in Networks and Systems
SP - 443
EP - 449
BT - Advances in Intelligent Manufacturing and Robotics - Selected Articles from ICIMR 2023
A2 - Tan, Andrew
A2 - Zhu, Fan
A2 - Jiang, Haochuan
A2 - Mostafa, Kazi
A2 - Yap, Eng Hwa
A2 - Chen, Leo
A2 - Olule, Lillian J. A.
A2 - Myung, Hyun
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 22 August 2023 through 23 August 2023
ER -