TY - GEN
T1 - Deep Learning Based Concealed Object Recognition in Active Millimeter Wave Imaging
AU - Myint, San Hlaing
AU - Katsuyama, Yutaka
AU - Sato, Toshio
AU - Qi, Xin
AU - Tamesue, Kazuhiko
AU - Wen, Zheng
AU - Yu, Keping
AU - Tokuda, Kiyohito
AU - Sato, Takuro
N1 - Funding Information:
V. CONCLUSIONS This paper focused on concealed object recognition based on deep learning with an active millimeter wave imaging system. This work attempted to address the lack of an active millimeter wave image dataset because of the high cost of real active millimeter wave imaging devices. Given the lack of comprehensive active millimeter wave image datasets, training dataset generation methods were proposed based on different active millimeter wave imaging simulation models. We compared two approaches to validate the accuracy of the CNN. Experiments were performed iteratively and evaluated at each epoch. The results indicated that the distance-based active millimeter wave image dataset outperformed the depth-based active millimeter wave image dataset by achieving high accuracy, 0.79 and 0.82 at each epoch 20 and 30 respectively, and a correct classification rate. In the near future, we will consider a hybrid millimeter wave imaging system and try to improve the recognition accuracy of deep learning methods by applying a comprehensive hybrid millimeter wave image dataset ACKNOWLEDGMENTS This work was supported by a research grant for expanding radio wave resources (JPJ000254) from the Ministry of Internal Affairs and Communications under the contract for “Research and development of radar fundamental technology for advanced recognition of moving objects for security enhancement”.
Publisher Copyright:
© 2021 IEEE.
PY - 2021
Y1 - 2021
N2 - In application related to public security check system, passive and active imaging of millimeter wave still faces critical challenges in providing high resolution quality images. Improving the detection, localization, and recognition accuracy of concealed object detection systems is very challenging due to the lack of a dataset of millimeter wave images with good resolution. Although previous studies proposed artificial intelligence-based concealed object recognition systems, improving accuracy remains a critical challenge. Therefore, in this paper, we propose two kinds of training dataset generation methods based on the proposed active millimeter wave imaging (AMWI) approaches presented in our previous work to improve the accuracy of convolutional neural networks (CNN)-based concealed object recognition systems. First, a depth-based training dataset generation method and a distance-based training dataset generation method are proposed for specular images and nonspecular images. Finally, a CNN-based concealed object recognition system is proposed by using generated active millimeter wave images and interferometer active images to improve the recognition accuracy.
AB - In application related to public security check system, passive and active imaging of millimeter wave still faces critical challenges in providing high resolution quality images. Improving the detection, localization, and recognition accuracy of concealed object detection systems is very challenging due to the lack of a dataset of millimeter wave images with good resolution. Although previous studies proposed artificial intelligence-based concealed object recognition systems, improving accuracy remains a critical challenge. Therefore, in this paper, we propose two kinds of training dataset generation methods based on the proposed active millimeter wave imaging (AMWI) approaches presented in our previous work to improve the accuracy of convolutional neural networks (CNN)-based concealed object recognition systems. First, a depth-based training dataset generation method and a distance-based training dataset generation method are proposed for specular images and nonspecular images. Finally, a CNN-based concealed object recognition system is proposed by using generated active millimeter wave images and interferometer active images to improve the recognition accuracy.
KW - Active Imaging
KW - Deep Learning
KW - Millimeter Wave
UR - http://www.scopus.com/inward/record.url?scp=85124423806&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85124423806&partnerID=8YFLogxK
U2 - 10.1109/APMC52720.2021.9662033
DO - 10.1109/APMC52720.2021.9662033
M3 - Conference contribution
AN - SCOPUS:85124423806
T3 - Asia-Pacific Microwave Conference Proceedings, APMC
SP - 434
EP - 436
BT - 2021 IEEE Asia-Pacific Microwave Conference, APMC 2021
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2021 IEEE Asia-Pacific Microwave Conference, APMC 2021
Y2 - 28 November 2021 through 1 December 2021
ER -