Browse > Article
http://dx.doi.org/10.7780/kjrs.2022.38.6.4.6

Comparison of Semantic Segmentation Performance of U-Net according to the Ratio of Small Objects for Nuclear Activity Monitoring  

Lee, Jinmin (Department of Civil Engineering, Seoul National University of Science and Technology)
Kim, Taeheon (Department of Civil Engineering, Seoul National University of Science and Technology)
Lee, Changhui (Department of Civil Engineering, Seoul National University of Science and Technology)
Lee, Hyunjin (Department of Civil Engineering, Seoul National University of Science and Technology)
Song, Ahram (Department of Location-Based Information System, Kyungpook National University)
Han, Youkyung (Department of Civil Engineering, Seoul National University of Science and Technology)
Publication Information
Korean Journal of Remote Sensing / v.38, no.6_4, 2022 , pp. 1925-1934 More about this Journal
Abstract
Monitoring nuclear activity for inaccessible areas using remote sensing technology is essential for nuclear non-proliferation. In recent years, deep learning has been actively used to detect nuclear-activity-related small objects. However, high-resolution satellite imagery containing small objects can result in class imbalance. As a result, there is a performance degradation problem in detecting small objects. Therefore, this study aims to improve detection accuracy by analyzing the effect of the ratio of small objects related to nuclear activity in the input data for the performance of the deep learning model. To this end, six case datasets with different ratios of small object pixels were generated and a U-Net model was trained for each case. Following that, each trained model was evaluated quantitatively and qualitatively using a test dataset containing various types of small object classes. The results of this study confirm that when the ratio of object pixels in the input image is adjusted, small objects related to nuclear activity can be detected efficiently. This study suggests that the performance of deep learning can be improved by adjusting the object pixel ratio of input data in the training dataset.
Keywords
Nuclear non-proliferation; Semantic segmentation; U-Net; Small object; Class imbalance;
Citations & Related Records
Times Cited By KSCI : 3  (Citation Analysis)
연도 인용수 순위
1 Lee, J.W., J.Y. Kim, J.K. Kim, and C.H. Kwon, 2020. Tiny object recognition in Aerial imagery Using Convolution Neural Network, Proc. of the 2020 Institute of Electronics and Information Engineers, Jeju, Korea, Aug. 19-21, pp. 1990-1991 (in Korean with English abstract).
2 Pabian, F., O. Heinonen, J. Liu, and S. J. Pitz, 2021. Yongbyon Nuclear Research Center: Construction Activity Near ELWR, https://www.38north.org/2021/09/yongbyon-nuclear-research-centerconstruction-activity-near-elwr/, Accessed on Dec. 10, 2022.
3 Seong, S., H. Choi, J. Mo, and J. Choi, 2021. Availability Evaluation of Object Detection Based on Deep Learning Method by Using Multitemporal and Multisensor Data for Nuclear Activity Analysis, Korean Journal of Remote Sensing, 37(5-1): 1083-1094 (in Korean with English abstract). https://doi.org/10.7780/kjrs.2021.37.5.1.20   DOI
4 Japkowicz, N., 2000. The class imbalance problem: Significance and strategies, Proc. of the 2000 International Conference on Artificial Intelligence, Las Vegas, NV, USA, Jun. 28-Jul. 1, vol. 56, pp. 111-117.
5 Johnson, J.M. and T.M. Khoshgoftaar, 2019. Survey on Deep Learning with Class Imbalance, Journal of Big Data, 6(1): 1-54. https://doi.org/10.1186/s40537-019-0192-5   DOI
6 Lee, W.J., J. Sun, H.S. Jung, S.C. Park, D.K. Lee, and K.Y. Oh, 2018. Detection of Surface Changes by the 6th North Korea Nuclear Test Using High-resolution Satellite Imagery, Korean Journal of Remote Sensing, 34(6-4): 1479-1488 (in Korean with English abstract). http://dx.doi.org/10.7780/kjrs.2018.34.6.4.2   DOI
7 Pabian, F., J.S. Bermudez Jr., and J. Liu, 2018. North Korea's Yongbyon Nuclear Center: Key Activities in a 5-Mwe Nuclear Power Plant, https://www.38north.org/2018/04/yongbyon040418k/, Accessed on Dec. 10, 2022.
8 Park, I., J. An, J.H. Lee, H.K. Lee, J.H. Park, M.S. Kim, J.S. Sin, Y.M. Choi, and H.C. Jang, 2013. Technological Information Analysis on the Nuclear Activities of Surrounding Countries, 2012-M5A1A 1026309, Korea Institute of Nuclear Nonproliferation and Control, Daejeon, Korea (in Korean with English abstract).
9 Son, M.J., S.W. Jung, and E.J. Hwang, 2019. A Deep Learning Based Over-Sampling Scheme for Imbalanced Data Classification, KIPS Transactions on Software and Data Engineering, 8(7): 311-316 (in Korean with English abstract). https://doi.org/10.3745/KTSDE.2019.8.7.311   DOI
10 Ronneberger, O., P. Fischer, and T. Brox, 2015. U-net: Convolutional networks for biomedical image segmentation, U-Net: Convolutional Networks for Biomedical Image Segmentation. In: Navab, N., Hornegger, J., Wells, W., Frangi, A. (eds), Medical Image Computing and Computer-Assisted Intervention - MICCAI 2015, Springer, Cham, Switzerland, vol. 9351, pp. 234-241. https://doi.org/10.1007/978-3-319-24574-4_28   DOI