The Journal of The Korea Institute of Intelligent Transport Systems (한국ITS학회 논문지)

The Korea Institute of Intelligent Transport Systems (한국ITS학회)
권호 표지
DOI QR코드

DOI QR Code

A Selection Method of Backbone Network through Multi-Classification Deep Neural Network Evaluation of Road Surface Damage Images

도로 노면 파손 영상의 다중 분류 심층 신경망 평가를 통한 Backbone Network 선정 기법

  • Received : 2019.05.06
  • Accepted : 2019.06.25
  • Published : 2019.06.30
Download PDF
⟨ Previous Next ⟩

Abstract

In recent years, research and development on image object recognition using artificial intelligence have been actively carried out, and it is expected to be used for road maintenance. Among them, artificial intelligence models for object detection of road surface are continuously introduced. In order to develop such object recognition algorithms, a backbone network that extracts feature maps is essential. In this paper, we will discuss how to select the appropriate neural network. To accomplish it, we compared with 4 different deep neural networks using 6,000 road surface damage images. Based on three evaluation methods for analyzing characteristics of neural networks, we propose a method to determine optimal neural networks. In addition, we improved the performance through optimal tuning of hyper-parameters, and finally developed a light backbone network that can achieve 85.9% accuracy of road surface damage classification.

최근 들어 인공 지능을 이용한 영상 객체 인식에 대한 연구 및 개발이 활발하게 진행되고 있다. 그 연장선상에서 도로 유지 및 관리 분야에도 관련 연구의 활용도가 크게 향상될 것으로 기대된다. 그 중에서도 특히 도로 노면 파손 객체 인식 (Object Detection) 을 위한 인공 지능모델이 지속적으로 개발되고 있다. 이러한 객체 인식 알고리즘을 개발하려면 우선적으로 특징지도를 생성하는 Backbone Network가 반드시 필요한데, 본 논문에서는 이를 선정하는 방법을 제안하고자 한다. 이를 위해 6,000여 장의 도로 노면 파손 영상 데이터를 확보하고, 근래에 많이 사용되는 4종류의 심층 신경망을 활용하여 성능을 비교한다. 3가지의 성능 평가 방법을 적용하여 심층 신경망의 특징을 분석하고 최적의 심층 신경망을 결정한다. 또한 하이퍼 파라미터의 최적 조율을 통해 성능을 향상시키고, 최종적으로 도로 노면 파손 영상 분류를 위하여 85.9%의 정확도로 수행이 가능한 경량화된 Backbone Network용 심층 신경망을 제안한다.

Keywords

References

  1. Badrinarayanan V., Kendall A. and Cipolla R.(2017), "SegNet: A deep convolutional encoder-decoder architecture for image segmentation," The IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 39, no. 12, pp.2481-2495. https://doi.org/10.1109/TPAMI.2016.2644615
  2. Chun C., Shim S., Kang S. and Ryu S.(2018), "Development and evaluation of automatic pothole detection using fully convolutional neural networks," Journal of Korea Institute of Intelligent Transport System, vol. 17, no. 5, pp.55-64.
  3. Dozat T.(2016), Incorporating nesterov momentum into adam.
  4. Eisenbach M., Stricker R., Seichter D., Amende K., Debes K., Sesselmann M., Ebersbach D., Stoeckert U. and Gross H.(2017), "How to get pavement distress detection ready for deep learning? A systematic approach," in Proc. International Joint Conference on Neural Networks (IJCNN), Anchorage, Alaska, pp.2039-2047.
  5. Girshick R.(2015), "Fast r-cnn," in Proc. The IEEE International Conference on Computer Vision (ICCV), Santiago, Chile, pp.1440-1448.
  6. Glorot X. and Bengio Y.(2010), "Understanding the difficulty of training deep feedforward neural networks," in Proc. the Thirteenth International Conference on Artificial Intelligence and Statistics, Sardinia, Italy, pp.249-256.
  7. He K., Gkioxari G., Dollar P. and Girshick R.(2017), "Mask R-CNN," in Proc. The IEEE International Conference on Computer Vision (ICCV), Venezia, Italy, pp.2980-2988.
  8. He K., Zhang X., Ren S. and Sun J.(2016), "Deep residual learning for image recognition," in Proc. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, Nevada, pp.770-778.
  9. Huang G., Liu Z., Van Der Maaten L. and Weinberger K. Q.(2017), "Densely connected convolutional networks," in Proc. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Hawaii, Honolulu, pp.4700-4708.
  10. Jo Y. and Ryu S.(2015), "Pothole detection system using black-box camera," Sensors, vol. 15, no. 11, pp.29316-29331. https://doi.org/10.3390/s151129316
  11. Jo Y. and Ryu S.(2016), "Pothole detection algorithm based on saliency map for improving detection performance," Journal of Korea Institute of Intelligent Transport System, vol. 15, no. 4, pp.104-114. https://doi.org/10.12815/kits.2016.15.4.104
  12. Kim S. J., An K. E., Lee S. W., Ryu S. and Seo D.(2018), "A study of image processing and deep learning algorithm for efficient of mobile-based pothole detection system," in Proc. Korea Software Congress 2018 (KSC), Pyeongchang, Korea, pp.1595-1597.
  13. Kingma D. P. and Ba J. L.(2015), "ADAM: a method for stochastic optimization," in Proc. third International Conference on Learning Representations (ICLR), San Diego, California, pp.1-15.
  14. Krizhevsky A. and Hinton G.(2009), "Learning multiple layers of features from tiny images," Technical report, University of Toronto, vol. 1, no. 4, pp.7-66.
  15. Lee S. W., An K. E., Jeong Y. J., Kim S. J., Ryu S. and Seo D.(2018), "A proposal of a real-time pothole detection system based on machine learning and computer vision," in Proc. Korea Computer Congress 2018 (KCC), Jeju, Korea, pp.7-9.
  16. Li Z., Peng C., Yu G., Zhang X., Deng Y. and Sun J.(2018), "Detnet: A backbone network for object detection," arXiv preprint arXiv:1804.06215.
  17. Long J., Shelhamer E. and Darrell T.(2015), "Fully convolutional networks for semantic segmentation," in Proc. IEEE Conference on Computer Vision and Pattern Recognition, Boston, USA, pp.3431-3440.
  18. Maeda H., Sekimoto Y., Seto T., Kashiyama T. and Omata H.(2018), "Road damage detection and classification using deep neural networks with smartphone images," Computer-Aided Civil Infrastructure Engineering, vol. 33, no. 12, pp.1127-1141. https://doi.org/10.1111/mice.12387
  19. MOLIT (Ministry of Land, Infrastructure and Transport). (Oct., 31, 2018). Smart Construction Technology Road Map. Available From: http://www.molit.go.kr/USR/NEWS/m_71/dtl.jsp?id=95081506 (accessed Feb., 8, 2019).
  20. Netzer Y., Wang T., Coates A., Bissacco A., Wu B. and Ng A. Y.(2011), "Reading digits in natural images with unsupervised feature learning," In NIPS workshop.
  21. Ren S., He K., Girshick R. and Sun J.(2015), "Faster R-CNN: towards real-time object detection with region proposal networks," in Proc. Advances in Neural Information Processing Systems 28 (NIPS 2015), Montreal, Canada, pp.91-99.
  22. Simonyan K. and Zisserman A.(2014), "Very deep convolutional networks for large-scale image recognition," arXiv preprint arXiv:1409.1556.
  23. Yang X., Li H., Yu Y., Luo X., Huang X. and Yang X.(2018), "Automatic pixel-level crack detection and measurement using fully convolutional networks," Computer-Aided Civil and Infrastructure Engineering, vol. 33, no. 12, pp.1090-1109. https://doi.org/10.1111/mice.12412
  24. Yu F. and Koltun V.(2015), "Multi-scale context aggregation by dilated convolution," arXiv preprint arXiv:1511.07122.
  25. Zhang A., Wang K. C., Li B., Yang E., Dai X., Peng Y., Fei Y., Liu Y., Li J. Q. and Chen C.(2017), "Automated pixel-level pavement crack detection on 3D asphalt surfaces using a deep-learning network," Computer-Aided Civil and Infrastructure Engineering, vol. 32, no. 10, pp.805-819. https://doi.org/10.1111/mice.12297