Journal of Drive and Control (드라이브 ㆍ 컨트롤)

The Korean Society for Fluid Power and Construction Equipment (유공압건설기계학회)
권호 표지
DOI QR코드

DOI QR Code

Validation of Semantic Segmentation Dataset for Autonomous Driving

승용자율주행을 위한 의미론적 분할 데이터셋 유효성 검증

  • Gwak, Seoku (Department of Mechanical Engineering, Sungkyunkwan University) ;
  • Na, Hoyong (Department of Mechanical Engineering, Sungkyunkwan University) ;
  • Kim, Kyeong Su (Department of Mechanical Engineering, Sungkyunkwan University) ;
  • Song, EunJi (Department of Mechanical Engineering, Sungkyunkwan University) ;
  • Jeong, Seyoung (Department of Mechanical Engineering, Sungkyunkwan University) ;
  • Lee, Kyewon (INFINIQ) ;
  • Jeong, Jihyun (INFINIQ) ;
  • Hwang, Sung-Ho (Department of Mechanical Engineering, Sungkyunkwan University)
  • Received : 2022.11.07
  • Accepted : 2022.11.29
  • Published : 2022.12.01
Download PDF
⟨ Previous Next ⟩

Abstract

For autonomous driving research using AI, datasets collected from road environments play an important role. In other countries, various datasets such as CityScapes, A2D2, and BDD have already been released, but datasets suitable for the domestic road environment still need to be provided. This paper analyzed and verified the dataset reflecting the Korean driving environment. In order to verify the training dataset, the class imbalance was confirmed by comparing the number of pixels and instances of the dataset. A similar A2D2 dataset was trained with the same deep learning model, ConvNeXt, to compare and verify the constructed dataset. IoU was compared for the same class between two datasets with ConvNeXt and mIoU was compared. In this paper, it was confirmed that the collected dataset reflecting the driving environment of Korea is suitable for learning.

Keywords

Acknowledgement

본 연구는 국토교통부/국토교통과학기술진흥원 교통물류연구사업의 연구비지원 (22TLRP-C152478-04)과 과학기술정보통신부 및 한국지능정보사회진흥원 인공지능 학습용 데이터 구축 사업의 연구결과로 수행된 결과물입니다.

References

  1. Tan, M. and Le, Q. (2019). Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, p.6105-6114. PMLR.
  2. Elmenreich, W. (2002). An introduction to sensor fusion. Vienna University of Technology, Austria, 502, p.1-28
  3. Song, T., Jang, H., Ha, N., Yeon, Y., Kwon, K., and Sohn, K. (2019). Deep Multi-task Network for Simultaneous Hazy Image Semantic Segmentation and Dehazing. Journal of Korea Multimedia Society, 22(9), 1000-1010. https://doi.org/10.9717/KMMS.2019.22.9.1000
  4. Minaee, S., Boykov, Y., Porikli, F., Plaza, A., Kehtarnavaz, N., and Terzopoulos, D. (2022). Image segmentation using deep learning: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(7), p.3523-3542
  5. H. S. Baek, J. H. Shin and S. J. Kim, "Development of AI-Based Condition Monitoring System for Failure Diagnosis of Excavator's Travel Device", Journal of Drive and Control, Vol.18, No.1, pp.24-30, 2021. https://doi.org/10.7839/KSFC.2021.18.1.024
  6. Lee, J., Gwak, G., Kim, K., Kang, W., Shin, D., and Hwang, S. (2021). Development of virtual simulator and database for deep learning-based object detection. Journal of Drive and Control, 18(4), p.9-18. 2022. https://doi.org/10.7839/KSFC.2021.18.4.009
  7. Kim, K., Lee, J., Gwak, S., Kang, W., Shin, D., and Hwang, S. (2022). Construction of database for deep learning-based occlusion area detection in the virtual environment. Journal of Drive and Control, 19(3), p.9-15. 2022. https://doi.org/10.7839/KSFC.2022.19.3.009
  8. Weng, X., Man, Y., Cheng, D., Park, J., O'toole, M., and Kitani, K. (2020). All-in-one drive: A large-scale comprehensive perception dataset with high-density long-range point clouds.
  9. Toldo, M., Maracani, A., Michieli, U., and Zanuttigh, P. (2020). Unsupervised domain adaptation in semantic segmentation: a review. Technologies, 8(2), p.35. https://doi.org/10.3390/technologies8020035
  10. Chen, Y., Liu, L., Tao, J., Chen, X., Xia, R., Zhang, Q., Xiong, J., Yang, K., and Xie, J. (2021). The image annotation algorithm using convolutional features from intermediate layer of deep learning. Multimedia Tools and Applications, 80(3), p.4237-4261. https://doi.org/10.1007/s11042-020-09887-2
  11. Contributors, M. (2020). MMSegmentation: Openmmlab semantic segmentation toolbox and benchmark. https://github.com/open-mmlab/mmsegmentation.
  12. Cordts, M., Omran, M., Ramos, S., Rehfeld, T., Enzweiler, M., Benenson, R., Franke, U., Roth, S., and Schiele, B. (2016). The cityscapes dataset for semantic urban scene understanding. In Proceedings of the IEEE conference on computer vision and pattern recognition, p.3213-3223.
  13. Geyer, J., Kassahun, Y., Mahmudi, M., Ricou, X., Durgesh, R., Chung, A., Hauswald, L., Pham, H., et al. (2020). A2d2: Audi autonomous driving dataset. arXiv preprint arXiv:2004.06320.