Journal of the Korea Society of Computer and Information (한국컴퓨터정보학회논문지)

Korean Society of Computer Information (한국컴퓨터정보학회)
권호 표지
DOI QR코드

DOI QR Code

Improved CNN Algorithm for Object Detection in Large Images

  • Yang, Seong Bong (Dept. of Computer Science and Engineering, Korea National Defense University) ;
  • Lee, Soo Jin (Dept. of Computer Science and Engineering, Korea National Defense University)
  • Received : 2019.10.22
  • Accepted : 2019.12.11
  • Published : 2020.01.31
Download PDF
⟨ Previous Next ⟩

Abstract

Conventional Convolutional Neural Network(CNN) algorithms have limitations in detecting small objects in large image. In this paper, we propose an improved model which is based on Region Of Interest(ROI) selection and image dividing technique. We prepared YOLOv3 / Faster R-CNN algorithms which are transfer-learned by airfield and aircraft datasets. Also we prepared large images for testing. In order to verify our model, we selected airfield area from large image as ROI first and divided it in two power n orders. Then we compared the aircraft detection rates by number of divisions. We could get the best size of divided image pieces for efficient small object detection derived from the comparison of aircraft detection rates. As a result, we could verify that the improved CNN algorithm can detect small object in large images.

기존의 CNN 알고리즘은 위성영상과 같은 대형 이미지에서 소형 객체를 식별하는 것이 불가능하다는 문제점을 가지고 있었다. 본 연구에서는 이러한 문제를 해결하기 위해 관심영역 설정 및 이미지 분할 기법을 적용한 CNN 알고리즘 개선방안을 제시하였다. 실험은 비행장 및 항공기 데이터셋으로 전환학습한 YOLOv3 / Faster R-CNN 알고리즘과 테스트용 대형 이미지를 이용하여 진행하였으며, 우선 대형 이미지에서 관심영역을 식별하고 이를 순차적으로 분할해 나가며 CNN 알고리즘의 객체식별 결과를 비교하였다. 분할 이미지의 크기는 실험을 통해 최소 분할로 최대의 식별률을 얻을 수 있는 최적의 이미지 조각 크기를 도출하여 적용하였다. 실험 결과, 본 연구에서 제시한 방안을 통해 CNN 알고리즘으로 대형 이미지에서의 소형 객체를 식별하는 것이 충분히 가능함을 검증하였다.

Keywords

References

  1. J. Yu, et al., "Improving Performance of YOLO Network Using Multi-layer Overlapped Windows for Detecting Correct Position of Small Dense Objects," Journal of The Korea Society of Computer and Information Vol. 24 No. 3, pp. 19-27, March 2019. DOI: 10.9708/JKSCI.2019.24.03.019.
  2. M. Kisantal, et al,. "Augmentation for small object detection," Computer Vision and Pattern Recognition (CVPR) February 2019. arXiv:1902.07296v1
  3. G. Cao, et al,. "Feature-fused SSD: fast detection for small objects," 9th International Conference on Graphic and Image Processing (ICGIP), April. 2018. DOI: 10.1117/12.2304811
  4. Y. Ren, et al,. "Small Object Detection in Optical Remote Sensing Images via Modified Faster R-CNN," Multidisciplinary Digital Publishing Institute (MDPI), May 2018. DOI: 10.3390/app8050813
  5. M. Peng, C. Wang, T. Chen, G. Liu,. "NIRFaceNet: A Convolutional Neural Network for Near-Infrared Face Identification," Information, Vol. 7, No. 4, pp. 61, October. 2016. DOI: 10.3390/info7040061
  6. S. Ren, K. He, R. Girshick, and J. Sun, "Faster R-CNN: Towards real-time object detection with region proposal networks," IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 15, Issue. 6, pp. 1137-1149, June. 2017. DOI: 10.1109/TPAMI.2016.2577031
  7. J. Dai, Y. Li, K. He, and J. Sun, "R-FCN: Object Detection via Region-based Fully Convolutional Networks," Neural Information Processing Systems(NIPS), 29, May. 2016. arXiv:1605.06409
  8. W. Liu, et al., "SSD: Single Shot MultiBox Detector." European Conference on Computer Vision(ECCV) 2016, pp. 21-37, September. 2016. DOI: 10.1007/978-3-319-46448-0_2
  9. J. Redmon, et. al., "You Only Look Once:Unified, Real-Time Object Detection," The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 779-788. June. 2016. DOI: 10.1109/CVPR.2016.91
  10. I. Goodfellow, et al., "Generative Adversarial Networks" Neural Information Processing Systems(NIPS), June. 2014. arXiv:1406.2661v1
  11. A. Mikolajczyk, and M. Grochowsk., "Data augmentation for improving deep learning in image classification problem," Conference: 2018 International Interdisciplinary PhD Workshop (IIPhDW), pp. 117-122. May. 2018. DOI: 10.1109/IIPHDW.2018.8388338
  12. H. Shin, et. al., "Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning," The IEEE Transactions on Medical Imaging, Vol. 35, No. 5, pp. 1285-1298. May. 2016. DOI: 10.1109/TMI.2016.2528162
  13. Digitalglobe. com., "Worldview-3 Data sheet," 2017, https://dgv4-cms-production.s3.amazonaws.com/uploads/document/file/128/DG2017_WorldView-3_DS.pdf
  14. J. Redmon, and A. Farhadi, "YOLOv3: An Incremental Improvement," Computer Vision and Pattern Recognition (cs.CV), April. 2018. arXiv:1804.02767
  15. Tzutalin., "LabelImg," Git code, https://github.com/tzutalin/labelImg

Cited by

  1. Concept Drift Based on CNN Probability Vector in Data Stream Environment vol.13, pp.4, 2020, https://doi.org/10.13160/ricns.2020.13.4.147
  2. 기계학습 기법에 따른 KOMPSAT-3A 시가화 영상 분류 - 서울시 양재 지역을 중심으로 - vol.36, pp.6, 2020, https://doi.org/10.7780/kjrs.2020.36.6.2.7