Improving Performance of YOLO Network Using Multi-layer Overlapped Windows for Detecting Correct Position of Small Dense Objects (밀집한 소형 물체의 정확한 위치 검출을 위한 다계층 중첩 윈도우를 이용한 YOLO 네트워크의 성능개선)

Jae-Hyoung Yu (유재형); Youngjun Han (한영준); Hernsoo Hahn (한헌수)

doi:10.9708/jksci.2019.24.03.019

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
Abbr : JKSCI
2019, 24(3), pp.19-27
DOI : 10.9708/jksci.2019.24.03.019
Publisher : The Korean Society Of Computer And Information
Research Area : Engineering > Computer Science
Received : January 25, 2019
Accepted : February 22, 2019
Published : March 29, 2019

Jae-Hyoung Yu ¹, Youngjun Han ¹, Hernsoo Hahn ¹

¹숭실대학교

Accredited

ABSTRACT

This paper proposes a new method using multi-layer overlapped windows to improve the performance of YOLO network which is vulnerable to detect small dense objects. In particular, the proposed method uses the YOLO Network based on the multi-layer overlapped windows to track small dense vehicles that approach from long distances. The method improves the detection performance for location and size of small vehicles. It allows crossing area of two multi-layer overlapped windows to track moving vehicles from a long distance to a short distance. And the YOLO network is optimized so that GPU computation time due to multi-layer overlapped windows should be reduced. The superiority of the proposed algorithm has been proved through various experiments using captured images from road surveillance cameras.

KEYWORDS

Multi-layer Overlapped Window, YOLO network, Small Dense Objects, Crossing Area, Small Vehicle Tracking

Citation status

* References for papers published after 2023 are currently being built.

[book] S. Russell / 2009 / rtificial Intelligence : A Modern Approach / Prentice Hall Press

[journal] O. Russakovsky / 2015 / ImageNet large scale visual recognition challenge / Int. J. Comput. Vis 115(3) : 211~252

[confproc] S. Ioffe / 2015 / Batch normalization :Accelerating deep network training by reducing internal covariate shift / Proc. ICML

[other] T. Young / 2017 / Recent Trends in Deep Learning Based Natural Language Processing / arXiv preprint arXiv : 1708. 02709

[confproc] K. He / 2017 / Mask R-CNN / ICCV / IEEE : 2980~2988

[confproc] O. Ronneberger / 2015 / Unet : Convolutional Networks for Biomedical Image Segmentation / International Conference on Medical image computing and computer-assisted intervention : 234~241

[other] V. Badrinarayanan / 2015 / SegNet : A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation / arXiv preprint arXiv : 1511. 00561

[journal] Y. Lecun / 1998 / Gradientbased learning applied to document recognition / Proceedings of the IEEE 86(11) : 2278~2324

[confproc] K. Simonyan / 2015 / Very deep convolutional networks for large-scale image recognition / Proc. ICLR

[confproc] A. Krizhevsky / 2012 / Imagenet classification with deep convolutional neural networks / Advances in neural information processing systems : 1097~1105

[confproc] C. Szegedy / 2015 / Going deeper with convolutions / Proc. CVPR

[confproc] R. Girshick / 2014 / Rich feature hierarchies for accurate object detection and semantic segmentation / Proceedings of the IEEE conference on computer vision and pattern recognition : 580~587

[confproc] R. Girshick / 2015 / Fast r-cnn / Proceedings of the IEEE International Conference on Computer Vision : 1440~1448

[confproc] W. Liu / 2016 / SSD : Single shot multibox detector / Proc. ECCV : 21~37

[confproc] J. Redmon / 2016 / You only look once : Unified, real-time object detection / Proc. CVPR

[confproc] J. Redmon / 2016 / YOLO9000 : Better, Faster, Stronger / Proceedings of the IEEE conference on computer vision and pattern recognition : 779~788

[other] J. Redmon / 2018 / YOLOv3 : An Incremental Improvement / arXiv preprint arXiv : 1804. 02767

This paper was written with support from the National Research Foundation of Korea.

KJCKorea
Journal Central

Journal of The Korea Society of Computer and Information 2023 KCI Impact Factor : 0.65

Improving Performance of YOLO Network Using Multi-layer Overlapped Windows for Detecting Correct Position of Small Dense Objects

ABSTRACT

KEYWORDS

Citation status

* References for papers published after 2023 are currently being built.

Journal of The Korea Society of Computer and Information 2023 KCI Impact Factor : 0.65

Improving Performance of YOLO Network Using Multi-layer Overlapped Windows for Detecting Correct Position of Small Dense Objects

ABSTRACT

KEYWORDS

Statistics

Tools

Issue List

Citation status

KCI Citation Counts (2)

REFERENCES (17) * References for papers published after 2023 are currently being built.

Search PDF

Citation

* References for papers published after 2023 are currently being built.