Learning Rotation Sensitive Neural Network for Deformed Objects' Detection in Fisheye Images
Research output: Contributions to collected editions/works › Article in conference proceedings › Research › peer-review
Standard
2019 4th International Conference on Robotics and Automation Engineering (ICRAE 2019): November 22-24, 2019, Singapore . Piscataway: IEEE - Institute of Electrical and Electronics Engineers Inc., 2019. p. 125-129 9043800 (International Conference on Robotics and Automation Engineering, ICRAE ; Vol. 4).
Research output: Contributions to collected editions/works › Article in conference proceedings › Research › peer-review
Harvard
APA
Vancouver
Bibtex
}
RIS
TY - CHAP
T1 - Learning Rotation Sensitive Neural Network for Deformed Objects' Detection in Fisheye Images
AU - Chen, Zhen
AU - Georgiadis, Anthimos
N1 - Conference code: 4
PY - 2019/11
Y1 - 2019/11
N2 - Object detection plays a significant role in an intelligent system equipped with a fisheye camera. The fisheye image captures a wide field-of-view but deforms in the radial direction. The deformation changes the relative angle between the edge of objects and the image. Therefore, a horizontal bounding box cannot perform an accurate description of an object's location and dimension in advanced neural network training. In this paper, we build a rotation sensitive neural network targeting to realize one-stage regression on the fisheye image detection. The oriented bounding box is applied in the object's description and detection. To evaluate our proposed method, we develop a new labelled fisheye image dataset that contains two categories. The network model training takes around 3 hours and achieves 100% precious by the test set.
AB - Object detection plays a significant role in an intelligent system equipped with a fisheye camera. The fisheye image captures a wide field-of-view but deforms in the radial direction. The deformation changes the relative angle between the edge of objects and the image. Therefore, a horizontal bounding box cannot perform an accurate description of an object's location and dimension in advanced neural network training. In this paper, we build a rotation sensitive neural network targeting to realize one-stage regression on the fisheye image detection. The oriented bounding box is applied in the object's description and detection. To evaluate our proposed method, we develop a new labelled fisheye image dataset that contains two categories. The network model training takes around 3 hours and achieves 100% precious by the test set.
KW - Fisheye image
KW - Image distortion
KW - Orientated Box
KW - YOLO detector
KW - Engineering
UR - http://www.scopus.com/inward/record.url?scp=85083237395&partnerID=8YFLogxK
U2 - 10.1109/ICRAE48301.2019.9043800
DO - 10.1109/ICRAE48301.2019.9043800
M3 - Article in conference proceedings
AN - SCOPUS:85083237395
SN - 978-1-7281-4741-3
T3 - International Conference on Robotics and Automation Engineering, ICRAE
SP - 125
EP - 129
BT - 2019 4th International Conference on Robotics and Automation Engineering (ICRAE 2019)
PB - IEEE - Institute of Electrical and Electronics Engineers Inc.
CY - Piscataway
T2 - 4th International Conference on Robotics and Automation Engineering, ICRAE 2019
Y2 - 22 November 2019 through 24 November 2019
ER -