TY - GEN
T1 - Differentiable Neural Architecture Search for SAR Image Ship Object Detection
AU - Li, Zhiheng
AU - Chen, Liang
AU - Huang, Xiaoqi
AU - Zhang, Zhixin
AU - Shi, Hao
N1 - Publisher Copyright:
© 2020 IET Conference Proceedings. All rights reserved.
PY - 2020
Y1 - 2020
N2 - SAR image ship detection has important applications in military and civil fields. With the wide application of deep learning in the field of computer vision, more and more SAR ship detection algorithms have applied deep learning. Neural network based on deep learning can automatically extract features instead of hand-crafted features, and has achieved great success. But the design of deep neural networks is time-consuming and complex, which need a lot of professional knowledge and experience. Neural architecture search (NAS) has shown great potential in automated design networks. NAS can find the most suitable architecture on the specified dataset within a given search space. In object detection, the scale of the object varies, which is usually solved by designing feature fusion network. So in this work, we proposed to use differentiable architecture search to search the feature fusion module for SAR ship detection. The experiments on SAR ship detection dataset (SSDD) showed that the discovered architecture can replace the corresponding part of the detection network with 1% performance improvement.
AB - SAR image ship detection has important applications in military and civil fields. With the wide application of deep learning in the field of computer vision, more and more SAR ship detection algorithms have applied deep learning. Neural network based on deep learning can automatically extract features instead of hand-crafted features, and has achieved great success. But the design of deep neural networks is time-consuming and complex, which need a lot of professional knowledge and experience. Neural architecture search (NAS) has shown great potential in automated design networks. NAS can find the most suitable architecture on the specified dataset within a given search space. In object detection, the scale of the object varies, which is usually solved by designing feature fusion network. So in this work, we proposed to use differentiable architecture search to search the feature fusion module for SAR ship detection. The experiments on SAR ship detection dataset (SSDD) showed that the discovered architecture can replace the corresponding part of the detection network with 1% performance improvement.
KW - FEATURE FUSION
KW - NEURAL ARCHITECTURE SEARCH
KW - SAR
KW - SHIP DETECTION
UR - http://www.scopus.com/inward/record.url?scp=85138224047&partnerID=8YFLogxK
U2 - 10.1049/icp.2021.0794
DO - 10.1049/icp.2021.0794
M3 - Conference contribution
AN - SCOPUS:85138224047
VL - 2020
SP - 950
EP - 954
BT - IET Conference Proceedings
PB - Institution of Engineering and Technology
T2 - 5th IET International Radar Conference, IET IRC 2020
Y2 - 4 November 2020 through 6 November 2020
ER -