Journal
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
Volume 125, Issue -, Pages -Publisher
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.engappai.2023.106696
Keywords
Warship re-identification; Transfer learning; Domain adaptation; CNNs; Complex sea conditions
Ask authors/readers for more resources
This paper proposes a dynamic alignment re-identification network model with transfer learning to improve the recognition accuracy of warships by using various types of ships as source domain data. The simulated swaying situation of warships at sea is tested to cope with the impact caused by complex sea conditions. The experimental results show that the improved algorithm increases the mean average accuracy (mAP) by 10.2% and the first hit rate (Rank1) by 4.9% on average.
Marine vessel re-identification technology is an important component of intelligent shipping systems and an important part of the visual perception tasks required for marine surveillance. However, unlike the situation on land, the maritime environment is complex and volatile, and ships are prone to different degrees of swaying. Warships, as a class of ships, are characterized by fewer image samples and high similarity, and it is more difficult to re-identify warships at sea. Therefore, this paper proposes a dynamic alignment re-identification network model incorporating transfer learning methods. Various types of ships are used as source domain data to assist the network in learning the target domain warships so as to improve the recognition accuracy of warships. At the same time, the swaying situation of warships at sea is simulated and tested to improve the recognition difficulty so as to cope with the impact caused by complex sea conditions. The effect of different types of ships as transfer objects is also discussed. The experimental results show that the improved algorithm improves the mean average accuracy (mAP) by 10.2% and the first hit rate (Rank1) by 4.9% on average.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available