pp. 1613-1617
S&M3274 Letter of Special Issue https://doi.org/10.18494/SAM4077 Published: May 12, 2023 Approaches to Upgrading the Performance of Fishing Vessel Recognition Technology [PDF] Ching-Hai Lin, Chun-Cheng Lin, Yu-Cheng Zhan, Cheng-Yu Yeh, and Shaw-Hwa Hwang (Received August 5, 2022; Accepted April 13, 2023) Keywords: fishing vessel identification, image recognition, deep learning, ArcFace
Fishing vessel recognition using face recognition has recently been addressed for the first time. This paper is actually an improved version of the original proposal, and there are two steps to improve the performance of fishing vessel recognition. In the first step, the number of recognizable fishing vessels was increased considerably from 156 to 272 and the numbers of images of different vessels were made as uniform as possible for a higher generalization ability. In the second step, an EfficientNet model was employed, input images were resized to 480 × 160 pixels to undistortedly display the side views of fishing vessels, and finally, the ArcFace loss function was used as well to train the presented model. As it turned out, the overall recognition performance was improved.
Corresponding author: Cheng-Yu YehThis work is licensed under a Creative Commons Attribution 4.0 International License. Cite this article Ching-Hai Lin, Chun-Cheng Lin, Yu-Cheng Zhan, Cheng-Yu Yeh, and Shaw-Hwa Hwang, Approaches to Upgrading the Performance of Fishing Vessel Recognition Technology, Sens. Mater., Vol. 35, No. 5, 2023, p. 1613-1617. |