Paper
8 December 2022 Automatic identification of tiger puffer in an aquaculture tank using deep learning
Atsushi Ito, Yusuke Fukushima, Hirotsugu Yamamoto, Yukitoshi Otani, Shiro Suyama, Masaki Yasugi, Yasutoshi Yoshiura
Author Affiliations +
Abstract
This paper presents an experiment to realize an automatic identification system of tiger puffer (torafugu) using Deep Learning. To meet the operation of growing and selling aquaculture fish, we tried to use Transfer Learning to reduce the operation cost to identify torafugu. In this trial, we used three torafugu. We took a video of them swimming in an aquaculture tank then extracted figures of each torafugu. Moreover, to increase the number of data for learning and testing, we finally got 150 pictures of each torafugu by rotating and flipping. As the result of the experiment, 60 to 80 pictures for each torafugu are enough for automatic identification. The generated learning model can identify torafugu under severe conditions such as unclear photos by waves and change of brightness, and movement of back born during swimming.
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Atsushi Ito, Yusuke Fukushima, Hirotsugu Yamamoto, Yukitoshi Otani, Shiro Suyama, Masaki Yasugi, and Yasutoshi Yoshiura "Automatic identification of tiger puffer in an aquaculture tank using deep learning", Proc. SPIE 12480, Optical Technology and Measurement for Industrial Applications Conference 2022, 124800H (8 December 2022); https://doi.org/10.1117/12.2660183
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Biology

Virtual reality

Imaging systems

Video

Airborne remote sensing

Cameras

Data modeling

Back to Top