ASABE Technical Library - Abstract

Click on “Download PDF” for the PDF version or on the title for the HTML version.
If you are not an ASABE member or if your employer has not arranged for access to the full-text, view access and subscription options

Tracking of Japanese Ayu fish in a school using fine-tuned SORT and OC-SORT algorithms

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.org

Citation:  2026 ASABE Annual International Meeting  2600132.(doi:10.13031/aim.202600132)
Authors:   Alin Khaliduzzaman, Takato Shibayama, Hitoshi Habe, Takayuki Niizato, Hiroaki Kawashima
Keywords:   Fish school, computer vision, object detection, multiple object tracking, collective behavior analysis.

Abstract. Real-time detection and tracking of fish in a school might have multifold functions to contribute to collective behavior analysis and precision fish farming practices like individual monitoring for growth, anomaly detection, population counting, feed management, and guided and directional control for various measurements. Although the method called SORT (simple online and real-time tracking) and its extensions (e.g., SORT, DeepSORT, OC-SORT) are widely used for tracking humans (e.g., pedestrians), such algorithms might have a great challenge to track objects with a similar appearance (e.g., animals, fish). Among them, tracking fish in a school is much more difficult because of their similarity in shape and appearance. Therefore, this research aims to compare the performance of multiple object tracking (MOT) methods, specifically SORT and OC-SORT algorithms, since SORT is considered as a base algorithm whereas OC-SORT is advanced SORT extension. These SORT algorithms were applied for tracking a school of 20 Japanese juvenile ayu fish (~7 cm) in a square flat-water tank (3m x 3m). The performance of the SORT and OC-SORT algorithms were evaluated using MOT accuracy (MOTA), MOT precision (MOTP) and identity switching (IDsw). After hyperparameter fine-tuning, MOTA, MOTP and IDsw were found 0.90, 0.48 and 7 respectively for SORT whereas 0.91, 0.40 and 1 for OC-SORT. Thus, OC-SORT performed far better in such a complex scenario which can be used for further applied research on various physical and behavioral measurements of fish for individual, collective behavioral analysis, and precision fish farming practices in the future.

(Download PDF)    (Export to EndNotes)