Title:
DEVICE, SYSTEM, METHOD, AND PROGRAM FOR LEARNING FISH SPECIES
Document Type and Number:
WIPO Patent Application WO/2024/057626
Kind Code:
A1
Abstract:
To provide a fish species learning device, fish species learning system, fish species learning method and a program that can more easily and efficiently acquire annotation data for machine learning. To solve this problem, a server (20) (a fish species learning device) includes a storage unit (202) configured to store echo data of a given amount of time and a control unit (201). The control unit (201) is configured to calculate a feature value related to fish species for each fish school in the echo data of the given amount of time, group the fish schools based on a similarity of the calculated feature values, set a priority for each group based on a priority condition specified for improving an accuracy of fish species discrimination, and, based on the priority of each group, determine the group for which the echo data of the fish school is subjected to annotation processing.
Inventors:
KASAI AKINORI (JP)
HIRABAYASHI YUTA (JP)
MURAGAKI MASASHI (JP)
HIRABAYASHI YUTA (JP)
MURAGAKI MASASHI (JP)
Application Number:
PCT/JP2023/019778
Publication Date:
March 21, 2024
Filing Date:
May 26, 2023
Export Citation:
Assignee:
FURUNO ELECTRIC CO (JP)
International Classes:
G06Q50/02; A01K61/95; G01S15/96; G06N20/00
Foreign References:
| EP3761059A1 | 2021-01-06 | |||
| JP2014190787A | 2014-10-06 |
Other References:
HANDOKO, YEFFRY ET AL.: "Using Echo Ultrasound from Schooling Fish to Detect and Classify Fish Types", JOURNAL OF BIONIC ENGINEERING, vol. 6, 17 October 2009 (2009-10-17), pages 264 - 269, XP026698034, DOI: https://doi.org/10.1016/S1672-6529(08)60120-1
Download PDF: