Yolov8 model weights to detect unknown underwater sounds

Flanders Marine Institute (Flanders Marine Institute)

Abstract

Machine learning for efficient segregation and labeling of potential biological sounds in long-term underwater recordings.Trained weights of model, in Python.Code to reproduce publication results, (re)train the models, or use the models for inference can be found on: GitHub - lifewatch/sound-segregation-and-categorization (https://github.com/lifewatch/sound-segregation-and-categorization).

Authors 10

1. Clea Parcerisas

   Flanders Marine Institute · Ghent University

   Affiliation as printed

   Flanders Marine Institute (VLIZ), Belgium

   WAVES - UGent

2. Elena Schall

   Alfred-Wegener-Institut Helmholtz-Zentrum für Polar- und Meeresforschung

   Affiliation as printed

   Ocean Acoustics Group - AWI, Germany

3. Kees te Velde Aachen Institute of Biology, Universiteit Leiden

   Leiden University

   Affiliation as printed

   Institute of Biology - Universiteit Leiden, Netherlands

4. Dick B. Botteldooren

   Ghent University

   Affiliation as printed

   WAVES - UGent

5. Paul Devos

   Ghent University

   Affiliation as printed

   WAVES - UGent

6. Elisabeth Debusschere

   Flanders Marine Institute

   Affiliation as printed

   Flanders Marine Institute (VLIZ), Belgium

7. Belgium Flanders Marine Institute (VLIZ)
8. WAVES - UGent
9. Germany Ocean Acoustics Group - AWI
10. Netherlands Institute of Biology - Universiteit Leiden