EMSO ERIC is glad to celebrate a milestone of the OBSEA Regional Facility within the iMagine project!
The OBSEA team created a workflow to automatically process underwater pictures, for automatic fish detection and classification using deep learning techniques. The video has been produced with a YOLOv8 neural network trained with 500 tagged images. This work has been funded by
the iMagine project (grant agreement number 101058625), which is focused on imaging data and services for aquatic science.
The computational resources for training were provided by the DEEP Hybrid Data Cloud Platform. Colleagues from ICM CSIC supported this work by helping in the labelling phase.
The derived datasets will be crucial for studying species’ presence or absence over time and understanding changes in abundance concerning environmental parameters, providing insights into the impact of climate change on the local fish community.
Have a look to the video here
https://www.youtube.com/watch?v=SEw79Gd6m5o
Colour labels represent different fish species, while numbers refer to the precision in identifying them on a scale from 0 to 1 (the closer to 1, the more accurate the identification).
Find out more here on OBSEA Regional Facility