Trials for detecting and classifying fish with the help of AI at OBSEA

Updated 04/10/2023

We’re here to celebrate a milestone of OBSEA thanks to the iMagine project!

OBSEA is an underwater observatory in the area of Barcelona which collects images of various fish species. However, identifying fish species with the manual analysis of the extensive image dataset is time-consuming; moreover, analysing only a subset of the whole dataset would imply losing important information.

AI tools applied to the fish images allow for extracting valuable biological content and creating derived datasets that marine scientists can use to draw ecological conclusions. Participating in the iMagine project and utilising the iMagine platform, a Deep Learning service has been trained and deployed to obtain species abundance data from existing and future images. These 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.

The team at OBSEA created a workflow to automatically process underwater pictures, extracting fish abundance and taxa information.
The workflow consists of two steps: segmentation and classification. Segmentation selects the regions of interest where fish specimens are present, and classification determines the taxa. Colleagues from ICM CSIC supported the species labelling phase by providing labelled deep-sea images.
In addition, the iMagine AI Platform provided the computational resources needed to train the YOLOv8 neural network with 500 tagged photos via the CSIC and LIP compute centres of EGI.

Watch now

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.