metadata.yaml

metadata_version: 2.0.0
title: OBSEA Fish Detection
summary: AI-based fish detection and classification algorithm based on YOLOv8. The model has been finetuned to detect and classify fish at the OBSEA underwater observatory.
description: |-
  Provided by UPC, EMSO, and iMagine.

  This module helps users that want to train a deep learning object detection/classification algorithm (YOLOv8) with citizen fish data.
  It was developed with the intent to automatically detect and classify fish at [OBSEA Underwater Observatory](https://obsea.es), allowing the automatic 
  generation of biodiversity time-series. Currently it is able to detect and classify fish from the following species: Chromis chromis, 
  Coris julis, Dentex dentex, Diplodus cervinus, Diplodus puntazzo Diplodus sargus, Diplodus vulgaris, Epinephelus costae, 
  Epinephelus marginatus, Mullus surmuletus, Muraena helena, Parablennius gattorugine, Sarpa salpa, Seriola dumerili, Serranus cabrilla, 
  Sparus aurata, Symphodus mediterraneus, Aetomylaeus bovinus, Oblada melanurus 
  
  The user can:
  - Detect/classify fish in pictures/videos using the built-in model 
  - Train with custom data

  This module is mainly based on this already [existing module](https://github.com/ai4os-hub/ai4os-yolov8-torch)

  You can find more information on YOLOv8 here:
  * https://github.com/ultralytics/ultralytics
  * https://docs.ultralytics.com


dates:
  created: '2023-11-29'
  updated: '2024-11-19'
links:
  source_code: https://github.com/ai4os-hub/obsea-fish-detection/
  docker_image: ai4oshub/obsea-fish-detection
  dataset: https://doi.org/10.5281/zenodo.13903520
  ai4_template: ai4-template/1.9.9
tags:
  - deep learning
  - object detection
  - classification
  - vo.imagine-ai.eu
tasks:
  - Computer Vision
  - Classification
  - Object detection
  - Other
categories:
  - AI4 pre trained
  - AI4 trainable
  - AI4 inference
libraries:
  - PyTorch
data-type:
  - Image
  - Video
  - Other