Object detection with YOLO

Author(s) Mohammad Joudy avatar Mohammad Joudyorcid logoYvan Le Bras avatar Yvan Le Bras
Overview
Creative Commons License: CC-BY Questions:

  • How can I run YOLO models on marine images to detect species using Galaxy?

Objectives:

  • Detect marine species in underwater images using a pretrained YOLOv8 model

  • Understand the requirements and output types of a detection model

Requirements:
Time estimation: 45 minutes
Supporting Materials:
Published: May 6, 2026
Last modification: May 6, 2026
License: Tutorial Content is licensed under Creative Commons Attribution 4.0 International License. The GTN Framework is licensed under MIT
Revision: 1

YOLO (You Only Look Once) is a fast, deep learning-based algorithm for real-time object detection. It predicts object classes and bounding boxes in a single pass over the image. YOLOv8 is a specific version, offering improved accuracy and speed.

In this tutorial, you will use Galaxy to run YOLOv8 model for Object Detection using real underwater images from the SEANOE dataset.

Agenda

In this tutotial, we will cover:

  1. Detection of marine species
    1. Input Dataset
    2. Model
    3. Run YOLOv8 in detect mode
    4. Explore the Outputs
    5. What to Look For
  2. Next Steps

Detection of marine species

Input Dataset

We will use selected images from the SEANOE dataset Lebeaud et al. 2024.

The SEANOE collection features real underwater images captured by deep‑sea observatories as part of a citizen science initiative called Deep Sea Spy. These non‑destructive imaging stations continuously monitor marine ecosystems and provide snapshots of various fauna. In this dataset, multiple annotators, including trained scientists and enthusiastic citizen scientists, have manually labeled images with polygons, lines, or points highlighting marine organisms. These annotations were then cleaned and converted into bounding boxes to create a training-ready dataset for object detection with YOLOv8. Though the exact species vary, images often include deep-sea fish, species, making this dataset well-suited for practicing detection tasks.

sample buccinid data. sample bythongraede data. sample buccinid2 data. sample buccinid3 data.

Get data

Hands On: Data Upload
  1. Create a new history for this tutorial
    • Give it a name (example: “DeepSeaSpy Yolo tutorial”) for you to find it again later if needed.

    To create a new history simply click the icon at the top of the history panel:

    UI for creating new history

    1. Click on (Edit) next to the history name (which by default is “Unnamed history”)
    2. Type the new name
    3. Click on Save
    4. To cancel renaming, click the “Cancel” button

    If you do not have the (Edit) next to the history name (which can be the case if you are using an older version of Galaxy) do the following:

    1. Click on Unnamed history (or the current name of the history) (Click to rename history) at the top of your history panel
    2. Type the new name
    3. Press Enter

  2. Import image data files and models from SEANOE marine datawarehouse.

    DeepSeaSpy image data files and models as a zip file:

    https://www.seanoe.org/data/00907/101899/data/115473.zip
    • Copy the link location

    • Click Upload at the top of the activity panel

    • Select Paste/Fetch Data

    • Paste the link(s) into the text field

    • Press Start

    • Close the window

  3. Unzip ( Galaxy version 6.0+galaxy0) with the following parameters:
    • “Input file”: 115473.zip

    This will create a data collection in your history where all archive files will be unzipped.

  4. Unhide the models data files:
    • In the history search bar, enter: name:detection deleted:false visible:any
    • Then unhide the 2 model files
      • “dataset_seanoe_101899_YOLOv8-weights-for-Bythograeidae-detection” and
      • “dataset_seanoe_101899_YOLOv8-weights-for-Buccinidae-detection”.

    If your history contains hidden datasets you will see “Include hidden” button directly above the dataset display.

    To un-hide datasets:

    • Type visible:hidden in the search box
    • Select datasets you want to un-hide
    • Click the dropdown that would appear at the top of the history;
    • Select “Unhide” option.

    An animated gif showing how to unhide datasets

    Alternatively, you can:

    • click “Include hidden” button directly above dataset display. This will cause hidden datasets to appear in history along with normal (un-hidden) datasets;
    • hidden datasets are distinguished by having within dataset box. Clicking on this icon will un-hide a given dataset;

  5. Select a sample of 100 image files and create a dedicated data collection
    • History search extension:jpg deleted:false visible:any then
    • click on “Select” and “Select All”.
    • click on “All 3979 selected” and “Build Dataset List” (or “Auto build list” depending on the selection mode),
    • select 100 files and give a name of the data collection, “DeepSeaSpy 100 images sample” for example.

    To select only last 100 files, you can use the history search function and specify extension:jpg deleted:false hid>XXXX visible:any in the search bar where XXXX is the id of the last image dataset minus 100. For example extension:jpg deleted:false hid>3885visible:any if you have images until the history dataset ID 3985.

  6. Create class name file “Buccinide_classnames”, copying and pasting this content in the file uploader: Optional step - Only if using under “Galaxy Version 8.3.0+galaxy5” tool version

    Autre poisson
Couverture de moules
Couverture microbienne
Couverture vers tubicole
Crabe araignée
Crabe bythograeidé
Crevette alvinocarididae
Escargot buccinidé
Ophiure
Poisson Cataetyx
Poisson chimère
Poisson zoarcidé
Pycnogonide
Ver polynoidé
Vers polynoidés
    • Click Upload Data at the top of the tool panel
    • Select Paste/Fetch Data at the bottom
    • Paste the file contents into the text field
    • Press Start and Close the window

Model

This dataset provides two pretrained YOLOv8 detection models tailored for the marine species found in SEANOE. One model detects Buccinidae (a family of sea snails), and the other targets Bythograeidae (a family of deep-sea crabs). These models were trained on cleaned annotation sets that contain thousands of examples—for instance, the Buccinidae set includes over 14,900 annotations in total. For this tutorial, you’ll find two model files—*.pt files—each accompanied by the appropriate class_names.txt file. You can upload either or both to Galaxy to run detection experiments on your underwater images.

Run YOLOv8 in detect mode

Hands On: Detect Buccinid snails on images
  1. Perform YOLO image labeling ( Galaxy version 8.3.0+galaxy2) with the following parameters:
    • “Input images”: DeepSeaSpy 100 images sample (click the button, then select “DeepSeaSpy 100 images sample” from the list)
    • “YOLO class name file”: Buccinide_classnames (input plain text file that lists the names of the classes the model can detect)
    • “Model file”: dataset_seanoe_101899_YOLOv8-weights-for-Buccinidae-detection (click the “…” button, then search for the name of the file)
    • “Prediction mode”: detect
    • “Image size”: 1000
    • “Confidence”: 0.25
    • “IoU”: 0.45
    • “Max detections”: 300
    Warning: Models type

    The model is trained only for detection, not segmentation.

Try changing the confidence and IoU thresholds (this is used for Non-Maximum Suppression (NMS), which removes overlapping detections) to see how detection results vary. It helps you find a good balance between sensitivity and accuracy.

Comment: Additional information on class names file and parameters

Concerning the class names file: Each class name must be on its own line, in the same order used during model training. So the class ID 0 corresponds to Buccinidae, and 1 to Bythograeidae. Concerning tool parameters:

  • “Image size”: Use 1000 (or a smaller number like 640 if processing speed is important). This controls how much the image is resized before prediction. Smaller values = faster but possibly less accurate.
  • “Confidence”: Set to 0.25 (25%). This controls how confident the model must be to report a detection. If you increase this value (e.g., 0.5), you’ll get fewer detections, but they’ll be more confident. If you lower it (e.g., 0.1), you may get more results, but possibly more false positives.
  • “IoU”: Set to 0.45. This is used for Non-Maximum Suppression (NMS), which removes overlapping detections. A higher IoU value (e.g., 0.7) keeps more overlapping boxes. A lower IoU (e.g., 0.3) removes more overlaps, which may help clean up crowded images.
  • “Max detections”: Set a reasonable cap like 300. This limits the number of objects detected per image.

Explore the Outputs

After running the tool, Galaxy will give you several output files for each image. Let’s go through what each one means and how to use them:

Text files (*.txt)

These are plain text files containing the detection results. Each line in a file shows:

<class_id> <confidence_score> <x_center> <y_center> <width> <height>

For example: 0 0.82 350 200 100 120

Question
  1. What is the meaning of these lines ?
  1. Class ID 0 (in our case, Buccinidae)
  2. Detected with 82% confidence
  3. The bounding box is centered at (350, 200) and has a width of 100 and height of 120 (in pixels, relative to the image. You can use this file to do further analysis, like counting species or tracking locations over time

Overlay images (*.jpg)

These are your original images with colored boxes drawn around detected species. Each box also includes: The class name and the confidence score. For example, you might see a box labeled: Esxcargot buccinidé 0.40

buccinid.

And if you reexecute the Yolo labeling tool specifying the “dataset_seanoe_101899_YOLOv8-weights-for-Bythograeidae-detection” model file, you will obtain detection result for crabs. bythog.

These images are useful for visually checking whether detections are correct or if something was missed.

Comment: What's the `?` in the labels ?

The ? character you see in the annotation is due to an incompatibility between the class names in the pre-trained model and the character encoding used on the system where the model was originally trained. This typically happens when non-UTF-8 characters are not properly handled during training or export.

No masks or segmentation files

Since we used detect mode, this tool will not generate segmentation masks (like .tiff or polygon files). Those are only available in segment mode, which we’ll cover in a separate tutorial.

What to Look For

  • Are species detected correctly?
  • Any false positives or missed detections?
  • What confidence levels do you observe?
  • How many objects per image?

Next Steps

Want to go further?

  • Train your own model on SEANOE annotations using YOLO training tool in Galaxy ( Galaxy version 8.3.0+galaxy2)
  • Get more information from YOLOv8 training notebook