Input

The model is trained on an image dataset composed of photoquadrats
collected by drop-camera and SCUBA divers as part as the MMP
(Mckenzie et al., 2022a), the Seagrass-Watch Global Seagrass
Observing Network (Seagrass-Watch, 2022) and the Torres Strait
Ranger Subtidal Monitoring Program (Carter et al., 2021b). Images
were captured between 2014 and 2024 from 28 sites across 18 unique
locations within the coastal and reef subtidal habitats from Torres
Strait to Hervey Bay.

The training dataset included :
- 524 images with at least 30% of oval species
- 1450 images with at least 30% of strappy species
- 331 images with at least 30% of ferny species
- 268 images with at least 30% of cylindrical species

Model architecture and training parameters

The seagrass cover class model was developped using Efficientnet. We used the EfficientNetB0 as a base for our training. We standardised our input images size to 600x600 pixels. The model converged after 59 epochs.
This is a multi-label model therefore the model prediction will return a probability between 0 and 1 for all morphological classes.

Model Performance

The testing dataset included :
- 163 images with at least 30% of oval species
- 408 images with at least 30% of strappy species
- 97 images with at least 30% of ferny species
- 68 images with at least 30% of cylindrical species

Model performance metrics (using 0.3 threshold) :

Class-wise Metrics:
oval: Precision=0.78, Recall=0.90, F1=0.84
strappy: Precision=0.94, Recall=0.98, F1=0.96
ferny: Precision=0.81, Recall=0.95, F1=0.88
cylindrical: Precision=0.83, Recall=0.88, F1=0.86

Example

Example oval	Example strappy	Example ferny	Example cylindrical

Classes

This model detects the presence of the folowing seagrass morphology classes:

• Oval Seagrass - Ho, Hd
• Strap-like Seagrass - Cs, Th, Hu, Cr, Zc
• Fern-like Seagrass - Hs
• Cylindrical Seagrass - Si

Note: The codes in italics correspond to specific seagrass species.
For full species names and descriptions, refer to the Seagrass-Watch Species ID Guide.

Annotation Method

Each image was visually assessed for seagrass cover and species composition by a trained seagrass scientist and underwent QAQC according to Seagrass-Watch protocols.

Model Limitations

This model was trained on specific dataset and therefore is finetuned for these specific locations. Model performance will vary on new images and we cannot garanty satifactory results. Please use the result output with caution and determine if this model is suited for your data.
This model is still under developpement.

Model Metadata

Tags:

Credits: Lucas A. Langlois, Catherine J. Collier and Len J. McKenzie