---
license: cc-by-4.0
tags:
- ocean
- midwater
- object-detection
---

# MBARI Midwater Supercategory Detector

## Model Details
- Trained by researchers at [CVisionAI](https://www.cvisionai.com/) and [Monterey Bay Aquarium Research Institute](https://www.mbari.org/) (MBARI).
- [YOLOv5v6.2](https://github.com/ultralytics/yolov5/tree/v6.2)
- Object detection
- Fine tuned yolov5l to detect 22 morphotaxonomic categories of midwater animals in the Greater Monterey Bay Area off the coast of Central California.

## Intended Use
- Make real time detections on video feed from MBARI Remotely Operated Vehicles.
- Post-process video collected in the region by MBARI vehicles.

## Factors
- Distribution shifts related to sampling platform, camera parameters, illumination, and deployment environment are expected to impact model performance.
- Evaluation was performed on an IID subset of available training data. Data to test out of distribution performance not currently available.

## Metrics
- [Precision-Recall curve](https://huggingface.co/FathomNet/MBARI-midwater-supercategory-detector/blob/main/plots/PR_curve.png) and [per class accuracy]((https://huggingface.co/FathomNet/MBARI-midwater-supercategory-detector/blob/main/plots/confusion_matrix.png)) were evaluated at test time. 
- mAP@0.5 = 0.866
- Indicates reasonably good performance for target task.

## Training and Evaluation Data
- A combination of publicly available [FathomNet](https://fathomnet.org/fathomnet/#/) and internal MBARI data
- Class labels have a [long tail and localizations occur throughout the frame](https://huggingface.co/FathomNet/MBARI-midwater-supercategory-detector/blob/main/plots/labels.jpg).

## Deployment
In an environment running [YOLOv5v6.2](https://github.com/ultralytics/yolov5/tree/v6.2):

```
python classify/predict.py --weights best.pt --data data/images/
```