We use transfer learning on yolov7 to detect bounding boxes around potential dolphin whistles.
| Data Type | Training Labels | Precision | Recall | [email protected] |
|---|---|---|---|---|
| REP23 | 430 | 0.59 | 0.50 | 0.31 |
| REP23+aug | 1300 | 0.66 | 0.69 | 0.55 |
| REP23+21+aug | 5000 | 0.79 | 0.69 | 0.77 |
| Name | Name | Last commit date | ||
|---|---|---|---|---|
We use transfer learning on yolov7 to detect bounding boxes around potential dolphin whistles.
| Data Type | Training Labels | Precision | Recall | [email protected] |
|---|---|---|---|---|
| REP23 | 430 | 0.59 | 0.50 | 0.31 |
| REP23+aug | 1300 | 0.66 | 0.69 | 0.55 |
| REP23+21+aug | 5000 | 0.79 | 0.69 | 0.77 |