Add both demo Spaces with accuracy comparison
Browse files
README.md
CHANGED
|
@@ -31,7 +31,7 @@ PyTorch U-Net models for **gray leaf spot** (*Magnaporthe* and related fungal) c
|
|
| 31 |
| `best_area_w_0.1.pt` | SmallUNet | ~250 K | 0.1 | Light area regularisation |
|
| 32 |
| `best_area_w_0.3.pt` | SmallUNet | ~250 K | 0.3 | Moderate area regularisation |
|
| 33 |
| `best_area_w_0.5.pt` | SmallUNet | ~250 K | 0.5 | Balanced BCE + area |
|
| 34 |
-
| **`best_area_w_0.7.pt`** | **SmallUNet** | **~250 K** | **0.7** | **Strong area consistency (used by demo)** |
|
| 35 |
|
| 36 |
All SmallUNet variants share the same architecture:
|
| 37 |
|
|
@@ -173,16 +173,20 @@ python src/predict.py --input raw --weights models/best_finetuned.pt --output-di
|
|
| 173 |
|
| 174 |
---
|
| 175 |
|
| 176 |
-
## Demo
|
| 177 |
|
| 178 |
-
|
| 179 |
|
| 180 |
-
**
|
| 181 |
|
| 182 |
Features: dish detection β colony segmentation β crack & hyphae analysis β 16 morphometric measurements β time-series growth charts β CSV/JSON export.
|
| 183 |
|
| 184 |
Source code: [`rotsl/grayleafspot-segmentation-demo`](https://huggingface.co/rotsl/grayleafspot-segmentation-demo) (model repo)
|
| 185 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 186 |
---
|
| 187 |
|
| 188 |
## Citation
|
|
|
|
| 31 |
| `best_area_w_0.1.pt` | SmallUNet | ~250 K | 0.1 | Light area regularisation |
|
| 32 |
| `best_area_w_0.3.pt` | SmallUNet | ~250 K | 0.3 | Moderate area regularisation |
|
| 33 |
| `best_area_w_0.5.pt` | SmallUNet | ~250 K | 0.5 | Balanced BCE + area |
|
| 34 |
+
| **`best_area_w_0.7.pt`** | **SmallUNet** | **~250 K** | **0.7** | **Strong area consistency (used by demo) β
recommended** |
|
| 35 |
|
| 36 |
All SmallUNet variants share the same architecture:
|
| 37 |
|
|
|
|
| 173 |
|
| 174 |
---
|
| 175 |
|
| 176 |
+
## Demo Spaces
|
| 177 |
|
| 178 |
+
### β
Recommended: [`rotsl/grayleafspot-segmentation-demo`](https://huggingface.co/spaces/rotsl/grayleafspot-segmentation-demo)
|
| 179 |
|
| 180 |
+
Uses **`best_area_w_0.7.pt`** (SmallUNet with area-consistency loss). More accurate segmentation with better boundary adherence thanks to the area-consistency regularisation.
|
| 181 |
|
| 182 |
Features: dish detection β colony segmentation β crack & hyphae analysis β 16 morphometric measurements β time-series growth charts β CSV/JSON export.
|
| 183 |
|
| 184 |
Source code: [`rotsl/grayleafspot-segmentation-demo`](https://huggingface.co/rotsl/grayleafspot-segmentation-demo) (model repo)
|
| 185 |
|
| 186 |
+
### Legacy: [`rotsl/fungal-colony-input`](https://huggingface.co/spaces/rotsl/fungal-colony-input)
|
| 187 |
+
|
| 188 |
+
Uses **`grayleafspot.pt`** (smp.Unet with ResNet-34 encoder). This is the earlier, larger model trained with standard BCE loss only β it is less accurate than the area-consistency variant above, particularly for colony boundary delineation and area estimation. Kept available for reference and backward compatibility.
|
| 189 |
+
|
| 190 |
---
|
| 191 |
|
| 192 |
## Citation
|