IGNF
/

sgiordano commited on
Commit
f7f2c7c
1 Parent(s): 5cff3bb

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -5
README.md CHANGED
@@ -96,7 +96,7 @@ pipeline_tag: image-segmentation
96
  <ul style="list-style-type:disc;">
97
  <li>Trained with the FLAIR-INC dataset</li>
98
  <li>RGB images (true colours)</li>
99
- <li>U-Net with a mitb5 encoder</li>
100
  <li>15 class nomenclature : [building, pervious surface, impervious surface, bare soil, water, coniferous, deciduous, brushwood, vineyard, herbaceous, agricultural land, plowed land, swimming pool, snow, greenhouse]</li>
101
  </ul>
102
  </div>
@@ -193,7 +193,7 @@ Statistics of the TRAIN+VALIDATION set :
193
  #### Training Hyperparameters
194
 
195
  * Model architecture: Unet (implementation from the [Segmentation Models Pytorch library](https://segmentation-modelspytorch.readthedocs.io/en/latest/docs/api.html#unet))
196
- * Encoder : Resnet-34 pre-trained with ImageNet
197
  * Augmentation :
198
  * VerticalFlip(p=0.5)
199
  * HorizontalFlip(p=0.5)
@@ -264,7 +264,7 @@ The following table give the class-wise metrics :
264
  | plowed_land | 42.202 | 59.355 | 65.114 | 54.532 |
265
  | swimming_pool | 0.000 | 0.000 | 0.000 | 0.000 |
266
  | snow | _0.000_ | _0.000_ | _0.000_ | _0.000_ |
267
- | greenhouse | 60.884 | 75.687 | 66.62 | 87.609 |
268
  | **average** | **53.440** | **65.146** | **65.517** | **65.644** |
269
 
270
 
@@ -278,9 +278,9 @@ The following illustration gives the resulting confusion matrix :
278
 
279
  <div style="position: relative; text-align: center;">
280
  <p style="margin: 0;">Normalized Confusion Matrix (precision)</p>
281
- <img src="figs/FLAIR-INC_RVBIE_resnet34_unet_15cl_norm_cm-precision.png" alt="drawing" style="width: 70%; display: block; margin: 0 auto;"/>
282
  <p style="margin: 0;">Normalized Confusion Matrix (recall)</p>
283
- <img src="figs/FLAIR-INC_RVBIE_resnet34_unet_15cl_norm_cm-recall.png" alt="drawing" style="width: 70%; display: block; margin: 0 auto;"/>
284
  </div>
285
 
286
 
 
96
  <ul style="list-style-type:disc;">
97
  <li>Trained with the FLAIR-INC dataset</li>
98
  <li>RGB images (true colours)</li>
99
+ <li>U-Net with a mit-b5 encoder</li>
100
  <li>15 class nomenclature : [building, pervious surface, impervious surface, bare soil, water, coniferous, deciduous, brushwood, vineyard, herbaceous, agricultural land, plowed land, swimming pool, snow, greenhouse]</li>
101
  </ul>
102
  </div>
 
193
  #### Training Hyperparameters
194
 
195
  * Model architecture: Unet (implementation from the [Segmentation Models Pytorch library](https://segmentation-modelspytorch.readthedocs.io/en/latest/docs/api.html#unet))
196
+ * Encoder : mit-b5 pre-trained with ImageNet
197
  * Augmentation :
198
  * VerticalFlip(p=0.5)
199
  * HorizontalFlip(p=0.5)
 
264
  | plowed_land | 42.202 | 59.355 | 65.114 | 54.532 |
265
  | swimming_pool | 0.000 | 0.000 | 0.000 | 0.000 |
266
  | snow | _0.000_ | _0.000_ | _0.000_ | _0.000_ |
267
+ | greenhouse | 60.884 | 75.687 | 66.620 | 87.609 |
268
  | **average** | **53.440** | **65.146** | **65.517** | **65.644** |
269
 
270
 
 
278
 
279
  <div style="position: relative; text-align: center;">
280
  <p style="margin: 0;">Normalized Confusion Matrix (precision)</p>
281
+ <img src="FLAIR-INC_rgb_15cl_mitb5-unet_confmat_norm-precision.png" alt="drawing" style="width: 70%; display: block; margin: 0 auto;"/>
282
  <p style="margin: 0;">Normalized Confusion Matrix (recall)</p>
283
+ <img src="FLAIR-INC_rgb_15cl_mitb5-unet_confmat_norm-recall.png" alt="drawing" style="width: 70%; display: block; margin: 0 auto;"/>
284
  </div>
285
 
286