sgiordano commited on
Commit
3495cd2
1 Parent(s): 1f57c62

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +20 -20
README.md CHANGED
@@ -247,34 +247,34 @@ FLAIR-INC_RVBIE_resnet34_unet_15cl_norm was obtained for num_epoch=76 with corre
247
  The evaluation was performed on a TEST set of 31750 patches that are independant from the TRAIN and VALIDATION patches. They represent 15 spatio-temporal domains.
248
  The TEST set corresponds to the reunion of the TEST set of scientific challenges FLAIR#1 and FLAIR#2. See the [FLAIR challenge page](https://ignf.github.io/FLAIR/) for more details.
249
 
250
- The choice of a separate TEST set instead of cross validation was made to be coherent with the FLAIR challenges. Please note that the _Snow_ class is absent from the TEST set.
 
251
  <!-- {{ testing_data | default("[More Information Needed]", true)}} -->
252
 
253
  #### Metrics
254
 
 
255
  The following table give the class-wise metrics :
256
 
257
- | Modalities | IoU | Fscore | Precision | Recall |
258
  | ----------------------- | ----------|---------|---------|---------|
259
- | building | 82.63 | 00.00 | 00.00 | 00.00 |
260
- | pervious surface | 53.24 | 00.00 | 00.00 | 00.00 |
261
- | impervious surface | 74.17 | 00.00 | 00.00 | 00.00 |
262
- | bare soil | 60.40 | 00.00 | 00.00 | 00.00 |
263
- | water | 87.59 | 00.00 | 00.00 | 00.00 |
264
- | coniferous | 46.35 | 00.00 | 00.00 | 00.00 |
265
- | deciduous | 67.45 | 00.00 | 00.00 | 00.00 |
266
- | brushwood | 30.23 | 00.00 | 00.00 | 00.00 |
267
- | vineyard | 82.93 | 00.00 | 00.00 | 00.00 |
268
- | herbaceous vegetation | 55.03 | 00.00 | 00.00 | 00.00 |
269
- | agricultural land | 52.01 | 00.00 | 00.00 | 00.00 |
270
- | plowed land | 40.84 | 00.00 | 00.00 | 00.00 |
271
- | swimming_pool | 48.44 | 00.00 | 00.00 | 00.00 |
272
  | snow | 00.00 | 00.00 | 00.00 | 00.00 |
273
- | greenhouse | 39.45 | 00.00 | 00.00 | 00.00 |
274
-
275
-
276
-
277
-
278
 
279
  <!-- These are the evaluation metrics being used, ideally with a description of why. -->
280
 
 
247
  The evaluation was performed on a TEST set of 31750 patches that are independant from the TRAIN and VALIDATION patches. They represent 15 spatio-temporal domains.
248
  The TEST set corresponds to the reunion of the TEST set of scientific challenges FLAIR#1 and FLAIR#2. See the [FLAIR challenge page](https://ignf.github.io/FLAIR/) for more details.
249
 
250
+ The choice of a separate TEST set instead of cross validation was made to be coherent with the FLAIR challenges.
251
+ However the metrics for the Challenge were calculated on 12 classes and the TEST set acordingly. As a result the _Snow_ class is unfortunately absent from the TEST set.
252
  <!-- {{ testing_data | default("[More Information Needed]", true)}} -->
253
 
254
  #### Metrics
255
 
256
+ With the evaluation protocol, the **FLAIR-INC_RVBIE_resnet34_unet_15cl_norm** have been evaluated to **OA= 76.37%** and **mIoU=54.71%**.
257
  The following table give the class-wise metrics :
258
 
259
+ | Modalities | IoU (%) | Fscore (%) | Precision (%) | Recall (%) |
260
  | ----------------------- | ----------|---------|---------|---------|
261
+ | building | 82.63 | 90.49 | 90.26 | 90.72 |
262
+ | pervious surface | 53.24 | 69.48 | 68.97 | 70.00 |
263
+ | impervious surface | 74.17 | 85.17 | 86.28 | 84.09 |
264
+ | bare soil | 60.40 | 75.31 | 80.49 | 70.75 |
265
+ | water | 87.59 | 93.38 | 93.16 | 93.61 |
266
+ | coniferous | 46.35 | 63.34 | 63.52 | 63.16 |
267
+ | deciduous | 67.45 | 80.56 | 77.44 | 83.94 |
268
+ | brushwood | 30.23 | 46.43 | 63.55 | 36.58 |
269
+ | vineyard | 82.93 | 90.67 | 91.35 | 89.99 |
270
+ | herbaceous vegetation | 55.03 | 70.99 | 70.59 | 71.40 |
271
+ | agricultural land | 52.01 | 68.43 | 59.18 | 81.12 |
272
+ | plowed land | 40.84 | 57.99 | 68.28 | 50.40 |
273
+ | swimming_pool | 48.44 | 65.27 | 81.62 | 54.37 |
274
  | snow | 00.00 | 00.00 | 00.00 | 00.00 |
275
+ | greenhouse | 39.45 | 56.57 | 45.52 | 74.72 |
276
+ | ----------------------- | ----------|---------|---------|---------|
277
+ | average | 54.72 | 67.60 | 69.35 | 67.66 |
 
 
278
 
279
  <!-- These are the evaluation metrics being used, ideally with a description of why. -->
280