Detailled semantic labeling (2D) result


Name K. Nogueira et al.
Affiliation Universidade Federal de Minas Gerais, Brazil
Abbreviation UFMG_1
Strategy (u)nsupervised, (s)upervised, (h)ybrid s


Overall statistics, reference set: full_reference


Number of non processed tiles by this participant: 0
↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.878
0.017
0.043
0.016
0.002
0.043
building
0.021
0.946
0.012
0.005
0.002
0.015
low_veg
0.038
0.009
0.864
0.064
0.000
0.025
tree
0.047
0.009
0.184
0.747
0.004
0.010
car
0.078
0.005
0.004
0.009
0.878
0.026
clutter
0.286
0.099
0.128
0.026
0.020
0.442
Precision/Correctness
0.886
0.944
0.771
0.858
0.859
0.453
Recall/Completeness
0.878
0.946
0.864
0.747
0.878
0.442
F1
0.882
0.945
0.815
0.798
0.868
0.447

Overall accuracy 0.849

Overall statistics, reference set: no_boundary


Number of non processed tiles by this participant: 0
↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.895
0.013
0.037
0.012
0.001
0.042
building
0.018
0.953
0.009
0.004
0.002
0.014
low_veg
0.028
0.005
0.886
0.058
0.000
0.023
tree
0.037
0.008
0.169
0.773
0.003
0.009
car
0.013
0.006
0.002
0.006
0.946
0.028
clutter
0.290
0.084
0.113
0.022
0.019
0.471
Precision/Correctness
0.906
0.959
0.793
0.882
0.881
0.460
Recall/Completeness
0.895
0.953
0.886
0.773
0.946
0.471
F1
0.901
0.956
0.837
0.824
0.913
0.465

Overall accuracy 0.870
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8470
0.0158
0.0414
0.0112
0.0026
0.0820
building
0.0235
0.9370
0.0110
0.0094
0.0008
0.0183
low_veg
0.0318
0.0088
0.8768
0.0448
0.0002
0.0376
tree
0.0435
0.0177
0.1950
0.7274
0.0021
0.0143
car
0.0877
0.0043
0.0039
0.0032
0.8539
0.0469
clutter
0.1046
0.1450
0.1269
0.0181
0.0529
0.5524
Precision/Correctness
0.885
0.927
0.816
0.902
0.808
0.215
Recall/Completeness
0.847
0.937
0.877
0.727
0.854
0.552
F1
0.866
0.932
0.845
0.805
0.830
0.310

Overall accuracy 0.839

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_2_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8685
0.0111
0.0314
0.0076
0.0007
0.0807
building
0.0204
0.9467
0.0082
0.0084
0.0007
0.0156
low_veg
0.0257
0.0048
0.8969
0.0366
0.0001
0.0359
tree
0.0352
0.0162
0.1749
0.7591
0.0015
0.0131
car
0.0331
0.0043
0.0030
0.0008
0.9092
0.0496
clutter
0.0759
0.1374
0.0918
0.0120
0.0616
0.6213
Precision/Correctness
0.909
0.948
0.841
0.923
0.844
0.192
Recall/Completeness
0.869
0.947
0.897
0.759
0.909
0.621
F1
0.888
0.947
0.868
0.833
0.876
0.294

Overall accuracy 0.863

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_2_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8462
0.0076
0.0759
0.0095
0.0020
0.0588
building
0.0187
0.9495
0.0123
0.0066
0.0007
0.0122
low_veg
0.0185
0.0033
0.9226
0.0355
0.0001
0.0199
tree
0.0259
0.0065
0.2490
0.7065
0.0013
0.0108
car
0.0536
0.0083
0.0040
0.0042
0.9088
0.0211
clutter
0.0908
0.2494
0.1302
0.0517
0.0787
0.3992
Precision/Correctness
0.874
0.918
0.790
0.937
0.827
0.061
Recall/Completeness
0.846
0.949
0.923
0.706
0.909
0.399
F1
0.860
0.934
0.851
0.806
0.866
0.106

Overall accuracy 0.834

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_2_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8706
0.0050
0.0580
0.0059
0.0008
0.0597
building
0.0169
0.9617
0.0079
0.0045
0.0005
0.0085
low_veg
0.0140
0.0017
0.9356
0.0293
0.0001
0.0193
tree
0.0221
0.0059
0.2343
0.7264
0.0010
0.0104
car
0.0040
0.0090
0.0008
0.0019
0.9638
0.0205
clutter
0.0621
0.2869
0.0933
0.0590
0.0958
0.4028
Precision/Correctness
0.895
0.939
0.805
0.951
0.854
0.046
Recall/Completeness
0.871
0.962
0.936
0.726
0.964
0.403
F1
0.883
0.950
0.866
0.824
0.905
0.083

Overall accuracy 0.851

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_3_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8544
0.0182
0.0598
0.0173
0.0031
0.0471
building
0.0178
0.9616
0.0076
0.0039
0.0003
0.0088
low_veg
0.0491
0.0199
0.8418
0.0673
0.0003
0.0216
tree
0.0335
0.0101
0.1427
0.7990
0.0042
0.0105
car
0.0727
0.0044
0.0031
0.0104
0.8941
0.0153
clutter
0.2327
0.1462
0.2158
0.0284
0.0226
0.3543
Precision/Correctness
0.877
0.940
0.779
0.895
0.864
0.280
Recall/Completeness
0.854
0.962
0.842
0.799
0.894
0.354
F1
0.865
0.951
0.809
0.844
0.879
0.313

Overall accuracy 0.854

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_3_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8804
0.0117
0.0492
0.0116
0.0004
0.0467
building
0.0155
0.9683
0.0054
0.0030
0.0002
0.0076
low_veg
0.0396
0.0104
0.8723
0.0585
0.0002
0.0191
tree
0.0264
0.0084
0.1256
0.8262
0.0035
0.0100
car
0.0128
0.0041
0.0008
0.0062
0.9618
0.0142
clutter
0.2333
0.1391
0.1893
0.0201
0.0264
0.3919
Precision/Correctness
0.903
0.962
0.807
0.919
0.898
0.261
Recall/Completeness
0.880
0.968
0.872
0.826
0.962
0.392
F1
0.892
0.965
0.839
0.870
0.929
0.313

Overall accuracy 0.881

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_3_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8401
0.0209
0.0704
0.0198
0.0049
0.0439
building
0.0154
0.9636
0.0095
0.0035
0.0005
0.0075
low_veg
0.0294
0.0102
0.8903
0.0589
0.0001
0.0112
tree
0.0303
0.0128
0.2665
0.6774
0.0026
0.0104
car
0.0602
0.0096
0.0047
0.0150
0.8909
0.0196
clutter
0.1930
0.1464
0.1673
0.0326
0.0164
0.4443
Precision/Correctness
0.873
0.934
0.727
0.888
0.822
0.444
Recall/Completeness
0.840
0.964
0.890
0.677
0.891
0.444
F1
0.856
0.949
0.800
0.768
0.855
0.444

Overall accuracy 0.828

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_3_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8671
0.0147
0.0582
0.0142
0.0021
0.0436
building
0.0134
0.9703
0.0071
0.0027
0.0003
0.0061
low_veg
0.0205
0.0049
0.9155
0.0499
0.0000
0.0092
tree
0.0257
0.0117
0.2535
0.6972
0.0023
0.0096
car
0.0083
0.0095
0.0021
0.0089
0.9519
0.0193
clutter
0.1857
0.1377
0.1375
0.0256
0.0178
0.4958
Precision/Correctness
0.899
0.953
0.744
0.912
0.853
0.448
Recall/Completeness
0.867
0.970
0.915
0.697
0.952
0.496
F1
0.883
0.961
0.821
0.790
0.900
0.471

Overall accuracy 0.850

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_4_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9015
0.0130
0.0427
0.0219
0.0013
0.0196
building
0.0137
0.9606
0.0096
0.0063
0.0005
0.0093
low_veg
0.0336
0.0109
0.8276
0.1160
0.0001
0.0118
tree
0.0476
0.0058
0.0724
0.8619
0.0039
0.0084
car
0.0775
0.0035
0.0070
0.0132
0.8834
0.0155
clutter
0.3858
0.1119
0.1729
0.0593
0.0152
0.2548
Precision/Correctness
0.863
0.958
0.763
0.787
0.902
0.587
Recall/Completeness
0.901
0.961
0.828
0.862
0.883
0.255
F1
0.882
0.959
0.794
0.823
0.892
0.355

Overall accuracy 0.861

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_4_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9192
0.0099
0.0354
0.0165
0.0003
0.0187
building
0.0095
0.9706
0.0061
0.0052
0.0004
0.0083
low_veg
0.0247
0.0064
0.8505
0.1079
0.0000
0.0105
tree
0.0360
0.0041
0.0582
0.8913
0.0029
0.0075
car
0.0080
0.0030
0.0024
0.0077
0.9626
0.0162
clutter
0.3916
0.0990
0.1614
0.0537
0.0157
0.2787
Precision/Correctness
0.888
0.970
0.796
0.818
0.917
0.600
Recall/Completeness
0.919
0.971
0.850
0.891
0.963
0.279
F1
0.903
0.970
0.823
0.853
0.939
0.381

Overall accuracy 0.886

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_4_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8658
0.0164
0.0644
0.0123
0.0027
0.0385
building
0.0152
0.9638
0.0051
0.0045
0.0045
0.0070
low_veg
0.0727
0.0136
0.7815
0.0743
0.0006
0.0572
tree
0.0462
0.0088
0.1619
0.7663
0.0044
0.0123
car
0.0745
0.0073
0.0046
0.0115
0.8810
0.0211
clutter
0.5016
0.0422
0.1341
0.0107
0.0043
0.3071
Precision/Correctness
0.734
0.936
0.746
0.865
0.871
0.528
Recall/Completeness
0.866
0.964
0.782
0.766
0.881
0.307
F1
0.795
0.950
0.763
0.813
0.876
0.388

Overall accuracy 0.792

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_4_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8824
0.0117
0.0592
0.0087
0.0003
0.0377
building
0.0127
0.9710
0.0030
0.0034
0.0042
0.0057
low_veg
0.0612
0.0087
0.8038
0.0684
0.0002
0.0576
tree
0.0377
0.0075
0.1436
0.7963
0.0037
0.0112
car
0.0200
0.0077
0.0025
0.0076
0.9408
0.0214
clutter
0.5182
0.0386
0.1226
0.0079
0.0034
0.3093
Precision/Correctness
0.745
0.952
0.766
0.887
0.895
0.544
Recall/Completeness
0.882
0.971
0.804
0.796
0.941
0.309
F1
0.808
0.961
0.784
0.839
0.917
0.394

Overall accuracy 0.810

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_4_15_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8804
0.0151
0.0648
0.0059
0.0026
0.0312
building
0.0112
0.9694
0.0129
0.0030
0.0004
0.0031
low_veg
0.0416
0.0195
0.8770
0.0427
0.0004
0.0188
tree
0.0838
0.0099
0.1809
0.7039
0.0107
0.0109
car
0.0814
0.0041
0.0073
0.0023
0.8772
0.0277
clutter
0.3168
0.2156
0.1149
0.0178
0.0202
0.3147
Precision/Correctness
0.903
0.950
0.731
0.896
0.873
0.352
Recall/Completeness
0.880
0.969
0.877
0.704
0.877
0.315
F1
0.892
0.959
0.798
0.788
0.875
0.332

Overall accuracy 0.868

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_4_15_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8968
0.0088
0.0595
0.0040
0.0008
0.0300
building
0.0093
0.9752
0.0106
0.0023
0.0003
0.0022
low_veg
0.0257
0.0116
0.9089
0.0375
0.0001
0.0163
tree
0.0697
0.0082
0.1597
0.7428
0.0091
0.0103
car
0.0173
0.0043
0.0010
0.0007
0.9468
0.0298
clutter
0.3397
0.2164
0.0987
0.0143
0.0181
0.3129
Precision/Correctness
0.927
0.966
0.753
0.919
0.900
0.326
Recall/Completeness
0.897
0.975
0.909
0.743
0.947
0.313
F1
0.911
0.971
0.824
0.822
0.923
0.319

Overall accuracy 0.892

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_5_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8770
0.0222
0.0426
0.0184
0.0019
0.0380
building
0.0184
0.9450
0.0088
0.0045
0.0009
0.0224
low_veg
0.0449
0.0088
0.8567
0.0657
0.0003
0.0236
tree
0.0679
0.0052
0.1820
0.7357
0.0041
0.0050
car
0.1003
0.0024
0.0007
0.0050
0.8639
0.0277
clutter
0.2019
0.2183
0.1283
0.0277
0.0917
0.3322
Precision/Correctness
0.903
0.941
0.791
0.813
0.833
0.301
Recall/Completeness
0.877
0.945
0.857
0.736
0.864
0.332
F1
0.890
0.943
0.822
0.772
0.848
0.316

Overall accuracy 0.857

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_5_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8939
0.0179
0.0373
0.0142
0.0005
0.0362
building
0.0151
0.9518
0.0071
0.0038
0.0008
0.0215
low_veg
0.0389
0.0047
0.8765
0.0592
0.0002
0.0205
tree
0.0526
0.0037
0.1677
0.7680
0.0036
0.0043
car
0.0176
0.0028
0.0001
0.0035
0.9440
0.0320
clutter
0.1766
0.2053
0.1172
0.0270
0.1101
0.3638
Precision/Correctness
0.928
0.958
0.816
0.839
0.848
0.279
Recall/Completeness
0.894
0.952
0.876
0.768
0.944
0.364
F1
0.911
0.955
0.845
0.802
0.893
0.316

Overall accuracy 0.880

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_5_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8910
0.0071
0.0368
0.0224
0.0011
0.0415
building
0.0176
0.9627
0.0055
0.0020
0.0002
0.0120
low_veg
0.0507
0.0050
0.8197
0.0925
0.0003
0.0318
tree
0.0645
0.0090
0.1408
0.7621
0.0064
0.0172
car
0.0962
0.0119
0.0080
0.0061
0.8514
0.0264
clutter
0.1767
0.1567
0.1343
0.0404
0.0141
0.4777
Precision/Correctness
0.853
0.936
0.787
0.823
0.867
0.594
Recall/Completeness
0.891
0.963
0.820
0.762
0.851
0.478
F1
0.872
0.949
0.803
0.791
0.859
0.529

Overall accuracy 0.840

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_5_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9082
0.0047
0.0296
0.0179
0.0002
0.0395
building
0.0155
0.9687
0.0033
0.0015
0.0001
0.0109
low_veg
0.0379
0.0023
0.8412
0.0892
0.0001
0.0293
tree
0.0541
0.0079
0.1318
0.7850
0.0054
0.0159
car
0.0122
0.0141
0.0062
0.0032
0.9374
0.0270
clutter
0.1636
0.1311
0.1206
0.0399
0.0143
0.5305
Precision/Correctness
0.881
0.954
0.809
0.843
0.883
0.613
Recall/Completeness
0.908
0.969
0.841
0.785
0.937
0.531
F1
0.894
0.961
0.825
0.813
0.909
0.569

Overall accuracy 0.863

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_5_15_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8580
0.0067
0.0505
0.0190
0.0016
0.0641
building
0.0154
0.9493
0.0139
0.0066
0.0014
0.0134
low_veg
0.0419
0.0044
0.8101
0.1191
0.0001
0.0244
tree
0.0446
0.0078
0.1252
0.8122
0.0045
0.0058
car
0.0988
0.0048
0.0039
0.0113
0.8628
0.0184
clutter
0.3609
0.1035
0.1176
0.0271
0.0097
0.3811
Precision/Correctness
0.889
0.953
0.714
0.768
0.912
0.443
Recall/Completeness
0.858
0.949
0.810
0.812
0.863
0.381
F1
0.873
0.951
0.759
0.790
0.887
0.410

Overall accuracy 0.835

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_5_15_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8711
0.0037
0.0451
0.0150
0.0007
0.0645
building
0.0120
0.9581
0.0120
0.0050
0.0012
0.0116
low_veg
0.0273
0.0013
0.8327
0.1157
0.0000
0.0230
tree
0.0342
0.0062
0.1142
0.8363
0.0040
0.0051
car
0.0154
0.0056
0.0008
0.0095
0.9499
0.0188
clutter
0.3799
0.0729
0.1082
0.0259
0.0091
0.4040
Precision/Correctness
0.907
0.971
0.734
0.794
0.925
0.444
Recall/Completeness
0.871
0.958
0.833
0.836
0.950
0.404
F1
0.889
0.965
0.780
0.815
0.937
0.423

Overall accuracy 0.855

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_6_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9186
0.0247
0.0229
0.0118
0.0013
0.0208
building
0.0321
0.9392
0.0163
0.0036
0.0002
0.0086
low_veg
0.0190
0.0035
0.9448
0.0295
0.0000
0.0031
tree
0.1014
0.0032
0.1681
0.7195
0.0021
0.0056
car
0.0804
0.0095
0.0006
0.0022
0.7754
0.1320
clutter
0.3889
0.0610
0.1330
0.0327
0.0144
0.3699
Precision/Correctness
0.912
0.950
0.864
0.868
0.899
0.370
Recall/Completeness
0.919
0.939
0.945
0.720
0.775
0.370
F1
0.915
0.945
0.902
0.787
0.833
0.370

Overall accuracy 0.893

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_6_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9310
0.0223
0.0185
0.0081
0.0006
0.0196
building
0.0283
0.9485
0.0123
0.0031
0.0002
0.0076
low_veg
0.0117
0.0019
0.9584
0.0260
0.0000
0.0021
tree
0.0812
0.0022
0.1557
0.7540
0.0017
0.0051
car
0.0131
0.0120
0.0003
0.0015
0.8273
0.1457
clutter
0.3980
0.0560
0.1133
0.0318
0.0131
0.3878
Precision/Correctness
0.930
0.958
0.882
0.892
0.924
0.374
Recall/Completeness
0.931
0.949
0.958
0.754
0.827
0.388
F1
0.930
0.953
0.919
0.817
0.873
0.381

Overall accuracy 0.910

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_6_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9201
0.0045
0.0281
0.0097
0.0031
0.0344
building
0.0514
0.9183
0.0060
0.0018
0.0082
0.0143
low_veg
0.0480
0.0045
0.8770
0.0560
0.0003
0.0142
tree
0.0607
0.0078
0.1672
0.7531
0.0076
0.0036
car
0.0574
0.0039
0.0047
0.0089
0.9117
0.0133
clutter
0.2918
0.1471
0.2196
0.0468
0.0175
0.2771
Precision/Correctness
0.906
0.976
0.799
0.833
0.856
0.237
Recall/Completeness
0.920
0.918
0.877
0.753
0.912
0.277
F1
0.913
0.946
0.836
0.791
0.883
0.255

Overall accuracy 0.882

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_6_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9324
0.0021
0.0232
0.0076
0.0001
0.0345
building
0.0472
0.9270
0.0033
0.0012
0.0081
0.0132
low_veg
0.0325
0.0016
0.9026
0.0509
0.0000
0.0124
tree
0.0499
0.0062
0.1535
0.7808
0.0063
0.0034
car
0.0045
0.0036
0.0027
0.0062
0.9698
0.0132
clutter
0.2852
0.1603
0.2160
0.0474
0.0166
0.2746
Precision/Correctness
0.924
0.983
0.821
0.859
0.872
0.220
Recall/Completeness
0.932
0.927
0.903
0.781
0.970
0.275
F1
0.928
0.954
0.860
0.818
0.919
0.244

Overall accuracy 0.899

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_6_15_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8885
0.0064
0.0416
0.0145
0.0038
0.0452
building
0.0101
0.9128
0.0464
0.0119
0.0017
0.0171
low_veg
0.0319
0.0026
0.8327
0.0804
0.0008
0.0516
tree
0.0358
0.0072
0.1822
0.7644
0.0053
0.0051
car
0.0662
0.0028
0.0048
0.0102
0.8968
0.0191
clutter
0.1367
0.0166
0.1267
0.0185
0.0329
0.6686
Precision/Correctness
0.936
0.977
0.664
0.846
0.853
0.571
Recall/Completeness
0.889
0.913
0.833
0.764
0.897
0.669
F1
0.912
0.944
0.739
0.803
0.874
0.616

Overall accuracy 0.854

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_6_15_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9060
0.0043
0.0341
0.0113
0.0003
0.0441
building
0.0069
0.9212
0.0443
0.0112
0.0015
0.0149
low_veg
0.0182
0.0010
0.8554
0.0742
0.0001
0.0511
tree
0.0274
0.0058
0.1652
0.7929
0.0043
0.0043
car
0.0094
0.0029
0.0008
0.0070
0.9594
0.0206
clutter
0.1340
0.0143
0.1183
0.0152
0.0315
0.6867
Precision/Correctness
0.954
0.984
0.680
0.870
0.887
0.588
Recall/Completeness
0.906
0.921
0.855
0.793
0.959
0.687
F1
0.929
0.952
0.758
0.830
0.922
0.633

Overall accuracy 0.873

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_7_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8632
0.0430
0.0166
0.0261
0.0016
0.0495
building
0.0231
0.8994
0.0046
0.0024
0.0066
0.0639
low_veg
0.0487
0.0123
0.8046
0.1182
0.0002
0.0159
tree
0.0294
0.0080
0.1734
0.7806
0.0018
0.0069
car
0.0584
0.0072
0.0029
0.0127
0.9058
0.0130
clutter
0.1917
0.0235
0.0331
0.0020
0.0107
0.7391
Precision/Correctness
0.934
0.888
0.759
0.702
0.784
0.619
Recall/Completeness
0.863
0.899
0.805
0.781
0.906
0.739
F1
0.897
0.894
0.781
0.739
0.841
0.673

Overall accuracy 0.848

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_7_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8755
0.0396
0.0139
0.0207
0.0006
0.0498
building
0.0209
0.9031
0.0035
0.0017
0.0063
0.0644
low_veg
0.0372
0.0090
0.8260
0.1132
0.0000
0.0147
tree
0.0179
0.0061
0.1632
0.8046
0.0012
0.0069
car
0.0067
0.0074
0.0011
0.0084
0.9626
0.0138
clutter
0.1848
0.0214
0.0314
0.0016
0.0096
0.7513
Precision/Correctness
0.942
0.899
0.778
0.729
0.793
0.624
Recall/Completeness
0.875
0.903
0.826
0.805
0.963
0.751
F1
0.908
0.901
0.802
0.765
0.870
0.681

Overall accuracy 0.861

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image