Detailled semantic labeling (2D) result


Name Y. Sun
Affiliation Huazhong University of Science and Technology (HUST), Wuhan, China
Abbreviation HUSTW
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
0.918
0.019
0.032
0.018
0.002
0.010
building
0.026
0.961
0.006
0.003
0.000
0.004
low_veg
0.051
0.023
0.835
0.081
0.000
0.009
tree
0.039
0.007
0.107
0.839
0.002
0.005
car
0.123
0.015
0.001
0.017
0.818
0.027
clutter
0.237
0.153
0.104
0.029
0.007
0.471
Precision/Correctness
0.887
0.922
0.836
0.851
0.916
0.740
Recall/Completeness
0.918
0.961
0.835
0.839
0.818
0.471
F1
0.902
0.941
0.835
0.845
0.864
0.575

Overall accuracy 0.875

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
0.938
0.014
0.026
0.013
0.000
0.009
building
0.023
0.967
0.004
0.002
0.000
0.003
low_veg
0.038
0.017
0.866
0.071
0.000
0.008
tree
0.029
0.006
0.093
0.866
0.001
0.005
car
0.037
0.016
0.000
0.011
0.909
0.027
clutter
0.224
0.136
0.086
0.024
0.007
0.522
Precision/Correctness
0.912
0.942
0.862
0.879
0.955
0.769
Recall/Completeness
0.938
0.967
0.866
0.866
0.909
0.522
F1
0.925
0.954
0.864
0.872
0.932
0.622

Overall accuracy 0.900
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9285
0.0212
0.0288
0.0125
0.0019
0.0071
building
0.0200
0.9610
0.0085
0.0049
0.0000
0.0056
low_veg
0.0545
0.0122
0.8617
0.0602
0.0000
0.0114
tree
0.0431
0.0125
0.1153
0.8181
0.0012
0.0098
car
0.1307
0.0334
0.0008
0.0092
0.7836
0.0423
clutter
0.1781
0.2464
0.1266
0.0295
0.0161
0.4033
Precision/Correctness
0.867
0.913
0.873
0.892
0.895
0.463
Recall/Completeness
0.929
0.961
0.862
0.818
0.784
0.403
F1
0.896
0.936
0.867
0.854
0.836
0.431

Overall accuracy 0.877

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
0.9525
0.0142
0.0208
0.0077
0.0003
0.0046
building
0.0163
0.9694
0.0058
0.0041
0.0000
0.0044
low_veg
0.0444
0.0066
0.8920
0.0475
0.0000
0.0095
tree
0.0333
0.0107
0.0962
0.8501
0.0008
0.0089
car
0.0461
0.0398
0.0007
0.0042
0.8629
0.0463
clutter
0.1389
0.2560
0.0894
0.0204
0.0191
0.4762
Precision/Correctness
0.897
0.939
0.900
0.919
0.939
0.480
Recall/Completeness
0.952
0.969
0.892
0.850
0.863
0.476
F1
0.924
0.954
0.896
0.883
0.899
0.478

Overall accuracy 0.906

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
0.9203
0.0082
0.0538
0.0147
0.0014
0.0017
building
0.0121
0.9680
0.0111
0.0049
0.0000
0.0038
low_veg
0.0266
0.0059
0.9042
0.0618
0.0000
0.0015
tree
0.0239
0.0034
0.1101
0.8597
0.0007
0.0022
car
0.1101
0.0089
0.0001
0.0102
0.8493
0.0214
clutter
0.1358
0.3925
0.1548
0.0536
0.0091
0.2541
Precision/Correctness
0.868
0.913
0.885
0.915
0.912
0.309
Recall/Completeness
0.920
0.968
0.904
0.860
0.849
0.254
F1
0.894
0.940
0.894
0.887
0.879
0.279

Overall accuracy 0.893

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
0.9449
0.0042
0.0400
0.0097
0.0001
0.0011
building
0.0089
0.9786
0.0068
0.0031
0.0000
0.0026
low_veg
0.0188
0.0035
0.9252
0.0512
0.0000
0.0012
tree
0.0193
0.0026
0.0960
0.8797
0.0005
0.0020
car
0.0266
0.0099
0.0000
0.0060
0.9355
0.0220
clutter
0.0861
0.4512
0.1149
0.0521
0.0093
0.2864
Precision/Correctness
0.899
0.941
0.903
0.932
0.955
0.307
Recall/Completeness
0.945
0.979
0.925
0.880
0.936
0.286
F1
0.922
0.960
0.914
0.905
0.945
0.296

Overall accuracy 0.914

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
0.9095
0.0223
0.0395
0.0209
0.0032
0.0046
building
0.0135
0.9759
0.0038
0.0026
0.0000
0.0041
low_veg
0.0607
0.0247
0.8303
0.0785
0.0001
0.0057
tree
0.0286
0.0096
0.0867
0.8685
0.0020
0.0046
car
0.1312
0.0111
0.0005
0.0210
0.8211
0.0151
clutter
0.3826
0.1729
0.2498
0.0309
0.0088
0.1551
Precision/Correctness
0.866
0.931
0.839
0.889
0.905
0.435
Recall/Completeness
0.910
0.976
0.830
0.869
0.821
0.155
F1
0.887
0.953
0.835
0.878
0.861
0.229

Overall accuracy 0.879

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
0.9382
0.0146
0.0298
0.0136
0.0003
0.0035
building
0.0102
0.9831
0.0016
0.0019
0.0000
0.0033
low_veg
0.0478
0.0133
0.8686
0.0659
0.0000
0.0045
tree
0.0205
0.0078
0.0714
0.8948
0.0015
0.0041
car
0.0473
0.0127
0.0003
0.0133
0.9120
0.0144
clutter
0.4131
0.1689
0.2131
0.0215
0.0101
0.1733
Precision/Correctness
0.898
0.956
0.872
0.916
0.953
0.449
Recall/Completeness
0.938
0.983
0.869
0.895
0.912
0.173
F1
0.918
0.969
0.870
0.905
0.932
0.250

Overall accuracy 0.910

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
0.8910
0.0250
0.0490
0.0209
0.0026
0.0115
building
0.0175
0.9734
0.0040
0.0024
0.0000
0.0026
low_veg
0.0354
0.0230
0.8674
0.0707
0.0000
0.0035
tree
0.0273
0.0047
0.1529
0.8087
0.0014
0.0049
car
0.1216
0.0140
0.0004
0.0294
0.8063
0.0282
clutter
0.3650
0.2402
0.1916
0.0335
0.0019
0.1678
Precision/Correctness
0.850
0.914
0.806
0.891
0.908
0.478
Recall/Completeness
0.891
0.973
0.867
0.809
0.806
0.168
F1
0.870
0.943
0.836
0.848
0.854
0.248

Overall accuracy 0.859

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
0.9199
0.0177
0.0369
0.0143
0.0004
0.0108
building
0.0149
0.9798
0.0017
0.0016
0.0000
0.0020
low_veg
0.0230
0.0170
0.8990
0.0587
0.0000
0.0023
tree
0.0212
0.0035
0.1379
0.8319
0.0012
0.0043
car
0.0378
0.0153
0.0001
0.0203
0.8974
0.0290
clutter
0.3937
0.2373
0.1598
0.0257
0.0020
0.1815
Precision/Correctness
0.881
0.936
0.832
0.917
0.951
0.496
Recall/Completeness
0.920
0.980
0.899
0.832
0.897
0.182
F1
0.900
0.958
0.864
0.872
0.923
0.266

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_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9261
0.0130
0.0309
0.0248
0.0014
0.0038
building
0.0168
0.9694
0.0058
0.0043
0.0000
0.0038
low_veg
0.0538
0.0223
0.7934
0.1266
0.0000
0.0039
tree
0.0364
0.0042
0.0514
0.9027
0.0017
0.0036
car
0.1495
0.0042
0.0011
0.0286
0.7913
0.0253
clutter
0.4442
0.1495
0.1624
0.0721
0.0057
0.1661
Precision/Correctness
0.845
0.947
0.799
0.779
0.944
0.739
Recall/Completeness
0.926
0.969
0.793
0.903
0.791
0.166
F1
0.884
0.958
0.796
0.836
0.861
0.271

Overall accuracy 0.864

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
0.9463
0.0076
0.0248
0.0181
0.0001
0.0030
building
0.0123
0.9783
0.0029
0.0033
0.0000
0.0032
low_veg
0.0394
0.0139
0.8297
0.1140
0.0000
0.0029
tree
0.0241
0.0024
0.0393
0.9306
0.0010
0.0025
car
0.0502
0.0034
0.0005
0.0188
0.9026
0.0245
clutter
0.4564
0.1334
0.1523
0.0636
0.0062
0.1881
Precision/Correctness
0.874
0.964
0.834
0.816
0.968
0.782
Recall/Completeness
0.946
0.978
0.830
0.931
0.903
0.188
F1
0.909
0.971
0.832
0.869
0.934
0.303

Overall accuracy 0.892

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
0.8759
0.0223
0.0616
0.0141
0.0025
0.0236
building
0.0156
0.9721
0.0037
0.0043
0.0002
0.0040
low_veg
0.0874
0.0253
0.7715
0.0886
0.0001
0.0271
tree
0.0341
0.0077
0.1146
0.8336
0.0018
0.0082
car
0.1314
0.0120
0.0007
0.0254
0.8008
0.0297
clutter
0.1440
0.0691
0.0771
0.0153
0.0013
0.6932
Precision/Correctness
0.826
0.907
0.795
0.854
0.939
0.818
Recall/Completeness
0.876
0.972
0.772
0.834
0.801
0.693
F1
0.850
0.939
0.783
0.844
0.864
0.750

Overall accuracy 0.843

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
0.8949
0.0168
0.0557
0.0089
0.0002
0.0234
building
0.0126
0.9788
0.0018
0.0034
0.0002
0.0031
low_veg
0.0730
0.0192
0.8033
0.0782
0.0000
0.0263
tree
0.0252
0.0061
0.0983
0.8623
0.0014
0.0067
car
0.0452
0.0125
0.0005
0.0173
0.8943
0.0301
clutter
0.1369
0.0641
0.0618
0.0110
0.0010
0.7253
Precision/Correctness
0.854
0.927
0.821
0.882
0.973
0.836
Recall/Completeness
0.895
0.979
0.803
0.862
0.894
0.725
F1
0.874
0.952
0.812
0.872
0.932
0.777

Overall accuracy 0.869

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
0.9224
0.0249
0.0354
0.0074
0.0023
0.0075
building
0.0101
0.9821
0.0043
0.0017
0.0000
0.0017
low_veg
0.0627
0.0294
0.8428
0.0599
0.0002
0.0049
tree
0.0784
0.0114
0.1224
0.7770
0.0059
0.0049
car
0.1337
0.0216
0.0006
0.0094
0.8051
0.0296
clutter
0.3253
0.3056
0.1162
0.0258
0.0068
0.2203
Precision/Correctness
0.897
0.927
0.809
0.878
0.915
0.549
Recall/Completeness
0.922
0.982
0.843
0.777
0.805
0.220
F1
0.910
0.954
0.826
0.824
0.856
0.314

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_15_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9413
0.0175
0.0299
0.0046
0.0003
0.0064
building
0.0076
0.9871
0.0027
0.0011
0.0000
0.0014
low_veg
0.0391
0.0178
0.8885
0.0515
0.0000
0.0031
tree
0.0614
0.0095
0.1033
0.8175
0.0045
0.0038
car
0.0406
0.0245
0.0000
0.0042
0.9014
0.0293
clutter
0.3296
0.3189
0.0917
0.0197
0.0063
0.2337
Precision/Correctness
0.928
0.947
0.840
0.908
0.954
0.571
Recall/Completeness
0.941
0.987
0.889
0.818
0.901
0.234
F1
0.935
0.967
0.864
0.860
0.927
0.332

Overall accuracy 0.914

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
0.9163
0.0209
0.0345
0.0208
0.0025
0.0050
building
0.0146
0.9737
0.0046
0.0029
0.0000
0.0042
low_veg
0.0519
0.0206
0.8447
0.0768
0.0000
0.0060
tree
0.0501
0.0060
0.1111
0.8279
0.0022
0.0027
car
0.1294
0.0029
0.0002
0.0074
0.8404
0.0197
clutter
0.2645
0.3172
0.1318
0.0268
0.0300
0.2296
Precision/Correctness
0.905
0.927
0.840
0.813
0.912
0.604
Recall/Completeness
0.916
0.974
0.845
0.828
0.840
0.230
F1
0.911
0.950
0.842
0.821
0.875
0.333

Overall accuracy 0.884

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
0.9375
0.0135
0.0294
0.0155
0.0004
0.0036
building
0.0117
0.9789
0.0031
0.0024
0.0000
0.0039
low_veg
0.0453
0.0148
0.8684
0.0672
0.0000
0.0044
tree
0.0350
0.0041
0.0967
0.8604
0.0018
0.0021
car
0.0348
0.0030
0.0000
0.0051
0.9360
0.0211
clutter
0.2388
0.3144
0.1218
0.0238
0.0347
0.2665
Precision/Correctness
0.931
0.948
0.867
0.843
0.942
0.631
Recall/Completeness
0.938
0.979
0.868
0.860
0.936
0.266
F1
0.934
0.963
0.868
0.852
0.939
0.375

Overall accuracy 0.910

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
0.9244
0.0112
0.0247
0.0255
0.0017
0.0124
building
0.0109
0.9817
0.0042
0.0017
0.0000
0.0015
low_veg
0.0754
0.0165
0.7918
0.0990
0.0001
0.0172
tree
0.0460
0.0071
0.0736
0.8579
0.0031
0.0123
car
0.1330
0.0241
0.0009
0.0175
0.8019
0.0226
clutter
0.1871
0.2435
0.1026
0.0534
0.0058
0.4075
Precision/Correctness
0.853
0.904
0.853
0.824
0.914
0.754
Recall/Completeness
0.924
0.982
0.792
0.858
0.802
0.408
F1
0.887
0.941
0.821
0.841
0.854
0.529

Overall accuracy 0.859

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
0.9437
0.0074
0.0183
0.0195
0.0002
0.0109
building
0.0091
0.9857
0.0028
0.0013
0.0000
0.0011
low_veg
0.0579
0.0117
0.8216
0.0923
0.0000
0.0165
tree
0.0352
0.0059
0.0655
0.8804
0.0022
0.0109
car
0.0379
0.0258
0.0004
0.0105
0.9039
0.0214
clutter
0.1696
0.2209
0.0896
0.0517
0.0058
0.4623
Precision/Correctness
0.885
0.926
0.878
0.848
0.944
0.780
Recall/Completeness
0.944
0.986
0.822
0.880
0.904
0.462
F1
0.914
0.955
0.849
0.864
0.923
0.581

Overall accuracy 0.885

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
0.9136
0.0145
0.0374
0.0212
0.0018
0.0116
building
0.0124
0.9747
0.0053
0.0033
0.0004
0.0038
low_veg
0.0518
0.0146
0.7607
0.1498
0.0000
0.0231
tree
0.0364
0.0112
0.0581
0.8894
0.0019
0.0031
car
0.1371
0.0072
0.0008
0.0166
0.8034
0.0348
clutter
0.2184
0.1478
0.0863
0.0209
0.0032
0.5235
Precision/Correctness
0.916
0.924
0.789
0.760
0.944
0.784
Recall/Completeness
0.914
0.975
0.761
0.889
0.803
0.523
F1
0.915
0.949
0.775
0.820
0.868
0.628

Overall accuracy 0.874

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
0.9313
0.0087
0.0331
0.0161
0.0005
0.0104
building
0.0088
0.9812
0.0041
0.0026
0.0003
0.0031
low_veg
0.0317
0.0087
0.7946
0.1430
0.0000
0.0219
tree
0.0257
0.0086
0.0500
0.9116
0.0015
0.0027
car
0.0419
0.0066
0.0001
0.0127
0.9026
0.0361
clutter
0.2094
0.1113
0.0764
0.0180
0.0033
0.5816
Precision/Correctness
0.939
0.951
0.813
0.790
0.968
0.807
Recall/Completeness
0.931
0.981
0.795
0.912
0.903
0.582
F1
0.935
0.966
0.804
0.846
0.934
0.676

Overall accuracy 0.900

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
0.9391
0.0197
0.0204
0.0162
0.0015
0.0030
building
0.0597
0.9179
0.0187
0.0012
0.0002
0.0023
low_veg
0.0243
0.0127
0.9196
0.0429
0.0000
0.0004
tree
0.0782
0.0023
0.1096
0.8067
0.0013
0.0019
car
0.1088
0.0124
0.0002
0.0040
0.8140
0.0606
clutter
0.4141
0.1040
0.1065
0.0420
0.0074
0.3260
Precision/Correctness
0.900
0.944
0.888
0.846
0.917
0.706
Recall/Completeness
0.939
0.918
0.920
0.807
0.814
0.326
F1
0.919
0.931
0.904
0.826
0.862
0.446

Overall accuracy 0.900

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
0.9554
0.0159
0.0156
0.0107
0.0002
0.0021
building
0.0556
0.9256
0.0158
0.0008
0.0002
0.0019
low_veg
0.0141
0.0089
0.9401
0.0366
0.0000
0.0002
tree
0.0580
0.0015
0.0981
0.8399
0.0010
0.0016
car
0.0324
0.0159
0.0000
0.0021
0.8853
0.0642
clutter
0.4133
0.0971
0.0801
0.0385
0.0062
0.3649
Precision/Correctness
0.919
0.958
0.907
0.877
0.962
0.755
Recall/Completeness
0.955
0.926
0.940
0.840
0.885
0.365
F1
0.937
0.941
0.923
0.858
0.922
0.492

Overall accuracy 0.920

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
0.9447
0.0152
0.0230
0.0109
0.0043
0.0020
building
0.0498
0.9396
0.0049
0.0010
0.0002
0.0045
low_veg
0.0589
0.0251
0.8341
0.0781
0.0001
0.0038
tree
0.0481
0.0087
0.1039
0.8334
0.0037
0.0022
car
0.1000
0.0157
0.0005
0.0131
0.8485
0.0222
clutter
0.3545
0.2293
0.1296
0.0324
0.0171
0.2371
Precision/Correctness
0.902
0.945
0.846
0.817
0.914
0.603
Recall/Completeness
0.945
0.940
0.834
0.833
0.848
0.237
F1
0.923
0.942
0.840
0.825
0.880
0.340

Overall accuracy 0.896

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
0.9623
0.0104
0.0177
0.0080
0.0004
0.0013
building
0.0454
0.9468
0.0030
0.0007
0.0001
0.0040
low_veg
0.0394
0.0184
0.8689
0.0709
0.0000
0.0024
tree
0.0371
0.0074
0.0917
0.8593
0.0026
0.0019
car
0.0273
0.0152
0.0000
0.0085
0.9277
0.0212
clutter
0.3483
0.2433
0.1082
0.0281
0.0172
0.2550
Precision/Correctness
0.923
0.958
0.873
0.847
0.967
0.652
Recall/Completeness
0.962
0.947
0.869
0.859
0.928
0.255
F1
0.942
0.953
0.871
0.853
0.947
0.367

Overall accuracy 0.918

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
0.8967
0.0310
0.0334
0.0155
0.0048
0.0186
building
0.0275
0.9591
0.0072
0.0032
0.0000
0.0029
low_veg
0.0508
0.0988
0.7243
0.1049
0.0003
0.0209
tree
0.0308
0.0139
0.1236
0.8227
0.0036
0.0054
car
0.0994
0.0224
0.0006
0.0159
0.8354
0.0263
clutter
0.1184
0.1136
0.0800
0.0188
0.0082
0.6610
Precision/Correctness
0.922
0.860
0.737
0.843
0.897
0.761
Recall/Completeness
0.897
0.959
0.724
0.823
0.835
0.661
F1
0.909
0.907
0.731
0.833
0.865
0.707

Overall accuracy 0.858

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
0.9186
0.0254
0.0271
0.0110
0.0007
0.0171
building
0.0243
0.9661
0.0049
0.0025
0.0000
0.0022
low_veg
0.0304
0.0997
0.7566
0.0954
0.0001
0.0179
tree
0.0218
0.0115
0.1103
0.8493
0.0028
0.0044
car
0.0262
0.0236
0.0001
0.0097
0.9180
0.0223
clutter
0.1104
0.1091
0.0686
0.0153
0.0080
0.6886
Precision/Correctness
0.944
0.878
0.764
0.873
0.950
0.792
Recall/Completeness
0.919
0.966
0.757
0.849
0.918
0.689
F1
0.931
0.920
0.760
0.861
0.934
0.737

Overall accuracy 0.883

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
0.9242
0.0181
0.0104
0.0230
0.0016
0.0227
building
0.0833
0.8977
0.0044
0.0015
0.0002
0.0130
low_veg
0.0590
0.0371
0.7314
0.1530
0.0001
0.0194
tree
0.0319
0.0030
0.1292
0.8291
0.0008
0.0060
car
0.1037
0.0200
0.0008
0.0289
0.8308
0.0159
clutter
0.1409
0.0200
0.0233
0.0029
0.0053
0.8077
Precision/Correctness
0.917
0.930
0.802
0.699
0.891
0.813
Recall/Completeness
0.924
0.898
0.731
0.829
0.831
0.808
F1
0.921
0.914
0.765
0.758
0.860
0.810

Overall accuracy 0.880

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
0.9384
0.0145
0.0078
0.0166
0.0004
0.0222
building
0.0802
0.9033
0.0032
0.0010
0.0001
0.0122
low_veg
0.0440
0.0350
0.7577
0.1448
0.0000
0.0185
tree
0.0195
0.0018
0.1221
0.8516
0.0005
0.0044
car
0.0274
0.0206
0.0004
0.0206
0.9150
0.0160
clutter
0.1304
0.0181
0.0216
0.0024
0.0048
0.8228
Precision/Correctness
0.928
0.942
0.824
0.732
0.932
0.824
Recall/Completeness
0.938
0.903
0.758
0.852
0.915
0.823
F1
0.933
0.922
0.790
0.787
0.923
0.823

Overall accuracy 0.896

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image