Detailled semantic labeling (2D) result


Name Y. Sun
Affiliation Huazhong University of Science and Technology (HUST), Wuhan, China
Abbreviation HUSTW3
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
0.907
0.015
0.036
0.024
0.004
0.014
building
0.016
0.971
0.006
0.003
0.000
0.005
low_veg
0.031
0.020
0.860
0.080
0.000
0.009
tree
0.023
0.005
0.088
0.880
0.001
0.003
car
0.069
0.004
0.002
0.022
0.891
0.011
clutter
0.169
0.118
0.110
0.033
0.006
0.564
Precision/Correctness
0.925
0.939
0.846
0.847
0.910
0.758
Recall/Completeness
0.907
0.971
0.860
0.880
0.891
0.564
F1
0.916
0.954
0.853
0.863
0.900
0.647

Overall accuracy 0.892

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
0.931
0.010
0.030
0.017
0.000
0.012
building
0.012
0.978
0.003
0.002
0.000
0.004
low_veg
0.020
0.015
0.888
0.069
0.000
0.007
tree
0.015
0.003
0.074
0.904
0.001
0.002
car
0.011
0.004
0.000
0.016
0.957
0.012
clutter
0.158
0.101
0.093
0.027
0.006
0.615
Precision/Correctness
0.946
0.956
0.872
0.877
0.963
0.789
Recall/Completeness
0.931
0.978
0.888
0.904
0.957
0.615
F1
0.938
0.967
0.880
0.890
0.960
0.692

Overall accuracy 0.916
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8904
0.0159
0.0541
0.0243
0.0035
0.0119
building
0.0094
0.9712
0.0087
0.0068
0.0000
0.0039
low_veg
0.0221
0.0102
0.9016
0.0587
0.0001
0.0073
tree
0.0194
0.0080
0.0963
0.8732
0.0007
0.0024
car
0.0698
0.0043
0.0024
0.0201
0.8778
0.0256
clutter
0.1076
0.1733
0.1477
0.0460
0.0194
0.5059
Precision/Correctness
0.933
0.936
0.872
0.886
0.878
0.592
Recall/Completeness
0.890
0.971
0.902
0.873
0.878
0.506
F1
0.911
0.953
0.887
0.879
0.878
0.546

Overall accuracy 0.897

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
0.9213
0.0102
0.0438
0.0159
0.0004
0.0083
building
0.0058
0.9814
0.0045
0.0053
0.0000
0.0030
low_veg
0.0160
0.0057
0.9290
0.0441
0.0000
0.0052
tree
0.0131
0.0062
0.0784
0.9001
0.0004
0.0018
car
0.0189
0.0046
0.0010
0.0113
0.9362
0.0280
clutter
0.0779
0.1650
0.1076
0.0341
0.0235
0.5920
Precision/Correctness
0.956
0.959
0.899
0.918
0.942
0.634
Recall/Completeness
0.921
0.981
0.929
0.900
0.936
0.592
F1
0.938
0.970
0.914
0.909
0.939
0.612

Overall accuracy 0.926

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
0.8861
0.0071
0.0727
0.0248
0.0024
0.0069
building
0.0039
0.9741
0.0088
0.0088
0.0000
0.0044
low_veg
0.0122
0.0066
0.9138
0.0662
0.0001
0.0012
tree
0.0093
0.0027
0.0819
0.9048
0.0005
0.0007
car
0.0520
0.0011
0.0008
0.0213
0.9212
0.0036
clutter
0.0888
0.3051
0.1477
0.0779
0.0042
0.3763
Precision/Correctness
0.937
0.919
0.901
0.909
0.906
0.389
Recall/Completeness
0.886
0.974
0.914
0.905
0.921
0.376
F1
0.911
0.946
0.907
0.907
0.914
0.382

Overall accuracy 0.909

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
0.9174
0.0034
0.0567
0.0160
0.0002
0.0062
building
0.0017
0.9841
0.0046
0.0057
0.0000
0.0039
low_veg
0.0070
0.0045
0.9337
0.0541
0.0000
0.0008
tree
0.0063
0.0019
0.0684
0.9226
0.0003
0.0005
car
0.0031
0.0006
0.0000
0.0139
0.9792
0.0032
clutter
0.0565
0.3471
0.1097
0.0678
0.0038
0.4151
Precision/Correctness
0.962
0.946
0.920
0.929
0.970
0.379
Recall/Completeness
0.917
0.984
0.934
0.923
0.979
0.415
F1
0.939
0.964
0.927
0.926
0.974
0.396

Overall accuracy 0.930

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
0.8826
0.0206
0.0543
0.0290
0.0046
0.0088
building
0.0117
0.9790
0.0035
0.0034
0.0000
0.0024
low_veg
0.0458
0.0246
0.8467
0.0750
0.0001
0.0078
tree
0.0168
0.0062
0.0775
0.8950
0.0014
0.0031
car
0.0673
0.0062
0.0009
0.0278
0.8876
0.0101
clutter
0.2437
0.1503
0.2491
0.0350
0.0132
0.3088
Precision/Correctness
0.901
0.938
0.836
0.885
0.893
0.571
Recall/Completeness
0.883
0.979
0.847
0.895
0.888
0.309
F1
0.892
0.958
0.841
0.890
0.890
0.401

Overall accuracy 0.887

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
0.9152
0.0132
0.0448
0.0194
0.0004
0.0070
building
0.0090
0.9859
0.0011
0.0022
0.0000
0.0018
low_veg
0.0356
0.0142
0.8834
0.0613
0.0000
0.0054
tree
0.0111
0.0045
0.0621
0.9188
0.0009
0.0027
car
0.0091
0.0071
0.0004
0.0192
0.9544
0.0098
clutter
0.2557
0.1399
0.2248
0.0234
0.0156
0.3407
Precision/Correctness
0.928
0.962
0.868
0.916
0.957
0.599
Recall/Completeness
0.915
0.986
0.883
0.919
0.954
0.341
F1
0.921
0.974
0.876
0.917
0.956
0.434

Overall accuracy 0.917

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
0.8723
0.0183
0.0580
0.0330
0.0046
0.0139
building
0.0115
0.9782
0.0044
0.0032
0.0000
0.0027
low_veg
0.0205
0.0417
0.8556
0.0764
0.0000
0.0057
tree
0.0143
0.0046
0.0956
0.8825
0.0008
0.0021
car
0.0672
0.0044
0.0007
0.0427
0.8781
0.0070
clutter
0.2719
0.1526
0.1741
0.0386
0.0028
0.3601
Precision/Correctness
0.898
0.911
0.843
0.883
0.892
0.657
Recall/Completeness
0.872
0.978
0.856
0.883
0.878
0.360
F1
0.885
0.943
0.849
0.883
0.885
0.465

Overall accuracy 0.878

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
0.9077
0.0109
0.0454
0.0238
0.0006
0.0117
building
0.0096
0.9843
0.0020
0.0022
0.0000
0.0020
low_veg
0.0115
0.0379
0.8850
0.0622
0.0000
0.0033
tree
0.0103
0.0034
0.0802
0.9040
0.0006
0.0015
car
0.0144
0.0041
0.0000
0.0310
0.9445
0.0060
clutter
0.2963
0.1323
0.1476
0.0300
0.0030
0.3907
Precision/Correctness
0.923
0.931
0.872
0.912
0.961
0.704
Recall/Completeness
0.908
0.984
0.885
0.904
0.944
0.391
F1
0.915
0.957
0.878
0.908
0.953
0.502

Overall accuracy 0.905

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
0.9156
0.0099
0.0324
0.0326
0.0030
0.0064
building
0.0214
0.9665
0.0053
0.0042
0.0001
0.0025
low_veg
0.0395
0.0212
0.8112
0.1211
0.0003
0.0067
tree
0.0175
0.0033
0.0435
0.9341
0.0008
0.0008
car
0.0693
0.0016
0.0027
0.0359
0.8854
0.0052
clutter
0.3800
0.1208
0.1756
0.0770
0.0067
0.2399
Precision/Correctness
0.868
0.956
0.802
0.774
0.927
0.804
Recall/Completeness
0.916
0.967
0.811
0.934
0.885
0.240
F1
0.891
0.961
0.806
0.847
0.906
0.370

Overall accuracy 0.873

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
0.9393
0.0057
0.0258
0.0240
0.0004
0.0048
building
0.0163
0.9761
0.0024
0.0030
0.0000
0.0020
low_veg
0.0270
0.0148
0.8462
0.1073
0.0000
0.0047
tree
0.0098
0.0015
0.0328
0.9550
0.0005
0.0003
car
0.0119
0.0009
0.0007
0.0253
0.9567
0.0045
clutter
0.3888
0.1017
0.1697
0.0663
0.0072
0.2664
Precision/Correctness
0.892
0.971
0.835
0.814
0.964
0.846
Recall/Completeness
0.939
0.976
0.846
0.955
0.957
0.266
F1
0.915
0.974
0.841
0.879
0.960
0.405

Overall accuracy 0.901

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
0.8730
0.0178
0.0603
0.0212
0.0051
0.0227
building
0.0132
0.9721
0.0028
0.0058
0.0003
0.0057
low_veg
0.0496
0.0195
0.8230
0.0859
0.0004
0.0215
tree
0.0193
0.0054
0.0917
0.8740
0.0013
0.0083
car
0.0600
0.0034
0.0013
0.0341
0.8793
0.0219
clutter
0.1588
0.0483
0.0817
0.0173
0.0014
0.6925
Precision/Correctness
0.865
0.929
0.821
0.851
0.916
0.831
Recall/Completeness
0.873
0.972
0.823
0.874
0.879
0.693
F1
0.869
0.950
0.822
0.862
0.897
0.755

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_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8982
0.0127
0.0532
0.0146
0.0006
0.0207
building
0.0103
0.9789
0.0011
0.0046
0.0003
0.0049
low_veg
0.0371
0.0131
0.8567
0.0748
0.0001
0.0182
tree
0.0127
0.0037
0.0751
0.9007
0.0008
0.0069
car
0.0156
0.0030
0.0007
0.0271
0.9316
0.0220
clutter
0.1570
0.0430
0.0687
0.0126
0.0009
0.7177
Precision/Correctness
0.886
0.949
0.847
0.881
0.973
0.857
Recall/Completeness
0.898
0.979
0.857
0.901
0.932
0.718
F1
0.892
0.963
0.852
0.891
0.952
0.781

Overall accuracy 0.889

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
0.9181
0.0154
0.0400
0.0112
0.0035
0.0118
building
0.0078
0.9828
0.0050
0.0025
0.0000
0.0019
low_veg
0.0432
0.0253
0.8688
0.0540
0.0007
0.0081
tree
0.0469
0.0076
0.1014
0.8365
0.0044
0.0032
car
0.0710
0.0040
0.0014
0.0134
0.9071
0.0030
clutter
0.3080
0.2248
0.0998
0.0318
0.0077
0.3280
Precision/Correctness
0.923
0.947
0.821
0.880
0.911
0.596
Recall/Completeness
0.918
0.983
0.869
0.836
0.907
0.328
F1
0.921
0.965
0.844
0.858
0.909
0.423

Overall accuracy 0.902

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
0.9385
0.0097
0.0343
0.0073
0.0004
0.0099
building
0.0052
0.9884
0.0031
0.0017
0.0000
0.0016
low_veg
0.0251
0.0152
0.9097
0.0445
0.0001
0.0055
tree
0.0326
0.0057
0.0816
0.8747
0.0030
0.0025
car
0.0088
0.0034
0.0001
0.0072
0.9785
0.0021
clutter
0.3252
0.2224
0.0747
0.0237
0.0065
0.3476
Precision/Correctness
0.948
0.965
0.851
0.912
0.965
0.618
Recall/Completeness
0.939
0.988
0.910
0.875
0.978
0.348
F1
0.943
0.976
0.879
0.893
0.972
0.445

Overall accuracy 0.929

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
0.9103
0.0141
0.0307
0.0296
0.0028
0.0124
building
0.0114
0.9788
0.0035
0.0041
0.0000
0.0021
low_veg
0.0430
0.0363
0.8238
0.0880
0.0001
0.0087
tree
0.0322
0.0045
0.1009
0.8597
0.0014
0.0012
car
0.0751
0.0003
0.0002
0.0111
0.8892
0.0241
clutter
0.1979
0.2376
0.1188
0.0326
0.0062
0.4069
Precision/Correctness
0.927
0.933
0.851
0.785
0.944
0.646
Recall/Completeness
0.910
0.979
0.824
0.860
0.889
0.407
F1
0.919
0.956
0.837
0.820
0.916
0.499

Overall accuracy 0.890

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
0.9339
0.0082
0.0262
0.0226
0.0003
0.0088
building
0.0078
0.9848
0.0024
0.0032
0.0000
0.0018
low_veg
0.0374
0.0320
0.8468
0.0776
0.0000
0.0061
tree
0.0200
0.0026
0.0881
0.8874
0.0011
0.0008
car
0.0099
0.0001
0.0000
0.0081
0.9557
0.0261
clutter
0.1744
0.2181
0.1112
0.0296
0.0063
0.4604
Precision/Correctness
0.950
0.952
0.876
0.816
0.981
0.686
Recall/Completeness
0.934
0.985
0.847
0.887
0.956
0.460
F1
0.942
0.968
0.861
0.850
0.968
0.551

Overall accuracy 0.915

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
0.9014
0.0086
0.0262
0.0377
0.0028
0.0232
building
0.0116
0.9800
0.0038
0.0026
0.0001
0.0018
low_veg
0.0382
0.0200
0.8225
0.1001
0.0003
0.0191
tree
0.0246
0.0034
0.0655
0.8986
0.0018
0.0060
car
0.0741
0.0148
0.0038
0.0271
0.8700
0.0103
clutter
0.1153
0.2031
0.1102
0.0593
0.0044
0.5077
Precision/Correctness
0.906
0.916
0.860
0.812
0.921
0.760
Recall/Completeness
0.901
0.980
0.822
0.899
0.870
0.508
F1
0.904
0.947
0.841
0.853
0.895
0.609

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_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9259
0.0053
0.0188
0.0291
0.0003
0.0207
building
0.0100
0.9844
0.0022
0.0019
0.0001
0.0014
low_veg
0.0258
0.0167
0.8487
0.0920
0.0000
0.0167
tree
0.0172
0.0023
0.0576
0.9166
0.0012
0.0050
car
0.0085
0.0167
0.0018
0.0182
0.9448
0.0100
clutter
0.0971
0.1770
0.0970
0.0555
0.0042
0.5692
Precision/Correctness
0.933
0.937
0.885
0.840
0.963
0.788
Recall/Completeness
0.926
0.984
0.849
0.917
0.945
0.569
F1
0.929
0.960
0.867
0.877
0.954
0.661

Overall accuracy 0.900

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
0.9020
0.0107
0.0422
0.0252
0.0025
0.0173
building
0.0113
0.9736
0.0053
0.0047
0.0002
0.0048
low_veg
0.0293
0.0093
0.7895
0.1483
0.0002
0.0235
tree
0.0224
0.0055
0.0516
0.9171
0.0012
0.0021
car
0.0953
0.0007
0.0020
0.0206
0.8689
0.0126
clutter
0.0916
0.1147
0.1003
0.0243
0.0023
0.6668
Precision/Correctness
0.951
0.944
0.783
0.754
0.943
0.795
Recall/Completeness
0.902
0.974
0.789
0.917
0.869
0.667
F1
0.926
0.959
0.786
0.828
0.904
0.725

Overall accuracy 0.888

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
0.9212
0.0064
0.0375
0.0194
0.0005
0.0150
building
0.0076
0.9808
0.0040
0.0035
0.0002
0.0039
low_veg
0.0146
0.0049
0.8186
0.1403
0.0000
0.0215
tree
0.0142
0.0033
0.0432
0.9366
0.0008
0.0018
car
0.0189
0.0004
0.0002
0.0175
0.9497
0.0134
clutter
0.0726
0.0808
0.0916
0.0210
0.0023
0.7316
Precision/Correctness
0.972
0.966
0.806
0.785
0.975
0.820
Recall/Completeness
0.921
0.981
0.819
0.937
0.950
0.732
F1
0.946
0.973
0.812
0.854
0.962
0.773

Overall accuracy 0.912

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
0.9407
0.0102
0.0207
0.0200
0.0021
0.0062
building
0.0305
0.9406
0.0160
0.0011
0.0009
0.0110
low_veg
0.0184
0.0208
0.9198
0.0398
0.0000
0.0012
tree
0.0497
0.0025
0.0845
0.8602
0.0008
0.0024
car
0.0601
0.0007
0.0000
0.0076
0.9251
0.0065
clutter
0.2679
0.0522
0.0892
0.0368
0.0054
0.5485
Precision/Correctness
0.937
0.957
0.902
0.850
0.903
0.666
Recall/Completeness
0.941
0.941
0.920
0.860
0.925
0.549
F1
0.939
0.949
0.911
0.855
0.914
0.601

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_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9586
0.0069
0.0161
0.0134
0.0004
0.0047
building
0.0260
0.9490
0.0129
0.0005
0.0009
0.0106
low_veg
0.0096
0.0186
0.9381
0.0331
0.0000
0.0006
tree
0.0329
0.0015
0.0729
0.8903
0.0006
0.0018
car
0.0054
0.0005
0.0000
0.0043
0.9831
0.0066
clutter
0.2588
0.0475
0.0619
0.0328
0.0040
0.5949
Precision/Correctness
0.954
0.967
0.921
0.883
0.952
0.692
Recall/Completeness
0.959
0.949
0.938
0.890
0.983
0.595
F1
0.956
0.958
0.930
0.887
0.967
0.640

Overall accuracy 0.937

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
0.9443
0.0114
0.0230
0.0111
0.0059
0.0044
building
0.0273
0.9582
0.0037
0.0015
0.0003
0.0090
low_veg
0.0377
0.0150
0.8705
0.0720
0.0006
0.0042
tree
0.0431
0.0080
0.1006
0.8436
0.0037
0.0009
car
0.0607
0.0039
0.0014
0.0121
0.9140
0.0078
clutter
0.2022
0.1554
0.1895
0.0470
0.0100
0.3959
Precision/Correctness
0.937
0.962
0.848
0.822
0.903
0.629
Recall/Completeness
0.944
0.958
0.871
0.844
0.914
0.396
F1
0.941
0.960
0.859
0.833
0.909
0.486

Overall accuracy 0.914

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
0.9631
0.0073
0.0174
0.0081
0.0005
0.0036
building
0.0231
0.9652
0.0018
0.0010
0.0001
0.0087
low_veg
0.0219
0.0083
0.9018
0.0653
0.0001
0.0026
tree
0.0323
0.0065
0.0887
0.8692
0.0025
0.0007
car
0.0087
0.0032
0.0003
0.0077
0.9723
0.0079
clutter
0.1818
0.1616
0.1763
0.0469
0.0093
0.4241
Precision/Correctness
0.956
0.974
0.873
0.851
0.972
0.648
Recall/Completeness
0.963
0.965
0.902
0.869
0.972
0.424
F1
0.959
0.969
0.887
0.860
0.972
0.513

Overall accuracy 0.935

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
0.9004
0.0114
0.0378
0.0159
0.0056
0.0288
building
0.0120
0.9745
0.0053
0.0029
0.0001
0.0052
low_veg
0.0283
0.0201
0.8613
0.0701
0.0010
0.0191
tree
0.0285
0.0047
0.1370
0.8237
0.0028
0.0034
car
0.0693
0.0093
0.0037
0.0190
0.8898
0.0090
clutter
0.0375
0.0518
0.0998
0.0117
0.0075
0.7918
Precision/Correctness
0.956
0.952
0.749
0.875
0.896
0.751
Recall/Completeness
0.900
0.974
0.861
0.824
0.890
0.792
F1
0.927
0.963
0.801
0.848
0.893
0.771

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_15_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9238
0.0074
0.0295
0.0114
0.0007
0.0272
building
0.0084
0.9819
0.0030
0.0020
0.0001
0.0047
low_veg
0.0141
0.0165
0.8932
0.0598
0.0002
0.0163
tree
0.0206
0.0030
0.1213
0.8504
0.0020
0.0027
car
0.0125
0.0094
0.0007
0.0134
0.9548
0.0093
clutter
0.0304
0.0483
0.0851
0.0093
0.0068
0.8200
Precision/Correctness
0.974
0.965
0.777
0.904
0.959
0.772
Recall/Completeness
0.924
0.982
0.893
0.850
0.955
0.820
F1
0.948
0.973
0.831
0.876
0.957
0.795

Overall accuracy 0.917

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
0.9115
0.0266
0.0124
0.0273
0.0022
0.0200
building
0.0231
0.9624
0.0051
0.0021
0.0004
0.0068
low_veg
0.0336
0.0154
0.7768
0.1608
0.0002
0.0132
tree
0.0226
0.0043
0.1230
0.8451
0.0007
0.0043
car
0.0578
0.0052
0.0004
0.0296
0.8967
0.0103
clutter
0.0877
0.0603
0.0216
0.0020
0.0091
0.8193
Precision/Correctness
0.960
0.913
0.807
0.681
0.855
0.848
Recall/Completeness
0.912
0.962
0.777
0.845
0.897
0.819
F1
0.935
0.937
0.791
0.754
0.875
0.833

Overall accuracy 0.896

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
0.9270
0.0236
0.0096
0.0203
0.0005
0.0191
building
0.0196
0.9689
0.0039
0.0013
0.0003
0.0060
low_veg
0.0218
0.0107
0.8028
0.1530
0.0000
0.0117
tree
0.0142
0.0026
0.1156
0.8636
0.0004
0.0036
car
0.0067
0.0059
0.0001
0.0222
0.9564
0.0087
clutter
0.0799
0.0583
0.0198
0.0014
0.0085
0.8322
Precision/Correctness
0.969
0.923
0.829
0.713
0.902
0.860
Recall/Completeness
0.927
0.969
0.803
0.864
0.956
0.832
F1
0.947
0.946
0.815
0.781
0.928
0.846

Overall accuracy 0.912

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image