Detailled semantic labeling (2D) result


Name T. V. Duy
Affiliation Hanoi University of Science and Technology, Vietnam
Abbreviation BKHN_2
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
0.920
0.017
0.035
0.016
0.002
0.010
building
0.012
0.974
0.005
0.004
0.000
0.005
low_veg
0.038
0.010
0.885
0.059
0.000
0.008
tree
0.037
0.006
0.123
0.828
0.003
0.003
car
0.074
0.006
0.003
0.011
0.895
0.012
clutter
0.307
0.115
0.156
0.022
0.009
0.391
Precision/Correctness
0.899
0.943
0.821
0.877
0.921
0.730
Recall/Completeness
0.920
0.974
0.885
0.828
0.895
0.391
F1
0.909
0.958
0.852
0.852
0.908
0.509

Overall accuracy 0.885

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
0.937
0.013
0.030
0.012
0.000
0.009
building
0.009
0.981
0.003
0.003
0.000
0.004
low_veg
0.027
0.005
0.909
0.051
0.000
0.007
tree
0.027
0.005
0.108
0.856
0.002
0.002
car
0.011
0.007
0.001
0.007
0.961
0.013
clutter
0.311
0.099
0.146
0.018
0.010
0.416
Precision/Correctness
0.919
0.959
0.844
0.902
0.948
0.751
Recall/Completeness
0.937
0.981
0.909
0.856
0.961
0.416
F1
0.928
0.970
0.875
0.878
0.954
0.535

Overall accuracy 0.906
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9060
0.0180
0.0435
0.0103
0.0018
0.0204
building
0.0075
0.9559
0.0084
0.0224
0.0000
0.0057
low_veg
0.0289
0.0070
0.9093
0.0396
0.0002
0.0149
tree
0.0407
0.0101
0.1349
0.8074
0.0013
0.0055
car
0.0866
0.0011
0.0008
0.0064
0.8676
0.0375
clutter
0.1184
0.1670
0.1111
0.0213
0.0057
0.5767
Precision/Correctness
0.907
0.936
0.858
0.907
0.917
0.474
Recall/Completeness
0.906
0.956
0.909
0.807
0.868
0.577
F1
0.906
0.946
0.883
0.854
0.892
0.520

Overall accuracy 0.886

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
0.9264
0.0129
0.0355
0.0064
0.0002
0.0186
building
0.0042
0.9647
0.0048
0.0217
0.0000
0.0046
low_veg
0.0214
0.0035
0.9310
0.0302
0.0001
0.0139
tree
0.0320
0.0084
0.1141
0.8401
0.0009
0.0046
car
0.0336
0.0009
0.0001
0.0027
0.9193
0.0435
clutter
0.0787
0.1654
0.0724
0.0138
0.0065
0.6633
Precision/Correctness
0.933
0.956
0.883
0.929
0.959
0.459
Recall/Completeness
0.926
0.965
0.931
0.840
0.919
0.663
F1
0.930
0.960
0.906
0.882
0.939
0.543

Overall accuracy 0.911

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
0.9149
0.0066
0.0654
0.0086
0.0009
0.0036
building
0.0053
0.9796
0.0077
0.0050
0.0000
0.0024
low_veg
0.0121
0.0033
0.9508
0.0311
0.0000
0.0027
tree
0.0211
0.0040
0.1802
0.7929
0.0007
0.0011
car
0.0715
0.0008
0.0018
0.0066
0.9187
0.0006
clutter
0.0793
0.3340
0.1569
0.0596
0.0033
0.3669
Precision/Correctness
0.912
0.931
0.839
0.950
0.927
0.373
Recall/Completeness
0.915
0.980
0.951
0.793
0.919
0.367
F1
0.914
0.955
0.892
0.864
0.923
0.370

Overall accuracy 0.889

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
0.9391
0.0035
0.0502
0.0047
0.0000
0.0025
building
0.0026
0.9888
0.0039
0.0031
0.0000
0.0016
low_veg
0.0069
0.0014
0.9653
0.0243
0.0000
0.0022
tree
0.0169
0.0031
0.1632
0.8154
0.0005
0.0009
car
0.0156
0.0001
0.0001
0.0028
0.9815
0.0000
clutter
0.0374
0.3848
0.1189
0.0718
0.0022
0.3850
Precision/Correctness
0.938
0.955
0.857
0.963
0.960
0.366
Recall/Completeness
0.939
0.989
0.965
0.815
0.981
0.385
F1
0.938
0.972
0.908
0.883
0.971
0.375

Overall accuracy 0.908

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
0.9034
0.0202
0.0450
0.0170
0.0024
0.0120
building
0.0085
0.9814
0.0038
0.0036
0.0001
0.0026
low_veg
0.0482
0.0203
0.8679
0.0568
0.0001
0.0066
tree
0.0322
0.0065
0.0939
0.8612
0.0025
0.0037
car
0.0673
0.0042
0.0009
0.0151
0.9060
0.0065
clutter
0.3012
0.1495
0.2165
0.0192
0.0058
0.3078
Precision/Correctness
0.887
0.942
0.837
0.913
0.920
0.543
Recall/Completeness
0.903
0.981
0.868
0.861
0.906
0.308
F1
0.895
0.961
0.852
0.886
0.913
0.393

Overall accuracy 0.891

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
0.9281
0.0134
0.0368
0.0108
0.0001
0.0109
building
0.0060
0.9883
0.0011
0.0025
0.0001
0.0020
low_veg
0.0370
0.0113
0.9013
0.0457
0.0000
0.0047
tree
0.0240
0.0049
0.0774
0.8886
0.0019
0.0032
car
0.0055
0.0046
0.0003
0.0098
0.9726
0.0072
clutter
0.3200
0.1395
0.1893
0.0121
0.0067
0.3323
Precision/Correctness
0.915
0.964
0.867
0.938
0.957
0.544
Recall/Completeness
0.928
0.988
0.901
0.889
0.973
0.332
F1
0.921
0.976
0.884
0.912
0.965
0.412

Overall accuracy 0.918

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
0.8887
0.0187
0.0621
0.0195
0.0026
0.0085
building
0.0095
0.9821
0.0036
0.0024
0.0000
0.0024
low_veg
0.0282
0.0096
0.9065
0.0499
0.0000
0.0057
tree
0.0250
0.0061
0.1763
0.7875
0.0015
0.0035
car
0.0627
0.0216
0.0020
0.0204
0.8840
0.0093
clutter
0.2098
0.1605
0.1469
0.0298
0.0069
0.4461
Precision/Correctness
0.889
0.942
0.794
0.913
0.901
0.734
Recall/Completeness
0.889
0.982
0.907
0.788
0.884
0.446
F1
0.889
0.962
0.846
0.846
0.892
0.555

Overall accuracy 0.875

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
0.9167
0.0122
0.0512
0.0131
0.0002
0.0066
building
0.0077
0.9881
0.0011
0.0016
0.0000
0.0016
low_veg
0.0183
0.0042
0.9344
0.0393
0.0000
0.0039
tree
0.0198
0.0049
0.1608
0.8104
0.0013
0.0028
car
0.0079
0.0247
0.0004
0.0128
0.9445
0.0098
clutter
0.2026
0.1521
0.1187
0.0236
0.0084
0.4947
Precision/Correctness
0.917
0.962
0.816
0.937
0.939
0.778
Recall/Completeness
0.917
0.988
0.934
0.810
0.944
0.495
F1
0.917
0.975
0.871
0.869
0.942
0.605

Overall accuracy 0.899

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
0.9351
0.0103
0.0279
0.0227
0.0011
0.0029
building
0.0084
0.9789
0.0054
0.0045
0.0001
0.0027
low_veg
0.0414
0.0109
0.8285
0.1145
0.0001
0.0046
tree
0.0320
0.0033
0.0407
0.9216
0.0020
0.0004
car
0.0676
0.0022
0.0023
0.0173
0.9079
0.0026
clutter
0.4735
0.1204
0.2032
0.0651
0.0134
0.1244
Precision/Correctness
0.859
0.961
0.804
0.799
0.928
0.767
Recall/Completeness
0.935
0.979
0.829
0.922
0.908
0.124
F1
0.896
0.970
0.816
0.856
0.918
0.214

Overall accuracy 0.877

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
0.9520
0.0068
0.0224
0.0166
0.0001
0.0021
building
0.0040
0.9881
0.0022
0.0033
0.0001
0.0023
low_veg
0.0304
0.0058
0.8553
0.1047
0.0000
0.0038
tree
0.0203
0.0019
0.0280
0.9484
0.0012
0.0002
car
0.0040
0.0016
0.0002
0.0103
0.9812
0.0026
clutter
0.4933
0.1067
0.1989
0.0587
0.0148
0.1277
Precision/Correctness
0.883
0.973
0.836
0.831
0.939
0.792
Recall/Completeness
0.952
0.988
0.855
0.948
0.981
0.128
F1
0.916
0.980
0.846
0.886
0.960
0.220

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
0.8842
0.0193
0.0582
0.0129
0.0025
0.0228
building
0.0063
0.9828
0.0027
0.0046
0.0002
0.0035
low_veg
0.0809
0.0159
0.8138
0.0689
0.0002
0.0203
tree
0.0336
0.0065
0.1250
0.8260
0.0027
0.0061
car
0.0684
0.0068
0.0009
0.0174
0.8903
0.0161
clutter
0.4771
0.0436
0.1400
0.0115
0.0005
0.3272
Precision/Correctness
0.750
0.933
0.780
0.878
0.940
0.718
Recall/Completeness
0.884
0.983
0.814
0.826
0.890
0.327
F1
0.812
0.957
0.797
0.851
0.915
0.450

Overall accuracy 0.822

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
0.8995
0.0143
0.0546
0.0088
0.0003
0.0225
building
0.0040
0.9885
0.0011
0.0035
0.0002
0.0028
low_veg
0.0698
0.0101
0.8398
0.0612
0.0001
0.0190
tree
0.0247
0.0049
0.1075
0.8557
0.0021
0.0052
car
0.0191
0.0067
0.0000
0.0129
0.9435
0.0177
clutter
0.4927
0.0398
0.1293
0.0085
0.0003
0.3293
Precision/Correctness
0.761
0.950
0.799
0.901
0.971
0.737
Recall/Completeness
0.899
0.988
0.840
0.856
0.943
0.329
F1
0.825
0.969
0.819
0.878
0.957
0.455

Overall accuracy 0.840

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
0.9145
0.0218
0.0513
0.0051
0.0022
0.0051
building
0.0040
0.9898
0.0030
0.0026
0.0000
0.0005
low_veg
0.0531
0.0197
0.8927
0.0279
0.0004
0.0062
tree
0.0753
0.0113
0.1368
0.7639
0.0091
0.0037
car
0.0824
0.0128
0.0040
0.0023
0.8815
0.0170
clutter
0.2302
0.2446
0.2216
0.0113
0.0045
0.2878
Precision/Correctness
0.917
0.940
0.775
0.929
0.909
0.686
Recall/Completeness
0.914
0.990
0.893
0.764
0.881
0.288
F1
0.916
0.964
0.830
0.838
0.895
0.406

Overall accuracy 0.896

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
0.9292
0.0158
0.0472
0.0033
0.0006
0.0038
building
0.0019
0.9945
0.0013
0.0019
0.0000
0.0003
low_veg
0.0344
0.0107
0.9280
0.0225
0.0001
0.0044
tree
0.0587
0.0096
0.1140
0.8074
0.0073
0.0029
car
0.0125
0.0153
0.0005
0.0008
0.9508
0.0201
clutter
0.2204
0.2500
0.2390
0.0083
0.0037
0.2786
Precision/Correctness
0.944
0.957
0.798
0.949
0.936
0.700
Recall/Completeness
0.929
0.995
0.928
0.807
0.951
0.279
F1
0.936
0.975
0.858
0.872
0.943
0.399

Overall accuracy 0.920

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
0.9284
0.0134
0.0293
0.0228
0.0013
0.0048
building
0.0068
0.9843
0.0042
0.0033
0.0000
0.0014
low_veg
0.0552
0.0108
0.8513
0.0767
0.0000
0.0059
tree
0.0462
0.0060
0.1082
0.8350
0.0025
0.0021
car
0.0908
0.0044
0.0007
0.0069
0.8823
0.0149
clutter
0.2228
0.2794
0.1193
0.0209
0.0436
0.3139
Precision/Correctness
0.918
0.945
0.852
0.811
0.913
0.728
Recall/Completeness
0.928
0.984
0.851
0.835
0.882
0.314
F1
0.923
0.964
0.852
0.823
0.897
0.439

Overall accuracy 0.897

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
0.9463
0.0082
0.0248
0.0173
0.0001
0.0033
building
0.0032
0.9905
0.0027
0.0026
0.0000
0.0010
low_veg
0.0487
0.0058
0.8735
0.0678
0.0000
0.0043
tree
0.0315
0.0042
0.0944
0.8662
0.0020
0.0017
car
0.0126
0.0051
0.0001
0.0052
0.9604
0.0166
clutter
0.1949
0.2764
0.1100
0.0208
0.0549
0.3430
Precision/Correctness
0.942
0.963
0.878
0.839
0.926
0.756
Recall/Completeness
0.946
0.991
0.873
0.866
0.960
0.343
F1
0.944
0.976
0.876
0.853
0.943
0.472

Overall accuracy 0.921

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
0.9317
0.0077
0.0247
0.0256
0.0008
0.0096
building
0.0155
0.9778
0.0039
0.0017
0.0001
0.0011
low_veg
0.0462
0.0082
0.8454
0.0882
0.0002
0.0118
tree
0.0451
0.0068
0.0719
0.8685
0.0040
0.0037
car
0.0900
0.0123
0.0105
0.0123
0.8619
0.0130
clutter
0.2125
0.2005
0.1649
0.0425
0.0045
0.3751
Precision/Correctness
0.865
0.924
0.843
0.840
0.928
0.818
Recall/Completeness
0.932
0.978
0.845
0.869
0.862
0.375
F1
0.897
0.950
0.844
0.854
0.894
0.514

Overall accuracy 0.871

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
0.9493
0.0047
0.0176
0.0198
0.0000
0.0085
building
0.0138
0.9820
0.0021
0.0012
0.0000
0.0008
low_veg
0.0325
0.0046
0.8695
0.0827
0.0000
0.0107
tree
0.0348
0.0055
0.0624
0.8910
0.0031
0.0031
car
0.0089
0.0142
0.0091
0.0086
0.9453
0.0140
clutter
0.2018
0.1730
0.1589
0.0411
0.0045
0.4207
Precision/Correctness
0.893
0.945
0.867
0.862
0.946
0.839
Recall/Completeness
0.949
0.982
0.870
0.891
0.945
0.421
F1
0.920
0.963
0.868
0.876
0.945
0.560

Overall accuracy 0.895

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
0.9156
0.0092
0.0466
0.0174
0.0010
0.0103
building
0.0103
0.9679
0.0084
0.0054
0.0001
0.0079
low_veg
0.0352
0.0056
0.8285
0.1164
0.0001
0.0142
tree
0.0358
0.0045
0.0767
0.8785
0.0026
0.0019
car
0.1131
0.0064
0.0025
0.0099
0.8600
0.0080
clutter
0.2847
0.1199
0.1234
0.0131
0.0135
0.4454
Precision/Correctness
0.913
0.946
0.756
0.796
0.931
0.786
Recall/Completeness
0.916
0.968
0.829
0.878
0.860
0.445
F1
0.914
0.957
0.791
0.835
0.894
0.568

Overall accuracy 0.877

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
0.9291
0.0053
0.0421
0.0132
0.0004
0.0099
building
0.0068
0.9751
0.0066
0.0041
0.0001
0.0072
low_veg
0.0198
0.0018
0.8537
0.1112
0.0000
0.0135
tree
0.0251
0.0027
0.0662
0.9024
0.0020
0.0016
car
0.0231
0.0069
0.0002
0.0081
0.9522
0.0095
clutter
0.2843
0.0873
0.1117
0.0108
0.0150
0.4908
Precision/Correctness
0.934
0.967
0.777
0.823
0.936
0.799
Recall/Completeness
0.929
0.975
0.854
0.902
0.952
0.491
F1
0.931
0.971
0.814
0.861
0.944
0.608

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_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9287
0.0301
0.0204
0.0141
0.0009
0.0058
building
0.0280
0.9543
0.0137
0.0015
0.0000
0.0025
low_veg
0.0168
0.0054
0.9487
0.0278
0.0000
0.0012
tree
0.0623
0.0029
0.0933
0.8379
0.0016
0.0019
car
0.0628
0.0020
0.0001
0.0032
0.8941
0.0379
clutter
0.3643
0.0761
0.3493
0.0171
0.0037
0.1894
Precision/Correctness
0.930
0.940
0.885
0.888
0.945
0.508
Recall/Completeness
0.929
0.954
0.949
0.838
0.894
0.189
F1
0.929
0.947
0.916
0.862
0.919
0.276

Overall accuracy 0.913

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
0.9433
0.0270
0.0164
0.0088
0.0000
0.0044
building
0.0245
0.9627
0.0099
0.0009
0.0000
0.0020
low_veg
0.0101
0.0032
0.9632
0.0228
0.0000
0.0007
tree
0.0448
0.0019
0.0806
0.8697
0.0012
0.0018
car
0.0062
0.0025
0.0000
0.0019
0.9473
0.0420
clutter
0.3728
0.0734
0.3672
0.0160
0.0032
0.1673
Precision/Correctness
0.945
0.950
0.902
0.916
0.977
0.508
Recall/Completeness
0.943
0.963
0.963
0.870
0.947
0.167
F1
0.944
0.956
0.932
0.892
0.962
0.252

Overall accuracy 0.930

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
0.9538
0.0060
0.0256
0.0093
0.0031
0.0022
building
0.0325
0.9561
0.0043
0.0011
0.0001
0.0059
low_veg
0.0418
0.0072
0.8885
0.0583
0.0003
0.0040
tree
0.0435
0.0082
0.0853
0.8567
0.0055
0.0008
car
0.0506
0.0060
0.0013
0.0081
0.9278
0.0062
clutter
0.3177
0.2134
0.2225
0.0201
0.0154
0.2109
Precision/Correctness
0.927
0.968
0.848
0.856
0.932
0.584
Recall/Completeness
0.954
0.956
0.888
0.857
0.928
0.211
F1
0.940
0.962
0.868
0.856
0.930
0.310

Overall accuracy 0.917

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
0.9677
0.0028
0.0210
0.0067
0.0001
0.0017
building
0.0286
0.9634
0.0017
0.0007
0.0000
0.0057
low_veg
0.0280
0.0030
0.9139
0.0522
0.0000
0.0029
tree
0.0322
0.0067
0.0718
0.8842
0.0043
0.0008
car
0.0033
0.0057
0.0001
0.0046
0.9794
0.0070
clutter
0.3083
0.2361
0.2208
0.0161
0.0151
0.2036
Precision/Correctness
0.944
0.977
0.871
0.883
0.970
0.575
Recall/Completeness
0.968
0.963
0.914
0.884
0.979
0.204
F1
0.956
0.970
0.892
0.883
0.975
0.301

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
0.9306
0.0049
0.0347
0.0099
0.0034
0.0165
building
0.0097
0.9777
0.0043
0.0041
0.0001
0.0040
low_veg
0.0333
0.0038
0.9009
0.0427
0.0008
0.0186
tree
0.0372
0.0059
0.1523
0.7975
0.0046
0.0025
car
0.0603
0.0049
0.0045
0.0079
0.9167
0.0057
clutter
0.1611
0.0084
0.1889
0.0075
0.0136
0.6204
Precision/Correctness
0.935
0.983
0.731
0.914
0.905
0.780
Recall/Completeness
0.931
0.978
0.901
0.797
0.917
0.620
F1
0.933
0.980
0.807
0.852
0.911
0.691

Overall accuracy 0.897

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
0.9469
0.0028
0.0280
0.0068
0.0001
0.0153
building
0.0054
0.9863
0.0012
0.0037
0.0000
0.0034
low_veg
0.0194
0.0021
0.9264
0.0346
0.0000
0.0174
tree
0.0276
0.0045
0.1345
0.8278
0.0037
0.0020
car
0.0079
0.0056
0.0010
0.0050
0.9746
0.0060
clutter
0.1569
0.0066
0.1803
0.0055
0.0139
0.6369
Precision/Correctness
0.953
0.989
0.753
0.936
0.943
0.799
Recall/Completeness
0.947
0.986
0.926
0.828
0.975
0.637
F1
0.950
0.988
0.831
0.879
0.959
0.709

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
0.9110
0.0415
0.0084
0.0262
0.0014
0.0116
building
0.0090
0.9620
0.0016
0.0015
0.0002
0.0257
low_veg
0.0448
0.0193
0.7963
0.1348
0.0001
0.0047
tree
0.0198
0.0064
0.0883
0.8841
0.0010
0.0004
car
0.0466
0.0031
0.0012
0.0190
0.9241
0.0059
clutter
0.3125
0.0237
0.0596
0.0021
0.0060
0.5961
Precision/Correctness
0.924
0.895
0.827
0.715
0.904
0.817
Recall/Completeness
0.911
0.962
0.796
0.884
0.924
0.596
F1
0.918
0.927
0.811
0.791
0.914
0.689

Overall accuracy 0.881

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
0.9252
0.0382
0.0062
0.0193
0.0002
0.0110
building
0.0065
0.9664
0.0009
0.0008
0.0001
0.0253
low_veg
0.0332
0.0149
0.8214
0.1263
0.0000
0.0042
tree
0.0121
0.0043
0.0793
0.9035
0.0006
0.0002
car
0.0032
0.0025
0.0005
0.0131
0.9747
0.0060
clutter
0.3069
0.0215
0.0582
0.0018
0.0058
0.6058
Precision/Correctness
0.932
0.906
0.845
0.748
0.935
0.826
Recall/Completeness
0.925
0.966
0.821
0.904
0.975
0.606
F1
0.929
0.935
0.833
0.819
0.954
0.699

Overall accuracy 0.894

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image