Detailled semantic labeling (2D) result


Name T. V. Duy
Affiliation Hanoi University of Science and Technology, Vietnam
Abbreviation BKHN_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
0.916
0.017
0.034
0.018
0.002
0.012
building
0.015
0.968
0.006
0.004
0.000
0.007
low_veg
0.044
0.010
0.871
0.065
0.000
0.009
tree
0.039
0.007
0.126
0.822
0.003
0.004
car
0.079
0.006
0.003
0.013
0.883
0.015
clutter
0.330
0.110
0.155
0.025
0.010
0.370
Precision/Correctness
0.888
0.944
0.818
0.866
0.916
0.668
Recall/Completeness
0.916
0.968
0.871
0.822
0.883
0.370
F1
0.902
0.956
0.844
0.843
0.899
0.476

Overall accuracy 0.877

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
0.934
0.013
0.028
0.013
0.000
0.011
building
0.012
0.976
0.003
0.003
0.000
0.006
low_veg
0.033
0.006
0.897
0.057
0.000
0.008
tree
0.030
0.005
0.111
0.849
0.002
0.003
car
0.013
0.007
0.001
0.009
0.954
0.017
clutter
0.337
0.095
0.147
0.021
0.010
0.390
Precision/Correctness
0.909
0.959
0.841
0.891
0.945
0.691
Recall/Completeness
0.934
0.976
0.897
0.849
0.954
0.390
F1
0.921
0.967
0.868
0.869
0.949
0.498

Overall accuracy 0.899
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9037
0.0173
0.0412
0.0123
0.0021
0.0235
building
0.0101
0.9575
0.0087
0.0150
0.0000
0.0087
low_veg
0.0327
0.0068
0.9000
0.0472
0.0002
0.0130
tree
0.0433
0.0104
0.1343
0.8038
0.0013
0.0068
car
0.0870
0.0008
0.0018
0.0074
0.8529
0.0501
clutter
0.1215
0.1567
0.1058
0.0220
0.0195
0.5744
Precision/Correctness
0.898
0.938
0.858
0.901
0.887
0.450
Recall/Completeness
0.904
0.957
0.900
0.804
0.853
0.574
F1
0.901
0.947
0.879
0.849
0.870
0.505

Overall accuracy 0.882

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
0.9242
0.0126
0.0335
0.0080
0.0003
0.0214
building
0.0060
0.9685
0.0049
0.0137
0.0000
0.0069
low_veg
0.0245
0.0035
0.9236
0.0371
0.0001
0.0111
tree
0.0345
0.0089
0.1144
0.8355
0.0009
0.0058
car
0.0321
0.0004
0.0006
0.0034
0.9044
0.0590
clutter
0.0814
0.1551
0.0712
0.0146
0.0239
0.6538
Precision/Correctness
0.926
0.956
0.884
0.924
0.927
0.440
Recall/Completeness
0.924
0.968
0.924
0.835
0.904
0.654
F1
0.925
0.962
0.903
0.877
0.916
0.526

Overall accuracy 0.908

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
0.9114
0.0057
0.0658
0.0106
0.0011
0.0054
building
0.0088
0.9695
0.0072
0.0065
0.0000
0.0079
low_veg
0.0158
0.0031
0.9391
0.0363
0.0000
0.0057
tree
0.0231
0.0041
0.1807
0.7894
0.0008
0.0019
car
0.0814
0.0033
0.0016
0.0084
0.9021
0.0032
clutter
0.0777
0.3131
0.1301
0.0674
0.0104
0.4014
Precision/Correctness
0.897
0.934
0.837
0.942
0.915
0.245
Recall/Completeness
0.911
0.970
0.939
0.789
0.902
0.401
F1
0.904
0.951
0.885
0.859
0.909
0.304

Overall accuracy 0.882

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
0.9360
0.0030
0.0505
0.0064
0.0002
0.0039
building
0.0052
0.9813
0.0029
0.0044
0.0000
0.0061
low_veg
0.0102
0.0015
0.9544
0.0289
0.0000
0.0051
tree
0.0188
0.0033
0.1649
0.8109
0.0005
0.0016
car
0.0182
0.0039
0.0002
0.0052
0.9702
0.0023
clutter
0.0419
0.3653
0.0927
0.0790
0.0116
0.4094
Precision/Correctness
0.923
0.955
0.854
0.956
0.947
0.221
Recall/Completeness
0.936
0.981
0.954
0.811
0.970
0.409
F1
0.930
0.968
0.902
0.877
0.959
0.287

Overall accuracy 0.901

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
0.8974
0.0179
0.0463
0.0187
0.0027
0.0169
building
0.0091
0.9768
0.0048
0.0044
0.0002
0.0048
low_veg
0.0547
0.0191
0.8515
0.0640
0.0001
0.0105
tree
0.0339
0.0067
0.0998
0.8526
0.0024
0.0046
car
0.0703
0.0046
0.0007
0.0170
0.8996
0.0077
clutter
0.2995
0.1525
0.2007
0.0245
0.0057
0.3170
Precision/Correctness
0.879
0.945
0.829
0.902
0.915
0.454
Recall/Completeness
0.897
0.977
0.851
0.853
0.900
0.317
F1
0.888
0.960
0.840
0.876
0.907
0.373

Overall accuracy 0.882

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
0.9225
0.0117
0.0380
0.0123
0.0002
0.0153
building
0.0060
0.9852
0.0018
0.0032
0.0002
0.0036
low_veg
0.0435
0.0106
0.8852
0.0529
0.0000
0.0077
tree
0.0257
0.0053
0.0838
0.8794
0.0018
0.0039
car
0.0078
0.0053
0.0001
0.0110
0.9676
0.0082
clutter
0.3167
0.1480
0.1751
0.0172
0.0064
0.3366
Precision/Correctness
0.908
0.965
0.858
0.927
0.954
0.451
Recall/Completeness
0.922
0.985
0.885
0.879
0.968
0.337
F1
0.915
0.975
0.871
0.903
0.961
0.386

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
0.8848
0.0208
0.0570
0.0225
0.0026
0.0122
building
0.0091
0.9800
0.0042
0.0022
0.0001
0.0045
low_veg
0.0309
0.0099
0.8921
0.0596
0.0001
0.0074
tree
0.0274
0.0064
0.1744
0.7862
0.0015
0.0041
car
0.0701
0.0232
0.0016
0.0219
0.8712
0.0120
clutter
0.2117
0.1596
0.1383
0.0377
0.0083
0.4444
Precision/Correctness
0.883
0.940
0.796
0.899
0.893
0.665
Recall/Completeness
0.885
0.980
0.892
0.786
0.871
0.444
F1
0.884
0.959
0.841
0.839
0.882
0.533

Overall accuracy 0.869

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
0.9129
0.0147
0.0465
0.0158
0.0002
0.0099
building
0.0070
0.9870
0.0014
0.0014
0.0001
0.0031
low_veg
0.0205
0.0047
0.9212
0.0485
0.0000
0.0051
tree
0.0223
0.0052
0.1596
0.8084
0.0012
0.0032
car
0.0099
0.0259
0.0002
0.0147
0.9365
0.0127
clutter
0.2049
0.1514
0.1127
0.0307
0.0099
0.4904
Precision/Correctness
0.912
0.959
0.818
0.924
0.931
0.711
Recall/Completeness
0.913
0.987
0.921
0.808
0.937
0.490
F1
0.912
0.973
0.866
0.862
0.934
0.581

Overall accuracy 0.894

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
0.9289
0.0110
0.0278
0.0260
0.0013
0.0050
building
0.0098
0.9753
0.0061
0.0051
0.0003
0.0034
low_veg
0.0489
0.0110
0.8169
0.1164
0.0001
0.0068
tree
0.0342
0.0043
0.0483
0.9104
0.0020
0.0007
car
0.0769
0.0022
0.0024
0.0209
0.8914
0.0062
clutter
0.4820
0.1143
0.1904
0.0697
0.0125
0.1310
Precision/Correctness
0.852
0.961
0.800
0.786
0.924
0.697
Recall/Completeness
0.929
0.975
0.817
0.910
0.891
0.131
F1
0.889
0.968
0.809
0.844
0.907
0.221

Overall accuracy 0.871

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
0.0079
0.0223
0.0197
0.0001
0.0037
building
0.0048
0.9858
0.0025
0.0040
0.0002
0.0027
low_veg
0.0370
0.0063
0.8446
0.1069
0.0000
0.0052
tree
0.0224
0.0027
0.0353
0.9377
0.0013
0.0005
car
0.0075
0.0020
0.0001
0.0140
0.9702
0.0062
clutter
0.5044
0.1006
0.1857
0.0640
0.0134
0.1319
Precision/Correctness
0.877
0.972
0.833
0.819
0.941
0.731
Recall/Completeness
0.946
0.986
0.845
0.938
0.970
0.132
F1
0.910
0.979
0.839
0.874
0.955
0.223

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_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8865
0.0195
0.0540
0.0147
0.0027
0.0225
building
0.0073
0.9808
0.0031
0.0043
0.0000
0.0045
low_veg
0.0942
0.0165
0.7896
0.0809
0.0003
0.0185
tree
0.0359
0.0072
0.1284
0.8187
0.0027
0.0071
car
0.0735
0.0073
0.0011
0.0195
0.8785
0.0201
clutter
0.5147
0.0434
0.1425
0.0122
0.0006
0.2865
Precision/Correctness
0.732
0.931
0.775
0.861
0.938
0.691
Recall/Completeness
0.887
0.981
0.790
0.819
0.879
0.287
F1
0.802
0.955
0.782
0.839
0.907
0.405

Overall accuracy 0.811

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
0.9028
0.0145
0.0506
0.0101
0.0002
0.0218
building
0.0049
0.9871
0.0013
0.0033
0.0000
0.0035
low_veg
0.0829
0.0109
0.8157
0.0735
0.0002
0.0168
tree
0.0270
0.0055
0.1118
0.8475
0.0021
0.0061
car
0.0188
0.0069
0.0001
0.0151
0.9375
0.0216
clutter
0.5323
0.0396
0.1330
0.0088
0.0004
0.2859
Precision/Correctness
0.743
0.949
0.793
0.885
0.971
0.713
Recall/Completeness
0.903
0.987
0.816
0.847
0.938
0.286
F1
0.815
0.968
0.804
0.866
0.954
0.408

Overall accuracy 0.829

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
0.9170
0.0175
0.0485
0.0060
0.0025
0.0084
building
0.0080
0.9794
0.0058
0.0034
0.0000
0.0033
low_veg
0.0586
0.0185
0.8800
0.0345
0.0004
0.0081
tree
0.0768
0.0118
0.1409
0.7579
0.0085
0.0040
car
0.0915
0.0111
0.0040
0.0036
0.8697
0.0202
clutter
0.2691
0.2343
0.1915
0.0126
0.0056
0.2869
Precision/Correctness
0.908
0.945
0.774
0.913
0.904
0.571
Recall/Completeness
0.917
0.979
0.880
0.758
0.870
0.287
F1
0.912
0.962
0.824
0.828
0.887
0.382

Overall accuracy 0.890

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
0.9325
0.0118
0.0440
0.0040
0.0009
0.0068
building
0.0051
0.9857
0.0038
0.0027
0.0000
0.0027
low_veg
0.0390
0.0103
0.9163
0.0289
0.0000
0.0054
tree
0.0603
0.0102
0.1199
0.7996
0.0069
0.0031
car
0.0148
0.0134
0.0006
0.0014
0.9468
0.0229
clutter
0.2712
0.2386
0.1990
0.0093
0.0049
0.2772
Precision/Correctness
0.935
0.961
0.798
0.936
0.933
0.575
Recall/Completeness
0.932
0.986
0.916
0.800
0.947
0.277
F1
0.934
0.973
0.853
0.862
0.940
0.374

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_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9254
0.0145
0.0271
0.0246
0.0015
0.0069
building
0.0104
0.9804
0.0035
0.0034
0.0000
0.0024
low_veg
0.0690
0.0112
0.8307
0.0822
0.0000
0.0069
tree
0.0484
0.0069
0.1136
0.8259
0.0026
0.0027
car
0.0950
0.0069
0.0009
0.0078
0.8712
0.0181
clutter
0.2342
0.2905
0.1089
0.0203
0.0282
0.3179
Precision/Correctness
0.906
0.942
0.852
0.799
0.928
0.667
Recall/Completeness
0.925
0.980
0.831
0.826
0.871
0.318
F1
0.915
0.961
0.841
0.812
0.899
0.431

Overall accuracy 0.889

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
0.9433
0.0094
0.0233
0.0187
0.0001
0.0052
building
0.0066
0.9871
0.0020
0.0026
0.0000
0.0017
low_veg
0.0618
0.0062
0.8537
0.0731
0.0000
0.0051
tree
0.0340
0.0052
0.1011
0.8554
0.0021
0.0022
car
0.0150
0.0079
0.0001
0.0061
0.9513
0.0197
clutter
0.2065
0.2926
0.1011
0.0191
0.0348
0.3458
Precision/Correctness
0.931
0.959
0.876
0.828
0.946
0.691
Recall/Completeness
0.943
0.987
0.854
0.855
0.951
0.346
F1
0.937
0.973
0.865
0.841
0.949
0.461

Overall accuracy 0.913

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
0.9277
0.0081
0.0236
0.0286
0.0011
0.0108
building
0.0167
0.9754
0.0043
0.0019
0.0001
0.0017
low_veg
0.0570
0.0080
0.8311
0.0908
0.0002
0.0129
tree
0.0483
0.0063
0.0767
0.8611
0.0039
0.0036
car
0.0958
0.0111
0.0098
0.0145
0.8529
0.0158
clutter
0.2150
0.1853
0.1717
0.0487
0.0054
0.3739
Precision/Correctness
0.855
0.928
0.836
0.830
0.921
0.800
Recall/Completeness
0.928
0.975
0.831
0.861
0.853
0.374
F1
0.890
0.951
0.833
0.845
0.885
0.510

Overall accuracy 0.865

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
0.9458
0.0052
0.0168
0.0225
0.0001
0.0096
building
0.0148
0.9803
0.0022
0.0014
0.0000
0.0013
low_veg
0.0428
0.0048
0.8558
0.0852
0.0000
0.0113
tree
0.0381
0.0052
0.0675
0.8830
0.0031
0.0031
car
0.0125
0.0123
0.0080
0.0098
0.9401
0.0173
clutter
0.2059
0.1580
0.1667
0.0475
0.0055
0.4163
Precision/Correctness
0.882
0.948
0.860
0.853
0.940
0.823
Recall/Completeness
0.946
0.980
0.856
0.883
0.940
0.416
F1
0.913
0.964
0.858
0.868
0.940
0.553

Overall accuracy 0.889

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
0.8993
0.0093
0.0466
0.0196
0.0011
0.0241
building
0.0146
0.9583
0.0089
0.0057
0.0005
0.0119
low_veg
0.0398
0.0065
0.8206
0.1194
0.0001
0.0137
tree
0.0402
0.0056
0.0830
0.8665
0.0023
0.0024
car
0.1175
0.0087
0.0034
0.0120
0.8470
0.0114
clutter
0.2754
0.1174
0.1382
0.0181
0.0189
0.4320
Precision/Correctness
0.908
0.945
0.745
0.783
0.911
0.667
Recall/Completeness
0.899
0.958
0.821
0.867
0.847
0.432
F1
0.904
0.952
0.781
0.823
0.878
0.524

Overall accuracy 0.865

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
0.9127
0.0058
0.0421
0.0151
0.0004
0.0240
building
0.0106
0.9671
0.0066
0.0044
0.0005
0.0107
low_veg
0.0244
0.0031
0.8454
0.1144
0.0000
0.0126
tree
0.0299
0.0037
0.0723
0.8903
0.0017
0.0021
car
0.0247
0.0095
0.0006
0.0101
0.9421
0.0130
clutter
0.2755
0.0874
0.1287
0.0161
0.0209
0.4715
Precision/Correctness
0.929
0.965
0.766
0.810
0.916
0.677
Recall/Completeness
0.913
0.967
0.845
0.890
0.942
0.471
F1
0.921
0.966
0.804
0.848
0.929
0.556

Overall accuracy 0.887

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
0.9250
0.0299
0.0192
0.0159
0.0012
0.0089
building
0.0295
0.9491
0.0150
0.0016
0.0001
0.0046
low_veg
0.0199
0.0062
0.9400
0.0318
0.0000
0.0020
tree
0.0645
0.0028
0.0994
0.8287
0.0015
0.0032
car
0.0668
0.0047
0.0003
0.0031
0.8836
0.0415
clutter
0.4628
0.0814
0.2458
0.0204
0.0031
0.1865
Precision/Correctness
0.921
0.939
0.889
0.873
0.939
0.397
Recall/Completeness
0.925
0.949
0.940
0.829
0.884
0.186
F1
0.923
0.944
0.914
0.850
0.910
0.254

Overall accuracy 0.907

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
0.9401
0.0268
0.0153
0.0103
0.0002
0.0073
building
0.0257
0.9584
0.0111
0.0011
0.0001
0.0037
low_veg
0.0122
0.0039
0.9561
0.0267
0.0000
0.0011
tree
0.0478
0.0019
0.0872
0.8590
0.0011
0.0029
car
0.0066
0.0059
0.0001
0.0017
0.9384
0.0474
clutter
0.4896
0.0781
0.2491
0.0191
0.0021
0.1621
Precision/Correctness
0.937
0.949
0.907
0.902
0.973
0.384
Recall/Completeness
0.940
0.958
0.956
0.859
0.938
0.162
F1
0.939
0.954
0.931
0.880
0.955
0.228

Overall accuracy 0.925

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
0.9517
0.0065
0.0246
0.0112
0.0031
0.0030
building
0.0433
0.9418
0.0052
0.0015
0.0002
0.0080
low_veg
0.0422
0.0073
0.8838
0.0612
0.0003
0.0051
tree
0.0447
0.0083
0.0923
0.8482
0.0053
0.0012
car
0.0587
0.0054
0.0011
0.0099
0.9158
0.0091
clutter
0.3228
0.1570
0.2118
0.0221
0.0116
0.2746
Precision/Correctness
0.918
0.972
0.845
0.842
0.932
0.573
Recall/Completeness
0.952
0.942
0.884
0.848
0.916
0.275
F1
0.935
0.956
0.864
0.845
0.924
0.371

Overall accuracy 0.911

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
0.9657
0.0033
0.0200
0.0084
0.0001
0.0023
building
0.0392
0.9499
0.0022
0.0010
0.0001
0.0075
low_veg
0.0268
0.0035
0.9109
0.0552
0.0000
0.0036
tree
0.0336
0.0069
0.0789
0.8753
0.0042
0.0010
car
0.0043
0.0048
0.0002
0.0060
0.9750
0.0098
clutter
0.3189
0.1710
0.2063
0.0177
0.0116
0.2745
Precision/Correctness
0.936
0.980
0.870
0.870
0.971
0.579
Recall/Completeness
0.966
0.950
0.911
0.875
0.975
0.274
F1
0.950
0.965
0.890
0.872
0.973
0.373

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_15_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9303
0.0053
0.0337
0.0122
0.0036
0.0149
building
0.0127
0.9737
0.0051
0.0022
0.0001
0.0061
low_veg
0.0390
0.0045
0.8876
0.0498
0.0009
0.0181
tree
0.0401
0.0075
0.1487
0.7958
0.0046
0.0033
car
0.0644
0.0011
0.0043
0.0098
0.9120
0.0083
clutter
0.1801
0.0107
0.2042
0.0102
0.0154
0.5794
Precision/Correctness
0.927
0.980
0.727
0.902
0.898
0.767
Recall/Completeness
0.930
0.974
0.888
0.796
0.912
0.579
F1
0.929
0.977
0.800
0.846
0.905
0.660

Overall accuracy 0.892

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
0.9473
0.0032
0.0270
0.0088
0.0002
0.0135
building
0.0081
0.9833
0.0019
0.0016
0.0001
0.0050
low_veg
0.0241
0.0030
0.9149
0.0410
0.0001
0.0169
tree
0.0308
0.0062
0.1317
0.8249
0.0037
0.0028
car
0.0086
0.0008
0.0011
0.0067
0.9740
0.0088
clutter
0.1768
0.0090
0.1972
0.0080
0.0157
0.5934
Precision/Correctness
0.945
0.987
0.749
0.927
0.938
0.791
Recall/Completeness
0.947
0.983
0.915
0.825
0.974
0.593
F1
0.946
0.985
0.824
0.873
0.956
0.678

Overall accuracy 0.912

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
0.9089
0.0416
0.0085
0.0275
0.0014
0.0122
building
0.0128
0.9515
0.0018
0.0022
0.0006
0.0310
low_veg
0.0520
0.0213
0.7798
0.1418
0.0001
0.0049
tree
0.0203
0.0088
0.0956
0.8733
0.0008
0.0012
car
0.0513
0.0026
0.0009
0.0246
0.9103
0.0103
clutter
0.3769
0.0219
0.0854
0.0031
0.0055
0.5071
Precision/Correctness
0.910
0.893
0.799
0.700
0.898
0.770
Recall/Completeness
0.909
0.951
0.780
0.873
0.910
0.507
F1
0.910
0.921
0.789
0.777
0.904
0.611

Overall accuracy 0.866

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
0.9232
0.0384
0.0062
0.0206
0.0002
0.0113
building
0.0103
0.9563
0.0011
0.0014
0.0005
0.0304
low_veg
0.0395
0.0174
0.8057
0.1334
0.0000
0.0040
tree
0.0122
0.0069
0.0873
0.8923
0.0004
0.0009
car
0.0035
0.0021
0.0002
0.0178
0.9664
0.0100
clutter
0.3731
0.0198
0.0848
0.0028
0.0051
0.5143
Precision/Correctness
0.918
0.904
0.815
0.732
0.934
0.782
Recall/Completeness
0.923
0.956
0.806
0.892
0.966
0.514
F1
0.920
0.929
0.810
0.804
0.950
0.621

Overall accuracy 0.880

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image