Detailled semantic labeling (2D) result


Name Y. Sun
Affiliation Huazhong University of Science and Technology (HUST), Wuhan, China
Abbreviation HUSTW4
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
0.925
0.012
0.033
0.021
0.002
0.007
building
0.011
0.980
0.004
0.003
0.000
0.002
low_veg
0.039
0.011
0.878
0.065
0.000
0.007
tree
0.028
0.007
0.106
0.856
0.001
0.002
car
0.112
0.004
0.003
0.029
0.837
0.014
clutter
0.286
0.120
0.133
0.029
0.010
0.422
Precision/Correctness
0.904
0.947
0.838
0.864
0.928
0.807
Recall/Completeness
0.925
0.980
0.878
0.856
0.837
0.422
F1
0.914
0.963
0.858
0.860
0.880
0.554

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
0.946
0.008
0.025
0.015
0.000
0.006
building
0.008
0.987
0.002
0.002
0.000
0.001
low_veg
0.029
0.006
0.905
0.054
0.000
0.006
tree
0.020
0.006
0.091
0.881
0.001
0.002
car
0.022
0.004
0.001
0.021
0.936
0.015
clutter
0.282
0.102
0.116
0.024
0.011
0.465
Precision/Correctness
0.927
0.965
0.865
0.894
0.956
0.835
Recall/Completeness
0.946
0.987
0.905
0.881
0.936
0.465
F1
0.936
0.976
0.885
0.888
0.946
0.598

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
0.9166
0.0130
0.0435
0.0197
0.0014
0.0058
building
0.0116
0.9635
0.0135
0.0071
0.0000
0.0043
low_veg
0.0286
0.0077
0.9094
0.0478
0.0002
0.0064
tree
0.0251
0.0108
0.1107
0.8494
0.0006
0.0034
car
0.1385
0.0028
0.0024
0.0281
0.7911
0.0372
clutter
0.1693
0.1613
0.1596
0.0395
0.0149
0.4554
Precision/Correctness
0.913
0.941
0.868
0.901
0.919
0.618
Recall/Completeness
0.917
0.963
0.909
0.849
0.791
0.455
F1
0.915
0.952
0.888
0.874
0.850
0.524

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
0.9432
0.0083
0.0324
0.0127
0.0001
0.0033
building
0.0069
0.9755
0.0084
0.0057
0.0000
0.0035
low_veg
0.0218
0.0039
0.9350
0.0344
0.0001
0.0048
tree
0.0180
0.0092
0.0916
0.8782
0.0004
0.0027
car
0.0502
0.0022
0.0011
0.0187
0.8843
0.0435
clutter
0.1311
0.1552
0.1167
0.0295
0.0196
0.5480
Precision/Correctness
0.941
0.961
0.897
0.931
0.951
0.664
Recall/Completeness
0.943
0.976
0.935
0.878
0.884
0.548
F1
0.942
0.968
0.915
0.904
0.917
0.600

Overall accuracy 0.925

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
0.9032
0.0054
0.0657
0.0230
0.0010
0.0016
building
0.0075
0.9757
0.0075
0.0083
0.0000
0.0010
low_veg
0.0159
0.0035
0.9271
0.0527
0.0000
0.0007
tree
0.0136
0.0045
0.1117
0.8691
0.0004
0.0007
car
0.1223
0.0003
0.0020
0.0288
0.8417
0.0049
clutter
0.1131
0.3469
0.1614
0.0835
0.0144
0.2807
Precision/Correctness
0.914
0.929
0.883
0.921
0.934
0.528
Recall/Completeness
0.903
0.976
0.927
0.869
0.842
0.281
F1
0.909
0.952
0.904
0.894
0.885
0.366

Overall accuracy 0.903

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
0.9350
0.0025
0.0460
0.0152
0.0001
0.0012
building
0.0035
0.9876
0.0025
0.0057
0.0000
0.0008
low_veg
0.0114
0.0017
0.9444
0.0421
0.0000
0.0005
tree
0.0104
0.0036
0.0971
0.8881
0.0003
0.0006
car
0.0322
0.0003
0.0000
0.0185
0.9424
0.0065
clutter
0.0616
0.4013
0.1162
0.0795
0.0160
0.3254
Precision/Correctness
0.939
0.953
0.902
0.940
0.969
0.554
Recall/Completeness
0.935
0.988
0.944
0.888
0.942
0.325
F1
0.937
0.970
0.923
0.913
0.955
0.410

Overall accuracy 0.924

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
0.9081
0.0202
0.0377
0.0278
0.0025
0.0037
building
0.0090
0.9832
0.0032
0.0027
0.0000
0.0019
low_veg
0.0523
0.0230
0.8485
0.0711
0.0001
0.0051
tree
0.0201
0.0113
0.0806
0.8855
0.0010
0.0016
car
0.1030
0.0070
0.0012
0.0371
0.8376
0.0142
clutter
0.3440
0.1799
0.2353
0.0320
0.0069
0.2019
Precision/Correctness
0.888
0.933
0.850
0.889
0.933
0.601
Recall/Completeness
0.908
0.983
0.849
0.885
0.838
0.202
F1
0.898
0.958
0.849
0.887
0.883
0.302

Overall accuracy 0.890

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
0.9379
0.0128
0.0278
0.0187
0.0001
0.0027
building
0.0061
0.9898
0.0007
0.0018
0.0000
0.0015
low_veg
0.0405
0.0123
0.8858
0.0577
0.0000
0.0037
tree
0.0139
0.0093
0.0657
0.9091
0.0006
0.0013
car
0.0193
0.0077
0.0006
0.0258
0.9322
0.0143
clutter
0.3703
0.1737
0.2015
0.0202
0.0078
0.2266
Precision/Correctness
0.917
0.957
0.883
0.919
0.975
0.633
Recall/Completeness
0.938
0.990
0.886
0.909
0.932
0.227
F1
0.928
0.973
0.885
0.914
0.953
0.334

Overall accuracy 0.920

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
0.8943
0.0190
0.0492
0.0304
0.0023
0.0048
building
0.0080
0.9862
0.0033
0.0018
0.0000
0.0007
low_veg
0.0287
0.0120
0.8929
0.0634
0.0000
0.0030
tree
0.0176
0.0078
0.1369
0.8347
0.0008
0.0023
car
0.1063
0.0073
0.0033
0.0476
0.8213
0.0142
clutter
0.3403
0.2002
0.1782
0.0416
0.0027
0.2370
Precision/Correctness
0.880
0.934
0.824
0.893
0.927
0.724
Recall/Completeness
0.894
0.986
0.893
0.835
0.821
0.237
F1
0.887
0.959
0.857
0.863
0.871
0.357

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
0.9256
0.0119
0.0370
0.0214
0.0004
0.0037
building
0.0060
0.9918
0.0008
0.0010
0.0000
0.0005
low_veg
0.0190
0.0057
0.9231
0.0503
0.0000
0.0018
tree
0.0133
0.0064
0.1215
0.8562
0.0007
0.0020
car
0.0215
0.0068
0.0012
0.0348
0.9209
0.0148
clutter
0.3637
0.1939
0.1465
0.0328
0.0031
0.2600
Precision/Correctness
0.908
0.955
0.850
0.921
0.963
0.767
Recall/Completeness
0.926
0.992
0.923
0.856
0.921
0.260
F1
0.917
0.973
0.885
0.888
0.941
0.388

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
0.9269
0.0108
0.0277
0.0306
0.0011
0.0030
building
0.0099
0.9803
0.0044
0.0039
0.0001
0.0014
low_veg
0.0444
0.0129
0.8248
0.1153
0.0001
0.0026
tree
0.0255
0.0047
0.0466
0.9225
0.0005
0.0002
car
0.1164
0.0029
0.0036
0.0485
0.8195
0.0090
clutter
0.5181
0.1345
0.1565
0.0750
0.0114
0.1044
Precision/Correctness
0.847
0.956
0.819
0.779
0.937
0.781
Recall/Completeness
0.927
0.980
0.825
0.922
0.819
0.104
F1
0.885
0.968
0.822
0.845
0.874
0.184

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
0.9484
0.0061
0.0207
0.0223
0.0000
0.0024
building
0.0053
0.9891
0.0015
0.0028
0.0001
0.0012
low_veg
0.0317
0.0053
0.8591
0.1020
0.0000
0.0020
tree
0.0164
0.0026
0.0350
0.9457
0.0002
0.0000
car
0.0210
0.0022
0.0007
0.0365
0.9299
0.0097
clutter
0.5434
0.1161
0.1455
0.0670
0.0129
0.1152
Precision/Correctness
0.873
0.972
0.855
0.819
0.951
0.814
Recall/Completeness
0.948
0.989
0.859
0.946
0.930
0.115
F1
0.909
0.980
0.857
0.878
0.940
0.202

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_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8830
0.0201
0.0646
0.0165
0.0022
0.0136
building
0.0056
0.9846
0.0027
0.0040
0.0006
0.0025
low_veg
0.0652
0.0178
0.8372
0.0588
0.0002
0.0208
tree
0.0264
0.0076
0.1341
0.8261
0.0009
0.0049
car
0.1011
0.0050
0.0037
0.0440
0.8274
0.0188
clutter
0.2311
0.0452
0.1210
0.0109
0.0038
0.5881
Precision/Correctness
0.829
0.929
0.780
0.883
0.937
0.852
Recall/Completeness
0.883
0.985
0.837
0.826
0.827
0.588
F1
0.855
0.956
0.808
0.854
0.879
0.696

Overall accuracy 0.852

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
0.9063
0.0142
0.0560
0.0106
0.0001
0.0127
building
0.0034
0.9899
0.0010
0.0030
0.0006
0.0020
low_veg
0.0551
0.0112
0.8639
0.0497
0.0000
0.0201
tree
0.0192
0.0057
0.1153
0.8553
0.0006
0.0038
car
0.0251
0.0042
0.0020
0.0343
0.9149
0.0195
clutter
0.2304
0.0406
0.1018
0.0079
0.0036
0.6156
Precision/Correctness
0.850
0.949
0.806
0.911
0.965
0.870
Recall/Completeness
0.906
0.990
0.864
0.855
0.915
0.616
F1
0.877
0.969
0.834
0.882
0.939
0.721

Overall accuracy 0.876

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
0.9218
0.0145
0.0447
0.0095
0.0018
0.0076
building
0.0035
0.9899
0.0037
0.0019
0.0001
0.0009
low_veg
0.0478
0.0219
0.8887
0.0377
0.0004
0.0035
tree
0.0524
0.0095
0.1267
0.8060
0.0038
0.0016
car
0.1213
0.0038
0.0037
0.0166
0.8445
0.0101
clutter
0.3806
0.2386
0.1373
0.0278
0.0052
0.2105
Precision/Correctness
0.914
0.948
0.799
0.902
0.938
0.610
Recall/Completeness
0.922
0.990
0.889
0.806
0.844
0.211
F1
0.918
0.969
0.841
0.851
0.889
0.313

Overall accuracy 0.900

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
0.9414
0.0076
0.0380
0.0060
0.0002
0.0067
building
0.0015
0.9947
0.0018
0.0013
0.0000
0.0007
low_veg
0.0304
0.0114
0.9266
0.0295
0.0000
0.0020
tree
0.0378
0.0074
0.1049
0.8463
0.0026
0.0011
car
0.0207
0.0026
0.0003
0.0090
0.9560
0.0114
clutter
0.4023
0.2398
0.1098
0.0212
0.0047
0.2222
Precision/Correctness
0.941
0.967
0.831
0.933
0.971
0.624
Recall/Completeness
0.941
0.995
0.927
0.846
0.956
0.222
F1
0.941
0.981
0.876
0.887
0.964
0.328

Overall accuracy 0.927

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
0.9267
0.0138
0.0292
0.0241
0.0019
0.0043
building
0.0070
0.9863
0.0031
0.0030
0.0000
0.0006
low_veg
0.0493
0.0121
0.8633
0.0694
0.0000
0.0059
tree
0.0366
0.0058
0.1129
0.8428
0.0014
0.0005
car
0.1089
0.0048
0.0008
0.0125
0.8680
0.0051
clutter
0.2593
0.2776
0.1344
0.0254
0.0606
0.2427
Precision/Correctness
0.919
0.944
0.851
0.817
0.890
0.728
Recall/Completeness
0.927
0.986
0.863
0.843
0.868
0.243
F1
0.923
0.965
0.857
0.830
0.879
0.364

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
0.9472
0.0078
0.0242
0.0178
0.0002
0.0029
building
0.0038
0.9918
0.0018
0.0023
0.0000
0.0004
low_veg
0.0427
0.0066
0.8866
0.0600
0.0000
0.0041
tree
0.0232
0.0037
0.0989
0.8729
0.0009
0.0004
car
0.0172
0.0050
0.0001
0.0095
0.9623
0.0058
clutter
0.2307
0.2740
0.1245
0.0230
0.0742
0.2735
Precision/Correctness
0.945
0.963
0.877
0.849
0.910
0.771
Recall/Completeness
0.947
0.992
0.887
0.873
0.962
0.274
F1
0.946
0.977
0.882
0.861
0.936
0.404

Overall accuracy 0.923

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
0.9247
0.0074
0.0261
0.0303
0.0009
0.0106
building
0.0060
0.9877
0.0039
0.0016
0.0001
0.0008
low_veg
0.0610
0.0104
0.8382
0.0772
0.0001
0.0130
tree
0.0372
0.0069
0.0825
0.8677
0.0015
0.0042
car
0.1213
0.0092
0.0104
0.0304
0.8081
0.0206
clutter
0.2195
0.1989
0.1290
0.0548
0.0042
0.3937
Precision/Correctness
0.863
0.924
0.844
0.839
0.949
0.809
Recall/Completeness
0.925
0.988
0.838
0.868
0.808
0.394
F1
0.893
0.955
0.841
0.853
0.873
0.530

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
0.9462
0.0036
0.0176
0.0232
0.0000
0.0093
building
0.0042
0.9917
0.0023
0.0011
0.0000
0.0007
low_veg
0.0477
0.0057
0.8645
0.0701
0.0000
0.0121
tree
0.0283
0.0056
0.0728
0.8888
0.0009
0.0037
car
0.0200
0.0092
0.0086
0.0205
0.9192
0.0225
clutter
0.2107
0.1679
0.1173
0.0522
0.0042
0.4476
Precision/Correctness
0.892
0.947
0.870
0.865
0.970
0.831
Recall/Completeness
0.946
0.992
0.864
0.889
0.919
0.448
F1
0.918
0.969
0.867
0.877
0.944
0.582

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_15_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9194
0.0099
0.0360
0.0238
0.0011
0.0097
building
0.0085
0.9800
0.0039
0.0049
0.0002
0.0025
low_veg
0.0385
0.0101
0.8122
0.1193
0.0001
0.0198
tree
0.0271
0.0078
0.0705
0.8919
0.0008
0.0019
car
0.1538
0.0034
0.0032
0.0266
0.7864
0.0266
clutter
0.2765
0.1322
0.0967
0.0230
0.0029
0.4687
Precision/Correctness
0.915
0.939
0.793
0.775
0.962
0.799
Recall/Completeness
0.919
0.980
0.812
0.892
0.786
0.469
F1
0.917
0.959
0.802
0.829
0.865
0.591

Overall accuracy 0.881

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
0.9367
0.0049
0.0309
0.0182
0.0005
0.0088
building
0.0049
0.9863
0.0027
0.0037
0.0002
0.0021
low_veg
0.0228
0.0050
0.8417
0.1113
0.0000
0.0191
tree
0.0184
0.0055
0.0611
0.9127
0.0006
0.0017
car
0.0345
0.0026
0.0006
0.0225
0.9111
0.0288
clutter
0.2759
0.0936
0.0873
0.0203
0.0030
0.5198
Precision/Correctness
0.937
0.964
0.817
0.807
0.975
0.819
Recall/Completeness
0.937
0.986
0.842
0.913
0.911
0.520
F1
0.937
0.975
0.829
0.857
0.942
0.636

Overall accuracy 0.906

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
0.9484
0.0106
0.0196
0.0171
0.0011
0.0032
building
0.0291
0.9625
0.0063
0.0008
0.0002
0.0011
low_veg
0.0201
0.0042
0.9480
0.0273
0.0000
0.0004
tree
0.0628
0.0039
0.0914
0.8405
0.0005
0.0009
car
0.0867
0.0039
0.0001
0.0112
0.8905
0.0076
clutter
0.4029
0.0644
0.3351
0.0326
0.0054
0.1595
Precision/Correctness
0.926
0.971
0.894
0.878
0.944
0.651
Recall/Completeness
0.948
0.962
0.948
0.840
0.891
0.160
F1
0.937
0.967
0.920
0.859
0.916
0.256

Overall accuracy 0.922

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
0.9655
0.0068
0.0144
0.0109
0.0002
0.0022
building
0.0257
0.9708
0.0020
0.0005
0.0002
0.0008
low_veg
0.0112
0.0016
0.9658
0.0212
0.0000
0.0002
tree
0.0459
0.0027
0.0793
0.8710
0.0004
0.0008
car
0.0093
0.0044
0.0000
0.0070
0.9728
0.0065
clutter
0.4061
0.0592
0.3444
0.0297
0.0050
0.1556
Precision/Correctness
0.943
0.982
0.914
0.911
0.976
0.696
Recall/Completeness
0.966
0.971
0.966
0.871
0.973
0.156
F1
0.954
0.976
0.939
0.890
0.974
0.254

Overall accuracy 0.941

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
0.9562
0.0067
0.0202
0.0117
0.0027
0.0025
building
0.0256
0.9669
0.0048
0.0014
0.0004
0.0008
low_veg
0.0496
0.0107
0.8759
0.0592
0.0003
0.0042
tree
0.0390
0.0092
0.0941
0.8553
0.0022
0.0003
car
0.1013
0.0029
0.0013
0.0218
0.8655
0.0072
clutter
0.3917
0.1879
0.1973
0.0299
0.0155
0.1777
Precision/Correctness
0.923
0.968
0.855
0.840
0.938
0.659
Recall/Completeness
0.956
0.967
0.876
0.855
0.866
0.178
F1
0.939
0.967
0.865
0.847
0.900
0.280

Overall accuracy 0.916

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
0.9718
0.0029
0.0147
0.0085
0.0001
0.0020
building
0.0213
0.9743
0.0024
0.0010
0.0004
0.0006
low_veg
0.0332
0.0043
0.9078
0.0515
0.0000
0.0032
tree
0.0287
0.0073
0.0809
0.8814
0.0014
0.0003
car
0.0158
0.0017
0.0001
0.0166
0.9584
0.0074
clutter
0.3867
0.2046
0.1862
0.0263
0.0158
0.1805
Precision/Correctness
0.944
0.978
0.882
0.871
0.974
0.693
Recall/Completeness
0.972
0.974
0.908
0.881
0.958
0.181
F1
0.958
0.976
0.895
0.876
0.966
0.286

Overall accuracy 0.938

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
0.9259
0.0058
0.0314
0.0148
0.0031
0.0189
building
0.0097
0.9819
0.0039
0.0025
0.0001
0.0019
low_veg
0.0415
0.0037
0.8719
0.0607
0.0007
0.0215
tree
0.0261
0.0072
0.1292
0.8329
0.0021
0.0025
car
0.0996
0.0030
0.0046
0.0263
0.8464
0.0200
clutter
0.1131
0.0139
0.1583
0.0119
0.0177
0.6849
Precision/Correctness
0.940
0.979
0.753
0.886
0.906
0.777
Recall/Completeness
0.926
0.982
0.872
0.833
0.846
0.685
F1
0.933
0.981
0.808
0.859
0.875
0.728

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_15_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9464
0.0025
0.0235
0.0099
0.0002
0.0175
building
0.0052
0.9906
0.0008
0.0019
0.0001
0.0015
low_veg
0.0263
0.0008
0.9036
0.0496
0.0001
0.0196
tree
0.0180
0.0055
0.1136
0.8593
0.0015
0.0020
car
0.0208
0.0027
0.0009
0.0194
0.9352
0.0210
clutter
0.1032
0.0112
0.1486
0.0090
0.0179
0.7101
Precision/Correctness
0.961
0.989
0.780
0.917
0.943
0.800
Recall/Completeness
0.946
0.991
0.904
0.859
0.935
0.710
F1
0.954
0.990
0.837
0.887
0.939
0.752

Overall accuracy 0.923

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
0.9412
0.0174
0.0098
0.0264
0.0012
0.0040
building
0.0114
0.9816
0.0027
0.0016
0.0009
0.0017
low_veg
0.0516
0.0141
0.8029
0.1280
0.0001
0.0033
tree
0.0202
0.0076
0.1161
0.8552
0.0004
0.0006
car
0.0915
0.0038
0.0018
0.0395
0.8589
0.0045
clutter
0.2712
0.0252
0.0275
0.0032
0.0073
0.6655
Precision/Correctness
0.930
0.944
0.827
0.710
0.887
0.956
Recall/Completeness
0.941
0.982
0.803
0.855
0.859
0.666
F1
0.936
0.963
0.815
0.776
0.873
0.785

Overall accuracy 0.904

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
0.9561
0.0136
0.0070
0.0194
0.0003
0.0037
building
0.0080
0.9870
0.0017
0.0010
0.0008
0.0015
low_veg
0.0387
0.0100
0.8304
0.1180
0.0000
0.0028
tree
0.0112
0.0051
0.1076
0.8756
0.0002
0.0004
car
0.0129
0.0033
0.0006
0.0315
0.9478
0.0039
clutter
0.2619
0.0227
0.0255
0.0027
0.0070
0.6802
Precision/Correctness
0.940
0.957
0.850
0.746
0.912
0.961
Recall/Completeness
0.956
0.987
0.830
0.876
0.948
0.680
F1
0.948
0.972
0.840
0.806
0.930
0.797

Overall accuracy 0.920

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image