Detailled semantic labeling (2D) result


Name L. Yang et al.
Affiliation BUCT, China
Abbreviation BUCTY1
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
0.911
0.010
0.059
0.012
0.001
0.007
building
0.014
0.967
0.010
0.003
0.000
0.007
low_veg
0.027
0.007
0.932
0.029
0.000
0.005
tree
0.038
0.005
0.197
0.754
0.003
0.003
car
0.104
0.001
0.004
0.010
0.850
0.030
clutter
0.266
0.093
0.317
0.014
0.005
0.305
Precision/Correctness
0.906
0.960
0.739
0.918
0.941
0.701
Recall/Completeness
0.911
0.967
0.932
0.754
0.850
0.305
F1
0.908
0.963
0.825
0.828
0.893
0.425

Overall accuracy 0.873

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
0.929
0.006
0.051
0.008
0.000
0.006
building
0.011
0.975
0.007
0.002
0.000
0.006
low_veg
0.018
0.003
0.951
0.024
0.000
0.004
tree
0.029
0.004
0.178
0.784
0.002
0.002
car
0.015
0.001
0.002
0.007
0.941
0.033
clutter
0.265
0.078
0.319
0.011
0.005
0.323
Precision/Correctness
0.927
0.973
0.761
0.940
0.961
0.723
Recall/Completeness
0.929
0.975
0.951
0.784
0.941
0.323
F1
0.928
0.974
0.845
0.855
0.951
0.447

Overall accuracy 0.894
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8861
0.0136
0.0736
0.0091
0.0016
0.0160
building
0.0089
0.9515
0.0279
0.0040
0.0000
0.0077
low_veg
0.0164
0.0054
0.9452
0.0223
0.0002
0.0105
tree
0.0375
0.0089
0.2014
0.7476
0.0013
0.0033
car
0.0872
0.0009
0.0024
0.0067
0.8384
0.0644
clutter
0.1071
0.1798
0.1941
0.0191
0.0009
0.4990
Precision/Correctness
0.922
0.943
0.795
0.945
0.928
0.489
Recall/Completeness
0.886
0.951
0.945
0.748
0.838
0.499
F1
0.904
0.947
0.864
0.835
0.881
0.494

Overall accuracy 0.876

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
0.9086
0.0088
0.0635
0.0058
0.0001
0.0132
building
0.0059
0.9611
0.0235
0.0030
0.0000
0.0064
low_veg
0.0105
0.0023
0.9622
0.0157
0.0001
0.0092
tree
0.0287
0.0072
0.1769
0.7837
0.0008
0.0026
car
0.0270
0.0006
0.0008
0.0034
0.8944
0.0738
clutter
0.0792
0.1854
0.1538
0.0127
0.0010
0.5678
Precision/Correctness
0.947
0.962
0.821
0.963
0.969
0.486
Recall/Completeness
0.909
0.961
0.962
0.784
0.894
0.568
F1
0.928
0.962
0.886
0.864
0.930
0.524

Overall accuracy 0.901

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
0.8840
0.0051
0.0961
0.0076
0.0007
0.0065
building
0.0065
0.9696
0.0108
0.0037
0.0000
0.0093
low_veg
0.0085
0.0026
0.9708
0.0164
0.0000
0.0017
tree
0.0155
0.0031
0.2555
0.7239
0.0008
0.0011
car
0.0855
0.0166
0.0074
0.0064
0.8490
0.0351
clutter
0.0848
0.2627
0.2191
0.0398
0.0048
0.3888
Precision/Correctness
0.930
0.944
0.789
0.968
0.922
0.320
Recall/Completeness
0.884
0.970
0.971
0.724
0.849
0.389
F1
0.906
0.956
0.871
0.828
0.884
0.351

Overall accuracy 0.867

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
0.9107
0.0024
0.0766
0.0046
0.0000
0.0057
building
0.0047
0.9801
0.0052
0.0023
0.0000
0.0077
low_veg
0.0047
0.0010
0.9809
0.0120
0.0000
0.0013
tree
0.0115
0.0024
0.2369
0.7477
0.0006
0.0009
car
0.0072
0.0211
0.0074
0.0044
0.9210
0.0388
clutter
0.0579
0.3020
0.1918
0.0424
0.0032
0.4027
Precision/Correctness
0.953
0.965
0.806
0.978
0.950
0.296
Recall/Completeness
0.911
0.980
0.981
0.748
0.921
0.403
F1
0.932
0.972
0.885
0.848
0.935
0.341

Overall accuracy 0.885

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
0.8760
0.0156
0.0840
0.0127
0.0018
0.0099
building
0.0099
0.9775
0.0058
0.0030
0.0000
0.0038
low_veg
0.0354
0.0147
0.9180
0.0278
0.0000
0.0041
tree
0.0303
0.0076
0.1590
0.7979
0.0024
0.0028
car
0.0961
0.0033
0.0027
0.0130
0.8619
0.0230
clutter
0.2650
0.1250
0.2943
0.0108
0.0043
0.3005
Precision/Correctness
0.896
0.952
0.762
0.945
0.929
0.568
Recall/Completeness
0.876
0.978
0.918
0.798
0.862
0.300
F1
0.886
0.965
0.833
0.865
0.894
0.393

Overall accuracy 0.879

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_3_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9017
0.0095
0.0725
0.0077
0.0001
0.0086
building
0.0080
0.9847
0.0022
0.0021
0.0000
0.0030
low_veg
0.0256
0.0066
0.9445
0.0206
0.0000
0.0027
tree
0.0223
0.0061
0.1378
0.8296
0.0018
0.0023
car
0.0145
0.0037
0.0009
0.0092
0.9468
0.0250
clutter
0.2745
0.1191
0.2690
0.0060
0.0050
0.3265
Precision/Correctness
0.925
0.971
0.791
0.964
0.960
0.577
Recall/Completeness
0.902
0.985
0.945
0.830
0.947
0.327
F1
0.913
0.978
0.861
0.892
0.953
0.417

Overall accuracy 0.906

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
0.8525
0.0129
0.1069
0.0151
0.0017
0.0111
building
0.0089
0.9793
0.0062
0.0025
0.0000
0.0031
low_veg
0.0171
0.0080
0.9507
0.0211
0.0000
0.0032
tree
0.0228
0.0042
0.2628
0.7049
0.0016
0.0038
car
0.0923
0.0002
0.0030
0.0242
0.8168
0.0635
clutter
0.1995
0.1720
0.2266
0.0133
0.0039
0.3847
Precision/Correctness
0.901
0.951
0.723
0.947
0.917
0.675
Recall/Completeness
0.852
0.979
0.951
0.705
0.817
0.385
F1
0.876
0.965
0.822
0.808
0.864
0.490

Overall accuracy 0.854

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
0.8817
0.0069
0.0919
0.0099
0.0001
0.0094
building
0.0073
0.9858
0.0027
0.0017
0.0000
0.0025
low_veg
0.0098
0.0033
0.9700
0.0149
0.0000
0.0020
tree
0.0177
0.0032
0.2436
0.7311
0.0013
0.0031
car
0.0152
0.0001
0.0010
0.0173
0.8944
0.0720
clutter
0.1915
0.1690
0.1925
0.0083
0.0048
0.4339
Precision/Correctness
0.929
0.968
0.745
0.966
0.944
0.709
Recall/Completeness
0.882
0.986
0.970
0.731
0.894
0.434
F1
0.905
0.977
0.843
0.832
0.919
0.538

Overall accuracy 0.878

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
0.9113
0.0060
0.0651
0.0146
0.0004
0.0027
building
0.0098
0.9730
0.0081
0.0045
0.0000
0.0046
low_veg
0.0264
0.0042
0.9134
0.0548
0.0000
0.0010
tree
0.0316
0.0024
0.0879
0.8751
0.0026
0.0004
car
0.1202
0.0002
0.0029
0.0158
0.8404
0.0204
clutter
0.4349
0.0934
0.2864
0.0458
0.0026
0.1370
Precision/Correctness
0.865
0.973
0.714
0.866
0.967
0.756
Recall/Completeness
0.911
0.973
0.913
0.875
0.840
0.137
F1
0.887
0.973
0.801
0.871
0.899
0.232

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
0.9283
0.0035
0.0561
0.0097
0.0001
0.0023
building
0.0057
0.9836
0.0032
0.0034
0.0000
0.0041
low_veg
0.0180
0.0011
0.9339
0.0463
0.0000
0.0008
tree
0.0193
0.0012
0.0686
0.9088
0.0018
0.0002
car
0.0129
0.0000
0.0002
0.0102
0.9541
0.0226
clutter
0.4526
0.0812
0.2772
0.0390
0.0027
0.1472
Precision/Correctness
0.890
0.982
0.747
0.898
0.976
0.771
Recall/Completeness
0.928
0.984
0.934
0.909
0.954
0.147
F1
0.909
0.983
0.830
0.903
0.965
0.247

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
0.8643
0.0175
0.1012
0.0085
0.0010
0.0074
building
0.0093
0.9739
0.0067
0.0037
0.0000
0.0063
low_veg
0.0483
0.0126
0.9010
0.0228
0.0000
0.0152
tree
0.0361
0.0049
0.2249
0.7258
0.0022
0.0061
car
0.1012
0.0025
0.0088
0.0136
0.8298
0.0441
clutter
0.3183
0.0370
0.3179
0.0075
0.0005
0.3188
Precision/Correctness
0.804
0.942
0.672
0.937
0.962
0.781
Recall/Completeness
0.864
0.974
0.901
0.726
0.830
0.319
F1
0.833
0.958
0.770
0.818
0.891
0.453

Overall accuracy 0.814

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
0.8800
0.0130
0.0948
0.0057
0.0000
0.0066
building
0.0069
0.9812
0.0038
0.0027
0.0000
0.0054
low_veg
0.0380
0.0077
0.9209
0.0181
0.0000
0.0153
tree
0.0277
0.0035
0.2012
0.7603
0.0017
0.0057
car
0.0156
0.0020
0.0079
0.0096
0.9166
0.0483
clutter
0.3247
0.0332
0.3073
0.0058
0.0003
0.3287
Precision/Correctness
0.822
0.958
0.687
0.955
0.980
0.801
Recall/Completeness
0.880
0.981
0.921
0.760
0.917
0.329
F1
0.850
0.970
0.787
0.846
0.947
0.466

Overall accuracy 0.833

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
0.9125
0.0111
0.0693
0.0026
0.0010
0.0036
building
0.0032
0.9864
0.0060
0.0013
0.0000
0.0032
low_veg
0.0362
0.0170
0.9318
0.0099
0.0001
0.0050
tree
0.0833
0.0083
0.2381
0.6579
0.0076
0.0048
car
0.1195
0.0006
0.0066
0.0050
0.8305
0.0378
clutter
0.2677
0.2077
0.2710
0.0053
0.0030
0.2452
Precision/Correctness
0.917
0.957
0.705
0.964
0.933
0.609
Recall/Completeness
0.912
0.986
0.932
0.658
0.831
0.245
F1
0.915
0.972
0.802
0.782
0.879
0.350

Overall accuracy 0.885

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
0.9284
0.0051
0.0624
0.0015
0.0001
0.0026
building
0.0012
0.9920
0.0033
0.0008
0.0000
0.0027
low_veg
0.0201
0.0091
0.9592
0.0074
0.0000
0.0042
tree
0.0671
0.0066
0.2128
0.7031
0.0062
0.0043
car
0.0201
0.0001
0.0009
0.0026
0.9330
0.0433
clutter
0.2684
0.2054
0.2719
0.0041
0.0025
0.2477
Precision/Correctness
0.944
0.973
0.728
0.977
0.954
0.609
Recall/Completeness
0.928
0.992
0.959
0.703
0.933
0.248
F1
0.936
0.983
0.828
0.818
0.943
0.352

Overall accuracy 0.909

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
0.9020
0.0107
0.0687
0.0143
0.0007
0.0035
building
0.0087
0.9777
0.0066
0.0024
0.0000
0.0046
low_veg
0.0301
0.0073
0.9211
0.0391
0.0000
0.0025
tree
0.0530
0.0045
0.1964
0.7415
0.0025
0.0020
car
0.1272
0.0003
0.0010
0.0056
0.8366
0.0293
clutter
0.2085
0.2027
0.2472
0.0124
0.0427
0.2865
Precision/Correctness
0.924
0.959
0.755
0.872
0.917
0.699
Recall/Completeness
0.902
0.978
0.921
0.742
0.837
0.286
F1
0.913
0.968
0.830
0.802
0.875
0.406

Overall accuracy 0.886

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
0.9206
0.0058
0.0610
0.0101
0.0001
0.0025
building
0.0054
0.9845
0.0040
0.0018
0.0000
0.0043
low_veg
0.0260
0.0034
0.9354
0.0336
0.0000
0.0017
tree
0.0377
0.0029
0.1789
0.7767
0.0020
0.0017
car
0.0187
0.0001
0.0001
0.0038
0.9450
0.0323
clutter
0.1831
0.1838
0.2395
0.0115
0.0532
0.3290
Precision/Correctness
0.950
0.975
0.784
0.898
0.926
0.713
Recall/Completeness
0.921
0.985
0.935
0.777
0.945
0.329
F1
0.935
0.980
0.853
0.833
0.936
0.450

Overall accuracy 0.910

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_5_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9278
0.0039
0.0437
0.0172
0.0004
0.0071
building
0.0137
0.9765
0.0051
0.0015
0.0001
0.0031
low_veg
0.0452
0.0040
0.8991
0.0446
0.0001
0.0070
tree
0.0550
0.0061
0.1370
0.7964
0.0038
0.0017
car
0.1292
0.0000
0.0112
0.0090
0.8184
0.0322
clutter
0.2130
0.1632
0.3403
0.0232
0.0012
0.2591
Precision/Correctness
0.859
0.942
0.749
0.896
0.947
0.792
Recall/Completeness
0.928
0.976
0.899
0.796
0.818
0.259
F1
0.892
0.959
0.817
0.844
0.878
0.390

Overall accuracy 0.859

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_potsdam_5_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9445
0.0014
0.0350
0.0128
0.0001
0.0061
building
0.0116
0.9821
0.0026
0.0011
0.0001
0.0026
low_veg
0.0327
0.0017
0.9194
0.0405
0.0000
0.0057
tree
0.0435
0.0051
0.1234
0.8237
0.0029
0.0013
car
0.0193
0.0000
0.0101
0.0060
0.9299
0.0346
clutter
0.2007
0.1364
0.3553
0.0208
0.0011
0.2858
Precision/Correctness
0.889
0.960
0.769
0.916
0.960
0.818
Recall/Completeness
0.945
0.982
0.919
0.824
0.930
0.286
F1
0.916
0.971
0.838
0.867
0.945
0.424

Overall accuracy 0.883

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
0.9085
0.0059
0.0602
0.0119
0.0007
0.0128
building
0.0153
0.9671
0.0071
0.0040
0.0003
0.0062
low_veg
0.0303
0.0037
0.9091
0.0511
0.0001
0.0057
tree
0.0438
0.0038
0.1398
0.8088
0.0017
0.0022
car
0.1344
0.0012
0.0046
0.0095
0.8197
0.0307
clutter
0.3241
0.1216
0.2194
0.0089
0.0011
0.3249
Precision/Correctness
0.902
0.953
0.693
0.871
0.972
0.728
Recall/Completeness
0.908
0.967
0.909
0.809
0.820
0.325
F1
0.905
0.960
0.787
0.839
0.889
0.449

Overall accuracy 0.868

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
0.9219
0.0027
0.0540
0.0085
0.0003
0.0125
building
0.0116
0.9755
0.0045
0.0029
0.0002
0.0053
low_veg
0.0176
0.0009
0.9316
0.0447
0.0000
0.0051
tree
0.0326
0.0022
0.1251
0.8369
0.0012
0.0020
car
0.0195
0.0010
0.0027
0.0074
0.9349
0.0346
clutter
0.3318
0.0910
0.2160
0.0072
0.0010
0.3530
Precision/Correctness
0.922
0.971
0.713
0.899
0.980
0.741
Recall/Completeness
0.922
0.976
0.932
0.837
0.935
0.353
F1
0.922
0.973
0.808
0.867
0.957
0.478

Overall accuracy 0.890

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
0.9426
0.0040
0.0319
0.0143
0.0006
0.0067
building
0.0355
0.9337
0.0182
0.0029
0.0001
0.0095
low_veg
0.0136
0.0014
0.9720
0.0118
0.0000
0.0011
tree
0.0660
0.0026
0.1519
0.7767
0.0018
0.0010
car
0.0858
0.0056
0.0006
0.0028
0.8861
0.0190
clutter
0.2616
0.0322
0.4691
0.0120
0.0023
0.2229
Precision/Correctness
0.932
0.987
0.839
0.910
0.955
0.458
Recall/Completeness
0.943
0.934
0.972
0.777
0.886
0.223
F1
0.937
0.960
0.901
0.838
0.919
0.300

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: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9565
0.0024
0.0265
0.0093
0.0000
0.0052
building
0.0310
0.9444
0.0133
0.0027
0.0001
0.0086
low_veg
0.0075
0.0004
0.9828
0.0086
0.0000
0.0007
tree
0.0479
0.0018
0.1364
0.8116
0.0014
0.0008
car
0.0075
0.0071
0.0003
0.0014
0.9632
0.0205
clutter
0.2535
0.0312
0.4916
0.0102
0.0018
0.2117
Precision/Correctness
0.948
0.992
0.859
0.936
0.976
0.459
Recall/Completeness
0.956
0.944
0.983
0.812
0.963
0.212
F1
0.952
0.968
0.917
0.870
0.970
0.290

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
0.9506
0.0033
0.0353
0.0068
0.0020
0.0020
building
0.0296
0.9540
0.0077
0.0011
0.0003
0.0073
low_veg
0.0345
0.0047
0.9204
0.0365
0.0002
0.0037
tree
0.0493
0.0084
0.1237
0.8120
0.0058
0.0008
car
0.0789
0.0002
0.0019
0.0059
0.8982
0.0149
clutter
0.2417
0.0677
0.3260
0.0109
0.0114
0.3422
Precision/Correctness
0.932
0.985
0.800
0.895
0.943
0.661
Recall/Completeness
0.951
0.954
0.920
0.812
0.898
0.342
F1
0.941
0.969
0.856
0.851
0.920
0.451

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_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9632
0.0011
0.0294
0.0047
0.0000
0.0015
building
0.0248
0.9629
0.0047
0.0008
0.0002
0.0066
low_veg
0.0224
0.0019
0.9419
0.0312
0.0000
0.0026
tree
0.0369
0.0071
0.1066
0.8442
0.0046
0.0007
car
0.0063
0.0000
0.0003
0.0035
0.9748
0.0150
clutter
0.2211
0.0723
0.3280
0.0080
0.0112
0.3593
Precision/Correctness
0.951
0.991
0.824
0.919
0.970
0.683
Recall/Completeness
0.963
0.963
0.942
0.844
0.975
0.359
F1
0.957
0.977
0.879
0.880
0.972
0.471

Overall accuracy 0.936

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
0.9406
0.0020
0.0434
0.0050
0.0022
0.0067
building
0.0205
0.9496
0.0220
0.0012
0.0001
0.0066
low_veg
0.0303
0.0013
0.9425
0.0203
0.0004
0.0052
tree
0.0494
0.0079
0.2375
0.6963
0.0060
0.0029
car
0.0911
0.0004
0.0042
0.0079
0.8748
0.0217
clutter
0.1815
0.0090
0.4092
0.0023
0.0042
0.3938
Precision/Correctness
0.921
0.988
0.628
0.951
0.927
0.798
Recall/Completeness
0.941
0.950
0.943
0.696
0.875
0.394
F1
0.931
0.968
0.754
0.804
0.900
0.527

Overall accuracy 0.870

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
0.9557
0.0006
0.0350
0.0028
0.0001
0.0058
building
0.0155
0.9592
0.0185
0.0008
0.0000
0.0061
low_veg
0.0165
0.0001
0.9635
0.0153
0.0000
0.0045
tree
0.0391
0.0068
0.2176
0.7290
0.0050
0.0025
car
0.0143
0.0002
0.0005
0.0059
0.9546
0.0245
clutter
0.1758
0.0078
0.4086
0.0015
0.0044
0.4018
Precision/Correctness
0.940
0.992
0.640
0.969
0.956
0.819
Recall/Completeness
0.956
0.959
0.964
0.729
0.955
0.402
F1
0.948
0.975
0.769
0.832
0.955
0.539

Overall accuracy 0.889

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
0.9192
0.0229
0.0238
0.0237
0.0008
0.0095
building
0.0144
0.9539
0.0074
0.0023
0.0000
0.0219
low_veg
0.0385
0.0058
0.8524
0.1005
0.0001
0.0026
tree
0.0176
0.0046
0.1597
0.8145
0.0007
0.0028
car
0.0702
0.0004
0.0005
0.0225
0.8877
0.0186
clutter
0.2302
0.0106
0.4075
0.0016
0.0046
0.3454
Precision/Correctness
0.938
0.941
0.579
0.735
0.933
0.750
Recall/Completeness
0.919
0.954
0.852
0.815
0.888
0.345
F1
0.928
0.948
0.689
0.772
0.910
0.473

Overall accuracy 0.860

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
0.9331
0.0201
0.0208
0.0170
0.0001
0.0089
building
0.0117
0.9597
0.0060
0.0017
0.0000
0.0209
low_veg
0.0301
0.0027
0.8722
0.0927
0.0001
0.0022
tree
0.0097
0.0028
0.1508
0.8333
0.0004
0.0029
car
0.0058
0.0003
0.0000
0.0159
0.9583
0.0198
clutter
0.2216
0.0092
0.4137
0.0013
0.0044
0.3497
Precision/Correctness
0.946
0.951
0.580
0.771
0.954
0.765
Recall/Completeness
0.933
0.960
0.872
0.833
0.958
0.350
F1
0.939
0.955
0.696
0.801
0.956
0.480

Overall accuracy 0.873

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image