Detailled semantic labeling (2D) result


Name S. Piramanayagam et al.
Affiliation RIT, USA
Abbreviation RIT
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
0.924
0.014
0.042
0.017
0.003
0.000
building
0.020
0.968
0.009
0.003
0.000
0.000
low_veg
0.047
0.008
0.885
0.059
0.000
0.000
tree
0.039
0.006
0.143
0.809
0.002
0.000
car
0.080
0.007
0.004
0.016
0.892
0.000
clutter
0.474
0.129
0.338
0.028
0.031
0.000
Precision/Correctness
0.867
0.946
0.771
0.871
0.858
-nan
Recall/Completeness
0.924
0.968
0.885
0.809
0.892
0.000
F1
0.894
0.957
0.824
0.839
0.875
-nan

Overall accuracy 0.864

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
0.942
0.010
0.035
0.012
0.000
0.000
building
0.016
0.976
0.005
0.003
0.000
0.000
low_veg
0.035
0.004
0.910
0.051
0.000
0.000
tree
0.030
0.004
0.128
0.836
0.002
0.000
car
0.019
0.008
0.002
0.011
0.960
0.000
clutter
0.487
0.117
0.341
0.024
0.032
0.000
Precision/Correctness
0.888
0.961
0.794
0.899
0.889
-nan
Recall/Completeness
0.942
0.976
0.910
0.836
0.960
0.000
F1
0.914
0.969
0.848
0.866
0.923
-nan

Overall accuracy 0.886
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9107
0.0149
0.0594
0.0123
0.0026
0.0000
building
0.0246
0.9521
0.0154
0.0071
0.0007
0.0000
low_veg
0.0349
0.0068
0.9184
0.0396
0.0003
0.0000
tree
0.0381
0.0106
0.1541
0.7957
0.0016
0.0000
car
0.1007
0.0085
0.0046
0.0092
0.8769
0.0000
clutter
0.3552
0.2340
0.2683
0.0485
0.0941
0.0000
Precision/Correctness
0.875
0.932
0.826
0.914
0.771
-nan
Recall/Completeness
0.911
0.952
0.918
0.796
0.877
0.000
F1
0.893
0.942
0.870
0.851
0.821
-nan

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
0.9325
0.0101
0.0498
0.0073
0.0003
0.0000
building
0.0201
0.9627
0.0109
0.0057
0.0006
0.0000
low_veg
0.0264
0.0033
0.9415
0.0287
0.0001
0.0000
tree
0.0295
0.0091
0.1339
0.8264
0.0011
0.0000
car
0.0458
0.0094
0.0028
0.0047
0.9372
0.0000
clutter
0.3441
0.2540
0.2418
0.0428
0.1173
0.0000
Precision/Correctness
0.905
0.952
0.853
0.940
0.803
-nan
Recall/Completeness
0.933
0.963
0.942
0.826
0.937
0.000
F1
0.918
0.957
0.895
0.880
0.865
-nan

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: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8940
0.0050
0.0858
0.0130
0.0022
0.0000
building
0.0106
0.9660
0.0158
0.0076
0.0000
0.0000
low_veg
0.0140
0.0032
0.9482
0.0345
0.0001
0.0000
tree
0.0193
0.0030
0.1979
0.7790
0.0009
0.0000
car
0.0726
0.0075
0.0076
0.0154
0.8969
0.0000
clutter
0.1510
0.3543
0.3833
0.0747
0.0368
0.0000
Precision/Correctness
0.906
0.937
0.820
0.942
0.871
-nan
Recall/Completeness
0.894
0.966
0.948
0.779
0.897
0.000
F1
0.900
0.951
0.879
0.853
0.884
-nan

Overall accuracy 0.877

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
0.9226
0.0021
0.0665
0.0078
0.0009
0.0000
building
0.0076
0.9784
0.0090
0.0049
0.0000
0.0000
low_veg
0.0093
0.0015
0.9621
0.0271
0.0000
0.0000
tree
0.0155
0.0023
0.1816
0.8000
0.0006
0.0000
car
0.0151
0.0084
0.0056
0.0093
0.9616
0.0000
clutter
0.0986
0.4188
0.3598
0.0799
0.0429
0.0000
Precision/Correctness
0.930
0.959
0.838
0.957
0.904
-nan
Recall/Completeness
0.923
0.978
0.962
0.800
0.962
0.000
F1
0.926
0.969
0.896
0.871
0.932
-nan

Overall accuracy 0.897

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
0.9050
0.0159
0.0550
0.0204
0.0037
0.0000
building
0.0139
0.9748
0.0075
0.0037
0.0000
0.0000
low_veg
0.0546
0.0163
0.8770
0.0519
0.0002
0.0000
tree
0.0331
0.0074
0.1182
0.8390
0.0022
0.0000
car
0.0728
0.0143
0.0049
0.0204
0.8876
0.0000
clutter
0.4167
0.1530
0.3689
0.0291
0.0323
0.0000
Precision/Correctness
0.867
0.947
0.796
0.910
0.871
-nan
Recall/Completeness
0.905
0.975
0.877
0.839
0.888
0.000
F1
0.886
0.961
0.835
0.873
0.879
-nan

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
0.9324
0.0099
0.0440
0.0134
0.0003
0.0000
building
0.0105
0.9835
0.0033
0.0025
0.0000
0.0000
low_veg
0.0427
0.0080
0.9082
0.0410
0.0001
0.0000
tree
0.0252
0.0059
0.1007
0.8666
0.0016
0.0000
car
0.0129
0.0170
0.0030
0.0124
0.9546
0.0000
clutter
0.4484
0.1511
0.3427
0.0194
0.0383
0.0000
Precision/Correctness
0.897
0.967
0.830
0.937
0.918
-nan
Recall/Completeness
0.932
0.984
0.908
0.867
0.955
0.000
F1
0.914
0.975
0.867
0.900
0.936
-nan

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_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8840
0.0152
0.0757
0.0216
0.0034
0.0000
building
0.0141
0.9748
0.0077
0.0034
0.0000
0.0000
low_veg
0.0332
0.0080
0.9092
0.0495
0.0001
0.0000
tree
0.0251
0.0050
0.1948
0.7739
0.0012
0.0000
car
0.0805
0.0216
0.0041
0.0362
0.8576
0.0000
clutter
0.4264
0.2169
0.2955
0.0431
0.0181
0.0000
Precision/Correctness
0.851
0.941
0.762
0.907
0.865
-nan
Recall/Completeness
0.884
0.975
0.909
0.774
0.858
0.000
F1
0.867
0.958
0.829
0.835
0.861
-nan

Overall accuracy 0.855

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
0.9126
0.0093
0.0632
0.0145
0.0004
0.0000
building
0.0117
0.9823
0.0037
0.0024
0.0000
0.0000
low_veg
0.0222
0.0033
0.9353
0.0392
0.0000
0.0000
tree
0.0198
0.0040
0.1785
0.7967
0.0009
0.0000
car
0.0252
0.0251
0.0013
0.0255
0.9229
0.0000
clutter
0.4588
0.2210
0.2655
0.0353
0.0194
0.0000
Precision/Correctness
0.880
0.959
0.787
0.933
0.914
-nan
Recall/Completeness
0.913
0.982
0.935
0.797
0.923
0.000
F1
0.896
0.970
0.855
0.859
0.918
-nan

Overall accuracy 0.881

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
0.9317
0.0079
0.0334
0.0244
0.0025
0.0000
building
0.0142
0.9739
0.0079
0.0039
0.0001
0.0000
low_veg
0.0635
0.0074
0.8288
0.1000
0.0003
0.0000
tree
0.0291
0.0044
0.0585
0.9063
0.0016
0.0000
car
0.0700
0.0060
0.0036
0.0246
0.8958
0.0000
clutter
0.5775
0.1179
0.2168
0.0644
0.0234
0.0000
Precision/Correctness
0.829
0.964
0.774
0.805
0.882
-nan
Recall/Completeness
0.932
0.974
0.829
0.906
0.896
0.000
F1
0.877
0.969
0.801
0.853
0.889
-nan

Overall accuracy 0.864

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_potsdam_4_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9502
0.0046
0.0273
0.0173
0.0006
0.0000
building
0.0084
0.9851
0.0036
0.0028
0.0001
0.0000
low_veg
0.0499
0.0026
0.8588
0.0887
0.0000
0.0000
tree
0.0181
0.0027
0.0445
0.9336
0.0010
0.0000
car
0.0094
0.0061
0.0002
0.0148
0.9695
0.0000
clutter
0.6045
0.1053
0.2091
0.0554
0.0256
0.0000
Precision/Correctness
0.855
0.976
0.810
0.842
0.901
-nan
Recall/Completeness
0.950
0.985
0.859
0.934
0.969
0.000
F1
0.900
0.980
0.834
0.886
0.934
-nan

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_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8900
0.0183
0.0733
0.0143
0.0041
0.0000
building
0.0130
0.9746
0.0062
0.0057
0.0006
0.0000
low_veg
0.0926
0.0130
0.8307
0.0631
0.0006
0.0000
tree
0.0376
0.0046
0.1499
0.8050
0.0030
0.0000
car
0.0687
0.0071
0.0037
0.0218
0.8987
0.0000
clutter
0.6240
0.0852
0.2677
0.0152
0.0079
0.0000
Precision/Correctness
0.706
0.920
0.719
0.877
0.885
-nan
Recall/Completeness
0.890
0.975
0.831
0.805
0.899
0.000
F1
0.788
0.947
0.771
0.839
0.892
-nan

Overall accuracy 0.789

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
0.9085
0.0137
0.0680
0.0093
0.0006
0.0000
building
0.0097
0.9818
0.0035
0.0045
0.0005
0.0000
low_veg
0.0794
0.0078
0.8581
0.0545
0.0002
0.0000
tree
0.0290
0.0031
0.1313
0.8341
0.0024
0.0000
car
0.0236
0.0074
0.0014
0.0149
0.9527
0.0000
clutter
0.6425
0.0839
0.2560
0.0109
0.0068
0.0000
Precision/Correctness
0.717
0.935
0.737
0.904
0.926
-nan
Recall/Completeness
0.908
0.982
0.858
0.834
0.953
0.000
F1
0.801
0.958
0.793
0.867
0.939
-nan

Overall accuracy 0.808

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
0.9217
0.0141
0.0547
0.0065
0.0030
0.0000
building
0.0066
0.9820
0.0086
0.0024
0.0003
0.0000
low_veg
0.0687
0.0192
0.8760
0.0355
0.0007
0.0000
tree
0.0801
0.0091
0.1559
0.7484
0.0064
0.0000
car
0.0929
0.0169
0.0030
0.0113
0.8759
0.0000
clutter
0.4173
0.2427
0.2665
0.0305
0.0430
0.0000
Precision/Correctness
0.893
0.948
0.745
0.906
0.864
-nan
Recall/Completeness
0.922
0.982
0.876
0.748
0.876
0.000
F1
0.907
0.965
0.805
0.820
0.870
-nan

Overall accuracy 0.883

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
0.9378
0.0085
0.0491
0.0041
0.0005
0.0000
building
0.0036
0.9884
0.0060
0.0018
0.0003
0.0000
low_veg
0.0463
0.0112
0.9142
0.0282
0.0002
0.0000
tree
0.0640
0.0075
0.1350
0.7886
0.0050
0.0000
car
0.0261
0.0194
0.0004
0.0059
0.9482
0.0000
clutter
0.4223
0.2507
0.2546
0.0270
0.0454
0.0000
Precision/Correctness
0.922
0.964
0.773
0.934
0.900
-nan
Recall/Completeness
0.938
0.988
0.914
0.789
0.948
0.000
F1
0.930
0.976
0.838
0.855
0.923
-nan

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
0.9309
0.0119
0.0323
0.0226
0.0023
0.0000
building
0.0153
0.9753
0.0052
0.0042
0.0001
0.0000
low_veg
0.0664
0.0098
0.8456
0.0781
0.0001
0.0000
tree
0.0476
0.0041
0.1159
0.8302
0.0022
0.0000
car
0.0924
0.0015
0.0013
0.0087
0.8961
0.0000
clutter
0.3810
0.2668
0.2283
0.0258
0.0981
0.0000
Precision/Correctness
0.892
0.949
0.827
0.806
0.844
-nan
Recall/Completeness
0.931
0.975
0.846
0.830
0.896
0.000
F1
0.911
0.962
0.836
0.818
0.869
-nan

Overall accuracy 0.884

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
0.9484
0.0070
0.0278
0.0166
0.0002
0.0000
building
0.0107
0.9826
0.0033
0.0033
0.0001
0.0000
low_veg
0.0585
0.0052
0.8669
0.0694
0.0000
0.0000
tree
0.0330
0.0028
0.1020
0.8604
0.0018
0.0000
car
0.0191
0.0014
0.0004
0.0054
0.9737
0.0000
clutter
0.3603
0.2659
0.2290
0.0238
0.1210
0.0000
Precision/Correctness
0.919
0.966
0.854
0.836
0.862
-nan
Recall/Completeness
0.948
0.983
0.867
0.860
0.974
0.000
F1
0.933
0.974
0.860
0.848
0.915
-nan

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_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9382
0.0054
0.0277
0.0267
0.0019
0.0000
building
0.0193
0.9736
0.0052
0.0017
0.0002
0.0000
low_veg
0.0595
0.0047
0.8496
0.0859
0.0004
0.0000
tree
0.0558
0.0064
0.1012
0.8331
0.0035
0.0000
car
0.0968
0.0127
0.0129
0.0189
0.8587
0.0000
clutter
0.3382
0.1974
0.3997
0.0473
0.0174
0.0000
Precision/Correctness
0.824
0.930
0.757
0.833
0.868
-nan
Recall/Completeness
0.938
0.974
0.850
0.833
0.859
0.000
F1
0.877
0.951
0.801
0.833
0.864
-nan

Overall accuracy 0.839

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
0.9558
0.0028
0.0206
0.0203
0.0004
0.0000
building
0.0167
0.9794
0.0026
0.0011
0.0001
0.0000
low_veg
0.0429
0.0019
0.8746
0.0806
0.0000
0.0000
tree
0.0449
0.0053
0.0915
0.8557
0.0027
0.0000
car
0.0212
0.0141
0.0105
0.0117
0.9425
0.0000
clutter
0.3366
0.1762
0.4227
0.0464
0.0181
0.0000
Precision/Correctness
0.852
0.948
0.777
0.857
0.891
-nan
Recall/Completeness
0.956
0.979
0.875
0.856
0.942
0.000
F1
0.901
0.963
0.823
0.856
0.916
-nan

Overall accuracy 0.862

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
0.9200
0.0101
0.0486
0.0194
0.0019
0.0000
building
0.0195
0.9626
0.0119
0.0052
0.0009
0.0000
low_veg
0.0556
0.0051
0.8268
0.1124
0.0001
0.0000
tree
0.0387
0.0035
0.0961
0.8596
0.0022
0.0000
car
0.1148
0.0061
0.0031
0.0144
0.8616
0.0000
clutter
0.4776
0.1193
0.3395
0.0200
0.0435
0.0000
Precision/Correctness
0.872
0.946
0.673
0.788
0.840
-nan
Recall/Completeness
0.920
0.963
0.827
0.860
0.862
0.000
F1
0.896
0.954
0.742
0.822
0.851
-nan

Overall accuracy 0.844

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
0.9346
0.0064
0.0436
0.0148
0.0006
0.0000
building
0.0144
0.9717
0.0092
0.0039
0.0009
0.0000
low_veg
0.0377
0.0021
0.8546
0.1056
0.0000
0.0000
tree
0.0286
0.0018
0.0853
0.8826
0.0017
0.0000
car
0.0346
0.0066
0.0003
0.0119
0.9466
0.0000
clutter
0.4931
0.0891
0.3522
0.0174
0.0481
0.0000
Precision/Correctness
0.893
0.965
0.690
0.818
0.842
-nan
Recall/Completeness
0.935
0.972
0.855
0.883
0.947
0.000
F1
0.913
0.968
0.763
0.849
0.891
-nan

Overall accuracy 0.866

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
0.9358
0.0220
0.0232
0.0174
0.0016
0.0000
building
0.0323
0.9510
0.0150
0.0013
0.0004
0.0000
low_veg
0.0221
0.0043
0.9373
0.0363
0.0000
0.0000
tree
0.0739
0.0034
0.1044
0.8170
0.0012
0.0000
car
0.0700
0.0011
0.0002
0.0049
0.9238
0.0000
clutter
0.4873
0.0675
0.3885
0.0302
0.0265
0.0000
Precision/Correctness
0.915
0.954
0.871
0.858
0.889
-nan
Recall/Completeness
0.936
0.951
0.937
0.817
0.924
0.000
F1
0.925
0.952
0.903
0.837
0.906
-nan

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: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9508
0.0194
0.0180
0.0116
0.0003
0.0000
building
0.0277
0.9612
0.0100
0.0007
0.0004
0.0000
low_veg
0.0135
0.0021
0.9533
0.0310
0.0000
0.0000
tree
0.0568
0.0021
0.0927
0.8475
0.0009
0.0000
car
0.0102
0.0013
0.0000
0.0015
0.9870
0.0000
clutter
0.4821
0.0659
0.3979
0.0286
0.0255
0.0000
Precision/Correctness
0.934
0.963
0.892
0.888
0.926
-nan
Recall/Completeness
0.951
0.961
0.953
0.848
0.987
0.000
F1
0.942
0.962
0.922
0.867
0.956
-nan

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
0.9567
0.0048
0.0240
0.0098
0.0047
0.0000
building
0.0436
0.9448
0.0095
0.0016
0.0005
0.0000
low_veg
0.0524
0.0060
0.8780
0.0629
0.0006
0.0000
tree
0.0482
0.0093
0.0976
0.8403
0.0047
0.0000
car
0.0620
0.0031
0.0031
0.0136
0.9181
0.0000
clutter
0.4247
0.2332
0.2743
0.0284
0.0394
0.0000
Precision/Correctness
0.909
0.969
0.829
0.841
0.894
-nan
Recall/Completeness
0.957
0.945
0.878
0.840
0.918
0.000
F1
0.932
0.957
0.853
0.840
0.906
-nan

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_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9718
0.0020
0.0189
0.0069
0.0005
0.0000
building
0.0379
0.9547
0.0059
0.0011
0.0004
0.0000
low_veg
0.0342
0.0027
0.9071
0.0560
0.0000
0.0000
tree
0.0367
0.0078
0.0845
0.8674
0.0035
0.0000
car
0.0118
0.0023
0.0008
0.0093
0.9757
0.0000
clutter
0.4137
0.2629
0.2635
0.0217
0.0382
0.0000
Precision/Correctness
0.929
0.976
0.856
0.872
0.944
-nan
Recall/Completeness
0.972
0.955
0.907
0.867
0.976
0.000
F1
0.950
0.965
0.881
0.869
0.959
-nan

Overall accuracy 0.927

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
0.9381
0.0050
0.0418
0.0103
0.0049
0.0000
building
0.0218
0.9659
0.0081
0.0031
0.0011
0.0000
low_veg
0.0517
0.0031
0.8985
0.0455
0.0012
0.0000
tree
0.0435
0.0048
0.1779
0.7693
0.0045
0.0000
car
0.0675
0.0012
0.0047
0.0111
0.9155
0.0000
clutter
0.3633
0.0465
0.5334
0.0161
0.0408
0.0000
Precision/Correctness
0.891
0.974
0.633
0.904
0.838
-nan
Recall/Completeness
0.938
0.966
0.899
0.769
0.915
0.000
F1
0.914
0.970
0.743
0.831
0.875
-nan

Overall accuracy 0.856

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
0.9559
0.0028
0.0340
0.0068
0.0005
0.0000
building
0.0159
0.9770
0.0040
0.0022
0.0010
0.0000
low_veg
0.0351
0.0013
0.9263
0.0372
0.0001
0.0000
tree
0.0343
0.0035
0.1599
0.7989
0.0035
0.0000
car
0.0136
0.0010
0.0011
0.0074
0.9769
0.0000
clutter
0.3632
0.0457
0.5377
0.0144
0.0390
0.0000
Precision/Correctness
0.910
0.981
0.647
0.930
0.880
-nan
Recall/Completeness
0.956
0.977
0.926
0.799
0.977
0.000
F1
0.932
0.979
0.762
0.859
0.926
-nan

Overall accuracy 0.876

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
0.9237
0.0327
0.0166
0.0248
0.0022
0.0000
building
0.0234
0.9682
0.0053
0.0026
0.0006
0.0000
low_veg
0.0552
0.0096
0.7966
0.1383
0.0002
0.0000
tree
0.0326
0.0066
0.1221
0.8381
0.0006
0.0000
car
0.0516
0.0076
0.0015
0.0291
0.9102
0.0000
clutter
0.5470
0.0212
0.4090
0.0033
0.0196
0.0000
Precision/Correctness
0.878
0.917
0.590
0.703
0.804
-nan
Recall/Completeness
0.924
0.968
0.797
0.838
0.910
0.000
F1
0.900
0.942
0.678
0.765
0.854
-nan

Overall accuracy 0.831

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
0.9382
0.0293
0.0141
0.0178
0.0007
0.0000
building
0.0203
0.9732
0.0041
0.0019
0.0005
0.0000
low_veg
0.0425
0.0057
0.8218
0.1300
0.0000
0.0000
tree
0.0236
0.0046
0.1143
0.8572
0.0003
0.0000
car
0.0064
0.0074
0.0004
0.0202
0.9656
0.0000
clutter
0.5439
0.0191
0.4157
0.0028
0.0184
0.0000
Precision/Correctness
0.886
0.928
0.592
0.738
0.826
-nan
Recall/Completeness
0.938
0.973
0.822
0.857
0.966
0.000
F1
0.911
0.950
0.688
0.793
0.890
-nan

Overall accuracy 0.843

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image