Detailled semantic labeling (2D) result


Name Y. Liu et al.
Affiliation Chinese Academy of Sciences, China
Abbreviation CASIA
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
0.933
0.014
0.036
0.015
0.002
0.000
building
0.018
0.972
0.007
0.003
0.000
0.000
low_veg
0.038
0.008
0.903
0.051
0.000
0.000
tree
0.033
0.005
0.134
0.825
0.002
0.000
car
0.072
0.005
0.002
0.013
0.907
0.000
clutter
0.418
0.143
0.386
0.028
0.025
0.000
Precision/Correctness
0.884
0.944
0.780
0.889
0.884
-nan
Recall/Completeness
0.933
0.972
0.903
0.825
0.907
0.000
F1
0.908
0.958
0.837
0.855
0.896
-nan

Overall accuracy 0.874

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
0.951
0.010
0.028
0.010
0.000
0.000
building
0.014
0.979
0.004
0.002
0.000
0.000
low_veg
0.027
0.004
0.925
0.043
0.000
0.000
tree
0.024
0.004
0.117
0.853
0.002
0.000
car
0.009
0.005
0.001
0.008
0.976
0.000
clutter
0.424
0.131
0.395
0.025
0.026
0.000
Precision/Correctness
0.905
0.959
0.803
0.913
0.910
-nan
Recall/Completeness
0.951
0.979
0.925
0.853
0.976
0.000
F1
0.928
0.969
0.860
0.882
0.942
-nan

Overall accuracy 0.896
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9073
0.0202
0.0615
0.0086
0.0025
0.0000
building
0.0138
0.9671
0.0162
0.0030
0.0000
0.0000
low_veg
0.0291
0.0097
0.9324
0.0286
0.0002
0.0000
tree
0.0380
0.0115
0.1504
0.7983
0.0018
0.0000
car
0.0752
0.0157
0.0038
0.0066
0.8987
0.0000
clutter
0.3036
0.3134
0.2827
0.0258
0.0744
0.0000
Precision/Correctness
0.892
0.914
0.828
0.940
0.805
-nan
Recall/Completeness
0.907
0.967
0.932
0.798
0.899
0.000
F1
0.900
0.940
0.877
0.863
0.849
-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_2_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9284
0.0145
0.0518
0.0049
0.0004
0.0000
building
0.0106
0.9753
0.0119
0.0022
0.0000
0.0000
low_veg
0.0220
0.0058
0.9514
0.0207
0.0001
0.0000
tree
0.0287
0.0094
0.1263
0.8343
0.0013
0.0000
car
0.0185
0.0181
0.0020
0.0022
0.9591
0.0000
clutter
0.2937
0.3423
0.2539
0.0182
0.0918
0.0000
Precision/Correctness
0.920
0.937
0.857
0.960
0.838
-nan
Recall/Completeness
0.928
0.975
0.951
0.834
0.959
0.000
F1
0.924
0.956
0.902
0.893
0.894
-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_2_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9105
0.0085
0.0722
0.0076
0.0013
0.0000
building
0.0024
0.9868
0.0083
0.0026
0.0000
0.0000
low_veg
0.0118
0.0038
0.9653
0.0190
0.0000
0.0000
tree
0.0197
0.0044
0.2180
0.7571
0.0009
0.0000
car
0.0610
0.0019
0.0024
0.0079
0.9267
0.0000
clutter
0.1558
0.3920
0.3235
0.0406
0.0881
0.0000
Precision/Correctness
0.915
0.921
0.815
0.966
0.869
-nan
Recall/Completeness
0.911
0.987
0.965
0.757
0.927
0.000
F1
0.913
0.953
0.884
0.849
0.897
-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_2_14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9396
0.0052
0.0509
0.0042
0.0000
0.0000
building
0.0006
0.9943
0.0035
0.0016
0.0000
0.0000
low_veg
0.0075
0.0018
0.9761
0.0146
0.0000
0.0000
tree
0.0153
0.0034
0.1988
0.7819
0.0006
0.0000
car
0.0058
0.0006
0.0001
0.0039
0.9896
0.0000
clutter
0.0951
0.4785
0.2862
0.0413
0.0990
0.0000
Precision/Correctness
0.940
0.946
0.835
0.976
0.908
-nan
Recall/Completeness
0.940
0.994
0.976
0.782
0.990
0.000
F1
0.940
0.969
0.900
0.868
0.947
-nan

Overall accuracy 0.900

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
0.9087
0.0210
0.0518
0.0157
0.0028
0.0000
building
0.0065
0.9840
0.0069
0.0027
0.0000
0.0000
low_veg
0.0480
0.0178
0.8933
0.0408
0.0002
0.0000
tree
0.0280
0.0068
0.1066
0.8561
0.0024
0.0000
car
0.0699
0.0084
0.0009
0.0132
0.9077
0.0000
clutter
0.4026
0.2021
0.3534
0.0216
0.0204
0.0000
Precision/Correctness
0.884
0.939
0.812
0.930
0.897
-nan
Recall/Completeness
0.909
0.984
0.893
0.856
0.908
0.000
F1
0.896
0.961
0.851
0.892
0.902
-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_3_13_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9348
0.0138
0.0417
0.0095
0.0001
0.0000
building
0.0043
0.9905
0.0036
0.0017
0.0000
0.0000
low_veg
0.0372
0.0088
0.9230
0.0310
0.0000
0.0000
tree
0.0196
0.0052
0.0875
0.8861
0.0017
0.0000
car
0.0084
0.0095
0.0002
0.0075
0.9744
0.0000
clutter
0.4356
0.2067
0.3172
0.0162
0.0243
0.0000
Precision/Correctness
0.913
0.960
0.846
0.953
0.938
-nan
Recall/Completeness
0.935
0.990
0.923
0.886
0.974
0.000
F1
0.924
0.975
0.883
0.919
0.956
-nan

Overall accuracy 0.918

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_potsdam_3_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9001
0.0199
0.0604
0.0168
0.0028
0.0000
building
0.0101
0.9829
0.0048
0.0023
0.0000
0.0000
low_veg
0.0234
0.0092
0.9379
0.0294
0.0001
0.0000
tree
0.0202
0.0043
0.2091
0.7650
0.0014
0.0000
car
0.0660
0.0073
0.0011
0.0235
0.9022
0.0000
clutter
0.2694
0.2283
0.3474
0.1429
0.0121
0.0000
Precision/Correctness
0.893
0.936
0.765
0.922
0.889
-nan
Recall/Completeness
0.900
0.983
0.938
0.765
0.902
0.000
F1
0.896
0.959
0.842
0.836
0.896
-nan

Overall accuracy 0.867

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
0.9289
0.0141
0.0458
0.0110
0.0002
0.0000
building
0.0080
0.9891
0.0014
0.0014
0.0000
0.0000
low_veg
0.0142
0.0042
0.9604
0.0213
0.0000
0.0000
tree
0.0149
0.0032
0.1904
0.7904
0.0012
0.0000
car
0.0100
0.0083
0.0001
0.0148
0.9669
0.0000
clutter
0.2613
0.2316
0.3273
0.1663
0.0135
0.0000
Precision/Correctness
0.922
0.954
0.790
0.943
0.930
-nan
Recall/Completeness
0.929
0.989
0.960
0.790
0.967
0.000
F1
0.925
0.971
0.867
0.860
0.948
-nan

Overall accuracy 0.891

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
0.9419
0.0071
0.0285
0.0213
0.0012
0.0000
building
0.0154
0.9723
0.0085
0.0037
0.0001
0.0000
low_veg
0.0453
0.0057
0.8604
0.0886
0.0001
0.0000
tree
0.0267
0.0018
0.0519
0.9180
0.0017
0.0000
car
0.0758
0.0003
0.0015
0.0232
0.8992
0.0000
clutter
0.6099
0.1103
0.1929
0.0675
0.0193
0.0000
Precision/Correctness
0.831
0.968
0.800
0.821
0.910
-nan
Recall/Completeness
0.942
0.972
0.860
0.918
0.899
0.000
F1
0.883
0.970
0.829
0.867
0.905
-nan

Overall accuracy 0.872

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
0.9592
0.0040
0.0217
0.0149
0.0003
0.0000
building
0.0105
0.9835
0.0036
0.0024
0.0001
0.0000
low_veg
0.0342
0.0013
0.8862
0.0783
0.0000
0.0000
tree
0.0153
0.0008
0.0373
0.9455
0.0010
0.0000
car
0.0052
0.0000
0.0001
0.0145
0.9802
0.0000
clutter
0.6493
0.0984
0.1718
0.0591
0.0214
0.0000
Precision/Correctness
0.854
0.979
0.842
0.857
0.919
-nan
Recall/Completeness
0.959
0.983
0.886
0.946
0.980
0.000
F1
0.903
0.981
0.864
0.899
0.949
-nan

Overall accuracy 0.898

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
0.9003
0.0187
0.0649
0.0131
0.0030
0.0000
building
0.0165
0.9724
0.0053
0.0054
0.0004
0.0000
low_veg
0.0752
0.0125
0.8471
0.0649
0.0003
0.0000
tree
0.0290
0.0049
0.1351
0.8287
0.0023
0.0000
car
0.0675
0.0066
0.0019
0.0211
0.9029
0.0000
clutter
0.5086
0.0569
0.4154
0.0147
0.0044
0.0000
Precision/Correctness
0.746
0.932
0.698
0.880
0.920
-nan
Recall/Completeness
0.900
0.972
0.847
0.829
0.903
0.000
F1
0.816
0.952
0.765
0.854
0.911
-nan

Overall accuracy 0.800

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
0.9172
0.0139
0.0604
0.0082
0.0003
0.0000
building
0.0132
0.9797
0.0025
0.0042
0.0004
0.0000
low_veg
0.0635
0.0074
0.8703
0.0586
0.0001
0.0000
tree
0.0204
0.0036
0.1143
0.8600
0.0017
0.0000
car
0.0188
0.0066
0.0014
0.0145
0.9587
0.0000
clutter
0.5205
0.0531
0.4107
0.0120
0.0037
0.0000
Precision/Correctness
0.758
0.948
0.712
0.903
0.954
-nan
Recall/Completeness
0.917
0.980
0.870
0.860
0.959
0.000
F1
0.830
0.964
0.783
0.881
0.956
-nan

Overall accuracy 0.817

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
0.9349
0.0246
0.0331
0.0054
0.0021
0.0000
building
0.0093
0.9823
0.0061
0.0023
0.0001
0.0000
low_veg
0.0570
0.0160
0.8976
0.0290
0.0004
0.0000
tree
0.0627
0.0055
0.1287
0.7968
0.0062
0.0000
car
0.0798
0.0104
0.0016
0.0068
0.9015
0.0000
clutter
0.4352
0.2627
0.2496
0.0178
0.0348
0.0000
Precision/Correctness
0.902
0.940
0.799
0.928
0.891
-nan
Recall/Completeness
0.935
0.982
0.898
0.797
0.901
0.000
F1
0.918
0.961
0.846
0.857
0.896
-nan

Overall accuracy 0.898

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
0.9497
0.0194
0.0276
0.0032
0.0002
0.0000
building
0.0067
0.9884
0.0033
0.0016
0.0001
0.0000
low_veg
0.0365
0.0083
0.9316
0.0236
0.0000
0.0000
tree
0.0454
0.0039
0.1051
0.8412
0.0045
0.0000
car
0.0103
0.0119
0.0002
0.0027
0.9749
0.0000
clutter
0.4443
0.2693
0.2350
0.0143
0.0372
0.0000
Precision/Correctness
0.930
0.955
0.832
0.950
0.923
-nan
Recall/Completeness
0.950
0.988
0.932
0.841
0.975
0.000
F1
0.940
0.972
0.879
0.892
0.948
-nan

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_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9367
0.0105
0.0314
0.0200
0.0014
0.0000
building
0.0123
0.9801
0.0046
0.0031
0.0000
0.0000
low_veg
0.0529
0.0098
0.8698
0.0675
0.0000
0.0000
tree
0.0469
0.0041
0.1039
0.8427
0.0025
0.0000
car
0.0887
0.0038
0.0008
0.0085
0.8982
0.0000
clutter
0.3602
0.2983
0.2454
0.0156
0.0805
0.0000
Precision/Correctness
0.904
0.947
0.837
0.831
0.872
-nan
Recall/Completeness
0.937
0.980
0.870
0.843
0.898
0.000
F1
0.920
0.963
0.853
0.837
0.885
-nan

Overall accuracy 0.893

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
0.9541
0.0055
0.0259
0.0144
0.0001
0.0000
building
0.0084
0.9866
0.0027
0.0024
0.0000
0.0000
low_veg
0.0459
0.0054
0.8900
0.0587
0.0000
0.0000
tree
0.0313
0.0024
0.0889
0.8754
0.0020
0.0000
car
0.0092
0.0044
0.0001
0.0054
0.9809
0.0000
clutter
0.3337
0.3052
0.2466
0.0139
0.1005
0.0000
Precision/Correctness
0.931
0.964
0.866
0.861
0.882
-nan
Recall/Completeness
0.954
0.987
0.890
0.875
0.981
0.000
F1
0.942
0.975
0.878
0.868
0.929
-nan

Overall accuracy 0.918

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
0.9438
0.0064
0.0249
0.0237
0.0011
0.0000
building
0.0070
0.9864
0.0048
0.0017
0.0001
0.0000
low_veg
0.0499
0.0087
0.8613
0.0798
0.0002
0.0000
tree
0.0431
0.0050
0.0805
0.8683
0.0031
0.0000
car
0.0912
0.0109
0.0037
0.0149
0.8793
0.0000
clutter
0.3312
0.2186
0.3907
0.0369
0.0226
0.0000
Precision/Correctness
0.846
0.922
0.777
0.852
0.870
-nan
Recall/Completeness
0.944
0.986
0.861
0.868
0.879
0.000
F1
0.892
0.953
0.817
0.860
0.875
-nan

Overall accuracy 0.852

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
0.9614
0.0037
0.0172
0.0175
0.0002
0.0000
building
0.0053
0.9912
0.0023
0.0012
0.0000
0.0000
low_veg
0.0358
0.0058
0.8843
0.0741
0.0000
0.0000
tree
0.0324
0.0038
0.0701
0.8914
0.0023
0.0000
car
0.0098
0.0122
0.0014
0.0084
0.9682
0.0000
clutter
0.3298
0.1983
0.4128
0.0343
0.0248
0.0000
Precision/Correctness
0.874
0.941
0.798
0.876
0.879
-nan
Recall/Completeness
0.961
0.991
0.884
0.891
0.968
0.000
F1
0.916
0.965
0.839
0.883
0.921
-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_5_15_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9157
0.0227
0.0424
0.0180
0.0013
0.0000
building
0.0129
0.9746
0.0080
0.0043
0.0003
0.0000
low_veg
0.0428
0.0050
0.8441
0.1079
0.0001
0.0000
tree
0.0296
0.0045
0.0695
0.8943
0.0021
0.0000
car
0.0965
0.0030
0.0029
0.0132
0.8844
0.0000
clutter
0.4088
0.1555
0.4072
0.0134
0.0151
0.0000
Precision/Correctness
0.893
0.919
0.685
0.806
0.924
-nan
Recall/Completeness
0.916
0.975
0.844
0.894
0.884
0.000
F1
0.904
0.946
0.756
0.848
0.904
-nan

Overall accuracy 0.853

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
0.9305
0.0190
0.0367
0.0133
0.0005
0.0000
building
0.0089
0.9820
0.0058
0.0031
0.0002
0.0000
low_veg
0.0294
0.0020
0.8679
0.1006
0.0001
0.0000
tree
0.0199
0.0025
0.0584
0.9177
0.0015
0.0000
car
0.0127
0.0032
0.0004
0.0105
0.9733
0.0000
clutter
0.4187
0.1248
0.4295
0.0112
0.0157
0.0000
Precision/Correctness
0.912
0.939
0.700
0.836
0.934
-nan
Recall/Completeness
0.931
0.982
0.868
0.918
0.973
0.000
F1
0.921
0.960
0.775
0.875
0.953
-nan

Overall accuracy 0.874

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
0.9579
0.0080
0.0199
0.0130
0.0012
0.0000
building
0.0366
0.9432
0.0189
0.0008
0.0005
0.0000
low_veg
0.0146
0.0035
0.9533
0.0285
0.0000
0.0000
tree
0.0597
0.0024
0.0730
0.8637
0.0012
0.0000
car
0.0662
0.0000
0.0008
0.0040
0.9290
0.0000
clutter
0.5015
0.0513
0.3962
0.0235
0.0275
0.0000
Precision/Correctness
0.922
0.978
0.887
0.893
0.897
-nan
Recall/Completeness
0.958
0.943
0.953
0.864
0.929
0.000
F1
0.940
0.960
0.919
0.878
0.913
-nan

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
0.9723
0.0056
0.0143
0.0077
0.0001
0.0000
building
0.0318
0.9535
0.0138
0.0004
0.0005
0.0000
low_veg
0.0081
0.0019
0.9670
0.0229
0.0000
0.0000
tree
0.0409
0.0014
0.0596
0.8971
0.0009
0.0000
car
0.0050
0.0000
0.0005
0.0017
0.9928
0.0000
clutter
0.4920
0.0490
0.4086
0.0217
0.0287
0.0000
Precision/Correctness
0.940
0.985
0.908
0.922
0.927
-nan
Recall/Completeness
0.972
0.954
0.967
0.897
0.993
0.000
F1
0.956
0.969
0.936
0.909
0.959
-nan

Overall accuracy 0.940

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
0.9542
0.0130
0.0199
0.0094
0.0035
0.0000
building
0.0369
0.9570
0.0051
0.0008
0.0002
0.0000
low_veg
0.0395
0.0078
0.8928
0.0594
0.0004
0.0000
tree
0.0404
0.0097
0.0688
0.8763
0.0048
0.0000
car
0.0519
0.0028
0.0018
0.0079
0.9356
0.0000
clutter
0.3824
0.3436
0.2309
0.0158
0.0273
0.0000
Precision/Correctness
0.922
0.950
0.866
0.859
0.919
-nan
Recall/Completeness
0.954
0.957
0.893
0.876
0.936
0.000
F1
0.938
0.953
0.879
0.867
0.927
-nan

Overall accuracy 0.915

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
0.9693
0.0093
0.0147
0.0064
0.0002
0.0000
building
0.0332
0.9644
0.0020
0.0004
0.0001
0.0000
low_veg
0.0239
0.0034
0.9202
0.0524
0.0000
0.0000
tree
0.0287
0.0083
0.0555
0.9039
0.0036
0.0000
car
0.0038
0.0021
0.0005
0.0045
0.9892
0.0000
clutter
0.3684
0.3810
0.2118
0.0119
0.0269
0.0000
Precision/Correctness
0.940
0.959
0.895
0.887
0.961
-nan
Recall/Completeness
0.969
0.964
0.920
0.904
0.989
0.000
F1
0.954
0.962
0.907
0.896
0.975
-nan

Overall accuracy 0.935

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_potsdam_6_15_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9434
0.0072
0.0366
0.0088
0.0040
0.0000
building
0.0124
0.9790
0.0060
0.0025
0.0001
0.0000
low_veg
0.0337
0.0016
0.9189
0.0446
0.0011
0.0000
tree
0.0369
0.0044
0.1490
0.8051
0.0046
0.0000
car
0.0582
0.0005
0.0048
0.0128
0.9237
0.0000
clutter
0.3728
0.0680
0.5143
0.0092
0.0357
0.0000
Precision/Correctness
0.904
0.967
0.663
0.916
0.861
-nan
Recall/Completeness
0.943
0.979
0.919
0.805
0.924
0.000
F1
0.923
0.973
0.770
0.857
0.891
-nan

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
0.9613
0.0046
0.0284
0.0055
0.0002
0.0000
building
0.0077
0.9887
0.0019
0.0017
0.0000
0.0000
low_veg
0.0193
0.0001
0.9439
0.0367
0.0001
0.0000
tree
0.0271
0.0030
0.1304
0.8359
0.0036
0.0000
car
0.0085
0.0001
0.0006
0.0095
0.9813
0.0000
clutter
0.3725
0.0692
0.5160
0.0073
0.0351
0.0000
Precision/Correctness
0.921
0.974
0.680
0.940
0.899
-nan
Recall/Completeness
0.961
0.989
0.944
0.836
0.981
0.000
F1
0.941
0.981
0.790
0.885
0.939
-nan

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
0.9525
0.0124
0.0096
0.0236
0.0020
0.0000
building
0.0464
0.9454
0.0048
0.0032
0.0002
0.0000
low_veg
0.0490
0.0065
0.7920
0.1523
0.0002
0.0000
tree
0.0240
0.0025
0.0926
0.8798
0.0011
0.0000
car
0.0511
0.0053
0.0008
0.0208
0.9220
0.0000
clutter
0.3764
0.0236
0.5713
0.0024
0.0263
0.0000
Precision/Correctness
0.901
0.958
0.553
0.707
0.783
-nan
Recall/Completeness
0.952
0.945
0.792
0.880
0.922
0.000
F1
0.926
0.952
0.651
0.784
0.847
-nan

Overall accuracy 0.843

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
0.9666
0.0090
0.0067
0.0168
0.0008
0.0000
building
0.0430
0.9509
0.0036
0.0023
0.0002
0.0000
low_veg
0.0375
0.0026
0.8148
0.1450
0.0000
0.0000
tree
0.0155
0.0013
0.0845
0.8982
0.0007
0.0000
car
0.0042
0.0054
0.0001
0.0132
0.9771
0.0000
clutter
0.3679
0.0223
0.5826
0.0019
0.0252
0.0000
Precision/Correctness
0.909
0.969
0.551
0.739
0.789
-nan
Recall/Completeness
0.967
0.951
0.815
0.898
0.977
0.000
F1
0.937
0.960
0.657
0.811
0.873
-nan

Overall accuracy 0.855

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image