Detailled semantic labeling (2D) result


Name J. Sherrah (DFCN IRRG CRF)
Affiliation DST Group, Australia
Abbreviation DST_2
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
0.918
0.021
0.029
0.021
0.006
0.006
building
0.022
0.961
0.010
0.003
0.000
0.004
low_veg
0.060
0.013
0.847
0.074
0.000
0.006
tree
0.037
0.006
0.108
0.846
0.001
0.002
car
0.181
0.009
0.005
0.024
0.776
0.005
clutter
0.302
0.136
0.180
0.028
0.007
0.347
Precision/Correctness
0.874
0.932
0.823
0.855
0.866
0.781
Recall/Completeness
0.918
0.961
0.847
0.846
0.776
0.347
F1
0.896
0.946
0.835
0.850
0.818
0.480

Overall accuracy 0.872

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
0.937
0.017
0.023
0.015
0.003
0.005
building
0.018
0.969
0.008
0.002
0.000
0.003
low_veg
0.046
0.008
0.877
0.064
0.000
0.005
tree
0.027
0.004
0.094
0.872
0.001
0.002
car
0.095
0.010
0.002
0.017
0.870
0.005
clutter
0.294
0.122
0.167
0.024
0.007
0.386
Precision/Correctness
0.900
0.949
0.848
0.883
0.915
0.803
Recall/Completeness
0.937
0.969
0.877
0.872
0.870
0.386
F1
0.918
0.959
0.863
0.877
0.892
0.521

Overall accuracy 0.897
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9004
0.0248
0.0366
0.0202
0.0035
0.0145
building
0.0212
0.9554
0.0159
0.0044
0.0000
0.0030
low_veg
0.0468
0.0105
0.8865
0.0509
0.0001
0.0052
tree
0.0314
0.0110
0.1438
0.8099
0.0009
0.0030
car
0.1544
0.0371
0.0050
0.0211
0.7587
0.0238
clutter
0.1945
0.2080
0.1795
0.0456
0.0010
0.3714
Precision/Correctness
0.877
0.915
0.846
0.894
0.888
0.517
Recall/Completeness
0.900
0.955
0.887
0.810
0.759
0.371
F1
0.889
0.935
0.866
0.850
0.818
0.432

Overall accuracy 0.873

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
0.9253
0.0194
0.0275
0.0136
0.0012
0.0130
building
0.0160
0.9662
0.0119
0.0033
0.0000
0.0025
low_veg
0.0380
0.0058
0.9135
0.0385
0.0000
0.0042
tree
0.0232
0.0092
0.1234
0.8412
0.0005
0.0025
car
0.0746
0.0449
0.0014
0.0139
0.8378
0.0274
clutter
0.1574
0.2071
0.1423
0.0392
0.0010
0.4530
Precision/Correctness
0.908
0.939
0.874
0.923
0.946
0.528
Recall/Completeness
0.925
0.966
0.913
0.841
0.838
0.453
F1
0.917
0.952
0.893
0.880
0.889
0.488

Overall accuracy 0.902

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
0.9010
0.0075
0.0669
0.0183
0.0053
0.0011
building
0.0061
0.9814
0.0073
0.0037
0.0000
0.0015
low_veg
0.0196
0.0044
0.9219
0.0520
0.0001
0.0021
tree
0.0139
0.0041
0.1408
0.8393
0.0005
0.0014
car
0.1350
0.0084
0.0043
0.0210
0.8314
0.0000
clutter
0.1266
0.2554
0.2905
0.0693
0.0012
0.2569
Precision/Correctness
0.904
0.926
0.860
0.923
0.830
0.360
Recall/Completeness
0.901
0.981
0.922
0.839
0.831
0.257
F1
0.903
0.953
0.890
0.879
0.831
0.300

Overall accuracy 0.890

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
0.9304
0.0042
0.0498
0.0121
0.0027
0.0009
building
0.0029
0.9913
0.0025
0.0021
0.0000
0.0013
low_veg
0.0137
0.0023
0.9395
0.0425
0.0000
0.0019
tree
0.0101
0.0032
0.1258
0.8593
0.0003
0.0013
car
0.0610
0.0073
0.0002
0.0146
0.9169
0.0000
clutter
0.0776
0.2780
0.2598
0.0719
0.0006
0.3122
Precision/Correctness
0.932
0.952
0.879
0.940
0.893
0.348
Recall/Completeness
0.930
0.991
0.940
0.859
0.917
0.312
F1
0.931
0.971
0.908
0.898
0.905
0.329

Overall accuracy 0.911

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
0.8881
0.0228
0.0478
0.0280
0.0065
0.0068
building
0.0117
0.9760
0.0056
0.0046
0.0000
0.0021
low_veg
0.0641
0.0246
0.8401
0.0657
0.0002
0.0054
tree
0.0257
0.0061
0.0933
0.8726
0.0013
0.0010
car
0.1725
0.0038
0.0038
0.0357
0.7789
0.0052
clutter
0.3658
0.1542
0.3115
0.0311
0.0034
0.1340
Precision/Correctness
0.863
0.936
0.821
0.891
0.866
0.456
Recall/Completeness
0.888
0.976
0.840
0.873
0.779
0.134
F1
0.876
0.955
0.830
0.882
0.820
0.207

Overall accuracy 0.875

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
0.9161
0.0159
0.0391
0.0195
0.0027
0.0067
building
0.0080
0.9844
0.0027
0.0032
0.0000
0.0017
low_veg
0.0503
0.0145
0.8779
0.0526
0.0001
0.0047
tree
0.0181
0.0046
0.0778
0.8978
0.0009
0.0008
car
0.0835
0.0040
0.0016
0.0261
0.8786
0.0063
clutter
0.3831
0.1519
0.2848
0.0213
0.0038
0.1551
Precision/Correctness
0.898
0.958
0.853
0.921
0.925
0.454
Recall/Completeness
0.916
0.984
0.878
0.898
0.879
0.155
F1
0.907
0.971
0.865
0.909
0.901
0.231

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
0.8793
0.0254
0.0551
0.0311
0.0052
0.0038
building
0.0076
0.9823
0.0062
0.0029
0.0001
0.0009
low_veg
0.0357
0.0142
0.8821
0.0648
0.0000
0.0032
tree
0.0191
0.0036
0.1354
0.8382
0.0006
0.0031
car
0.1537
0.0160
0.0038
0.0484
0.7759
0.0021
clutter
0.2410
0.1892
0.2322
0.0392
0.0018
0.2966
Precision/Correctness
0.879
0.931
0.813
0.891
0.876
0.770
Recall/Completeness
0.879
0.982
0.882
0.838
0.776
0.297
F1
0.879
0.956
0.846
0.864
0.823
0.428

Overall accuracy 0.873

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
0.9100
0.0191
0.0424
0.0228
0.0020
0.0036
building
0.0049
0.9899
0.0026
0.0019
0.0001
0.0007
low_veg
0.0236
0.0083
0.9129
0.0528
0.0000
0.0024
tree
0.0141
0.0025
0.1187
0.8617
0.0004
0.0025
car
0.0695
0.0175
0.0019
0.0367
0.8724
0.0019
clutter
0.2271
0.1799
0.1999
0.0331
0.0017
0.3583
Precision/Correctness
0.914
0.952
0.842
0.917
0.933
0.804
Recall/Completeness
0.910
0.990
0.913
0.862
0.872
0.358
F1
0.912
0.970
0.876
0.889
0.902
0.496

Overall accuracy 0.902

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
0.9251
0.0124
0.0246
0.0291
0.0062
0.0027
building
0.0183
0.9674
0.0070
0.0048
0.0000
0.0025
low_veg
0.0713
0.0124
0.7997
0.1121
0.0003
0.0042
tree
0.0299
0.0037
0.0519
0.9129
0.0012
0.0006
car
0.1897
0.0049
0.0046
0.0370
0.7637
0.0002
clutter
0.5473
0.1314
0.1482
0.0771
0.0045
0.0915
Precision/Correctness
0.820
0.955
0.814
0.781
0.875
0.726
Recall/Completeness
0.925
0.967
0.800
0.913
0.764
0.091
F1
0.869
0.961
0.807
0.842
0.815
0.163

Overall accuracy 0.859

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
0.9453
0.0084
0.0195
0.0217
0.0029
0.0023
building
0.0125
0.9782
0.0037
0.0034
0.0000
0.0022
low_veg
0.0551
0.0058
0.8353
0.1000
0.0000
0.0038
tree
0.0191
0.0022
0.0400
0.9376
0.0007
0.0003
car
0.1023
0.0045
0.0012
0.0271
0.8648
0.0001
clutter
0.5694
0.1184
0.1342
0.0687
0.0051
0.1043
Precision/Correctness
0.849
0.969
0.850
0.819
0.923
0.757
Recall/Completeness
0.945
0.978
0.835
0.938
0.865
0.104
F1
0.895
0.974
0.842
0.874
0.893
0.183

Overall accuracy 0.888

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
0.8769
0.0246
0.0662
0.0187
0.0067
0.0070
building
0.0212
0.9661
0.0040
0.0053
0.0000
0.0035
low_veg
0.1010
0.0201
0.7871
0.0750
0.0003
0.0165
tree
0.0301
0.0078
0.1203
0.8377
0.0010
0.0031
car
0.1829
0.0104
0.0057
0.0359
0.7534
0.0117
clutter
0.4560
0.0580
0.1677
0.0119
0.0009
0.3055
Precision/Correctness
0.730
0.914
0.760
0.863
0.894
0.809
Recall/Completeness
0.877
0.966
0.787
0.838
0.753
0.305
F1
0.797
0.939
0.773
0.850
0.818
0.444

Overall accuracy 0.807

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
0.8976
0.0192
0.0610
0.0127
0.0031
0.0063
building
0.0180
0.9731
0.0019
0.0040
0.0000
0.0031
low_veg
0.0869
0.0139
0.8171
0.0657
0.0000
0.0163
tree
0.0218
0.0062
0.1026
0.8661
0.0006
0.0026
car
0.1050
0.0109
0.0020
0.0280
0.8413
0.0128
clutter
0.4665
0.0531
0.1557
0.0089
0.0006
0.3152
Precision/Correctness
0.748
0.933
0.782
0.891
0.943
0.827
Recall/Completeness
0.898
0.973
0.817
0.866
0.841
0.315
F1
0.816
0.953
0.799
0.878
0.889
0.456

Overall accuracy 0.830

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
0.9229
0.0287
0.0319
0.0081
0.0053
0.0030
building
0.0230
0.9618
0.0053
0.0017
0.0001
0.0082
low_veg
0.0760
0.0250
0.8578
0.0362
0.0009
0.0042
tree
0.0652
0.0084
0.1276
0.7923
0.0036
0.0029
car
0.1847
0.0212
0.0047
0.0190
0.7661
0.0043
clutter
0.3376
0.2985
0.1639
0.0232
0.0012
0.1757
Precision/Correctness
0.882
0.925
0.806
0.908
0.880
0.517
Recall/Completeness
0.923
0.962
0.858
0.792
0.766
0.176
F1
0.902
0.943
0.831
0.846
0.819
0.262

Overall accuracy 0.882

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
0.9416
0.0230
0.0262
0.0050
0.0022
0.0021
building
0.0200
0.9681
0.0027
0.0010
0.0001
0.0081
low_veg
0.0537
0.0149
0.9010
0.0271
0.0003
0.0030
tree
0.0481
0.0063
0.1088
0.8318
0.0025
0.0024
car
0.0899
0.0255
0.0010
0.0126
0.8662
0.0047
clutter
0.3378
0.3115
0.1403
0.0186
0.0008
0.1910
Precision/Correctness
0.913
0.943
0.839
0.938
0.935
0.519
Recall/Completeness
0.942
0.968
0.901
0.832
0.866
0.191
F1
0.927
0.956
0.869
0.882
0.899
0.279

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_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9252
0.0188
0.0221
0.0252
0.0041
0.0047
building
0.0154
0.9739
0.0039
0.0034
0.0000
0.0034
low_veg
0.0960
0.0165
0.7718
0.1117
0.0000
0.0039
tree
0.0592
0.0060
0.0784
0.8543
0.0017
0.0002
car
0.1954
0.0011
0.0014
0.0103
0.7882
0.0037
clutter
0.3030
0.2711
0.1802
0.0160
0.0379
0.1918
Precision/Correctness
0.873
0.936
0.860
0.770
0.878
0.645
Recall/Completeness
0.925
0.974
0.772
0.854
0.788
0.192
F1
0.898
0.955
0.814
0.810
0.831
0.296

Overall accuracy 0.873

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
0.9437
0.0130
0.0185
0.0193
0.0017
0.0038
building
0.0108
0.9808
0.0026
0.0025
0.0000
0.0033
low_veg
0.0887
0.0112
0.7954
0.1017
0.0000
0.0031
tree
0.0429
0.0041
0.0681
0.8834
0.0013
0.0002
car
0.0965
0.0011
0.0004
0.0066
0.8914
0.0039
clutter
0.2699
0.2626
0.1795
0.0129
0.0473
0.2279
Precision/Correctness
0.902
0.955
0.885
0.798
0.909
0.662
Recall/Completeness
0.944
0.981
0.795
0.883
0.891
0.228
F1
0.922
0.968
0.838
0.839
0.900
0.339

Overall accuracy 0.899

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
0.9315
0.0092
0.0180
0.0300
0.0061
0.0051
building
0.0214
0.9714
0.0038
0.0021
0.0001
0.0012
low_veg
0.0814
0.0126
0.8030
0.0970
0.0004
0.0056
tree
0.0438
0.0040
0.0706
0.8770
0.0022
0.0023
car
0.1829
0.0099
0.0156
0.0191
0.7583
0.0142
clutter
0.2432
0.2174
0.1488
0.0637
0.0042
0.3227
Precision/Correctness
0.829
0.917
0.849
0.820
0.859
0.872
Recall/Completeness
0.932
0.971
0.803
0.877
0.758
0.323
F1
0.877
0.943
0.826
0.847
0.806
0.471

Overall accuracy 0.856

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
0.9513
0.0060
0.0121
0.0231
0.0030
0.0045
building
0.0187
0.9766
0.0021
0.0014
0.0001
0.0011
low_veg
0.0651
0.0080
0.8315
0.0905
0.0001
0.0048
tree
0.0330
0.0030
0.0623
0.8982
0.0016
0.0020
car
0.0965
0.0105
0.0134
0.0127
0.8508
0.0161
clutter
0.2312
0.1936
0.1368
0.0630
0.0045
0.3709
Precision/Correctness
0.860
0.938
0.874
0.845
0.906
0.890
Recall/Completeness
0.951
0.977
0.831
0.898
0.851
0.371
F1
0.903
0.957
0.852
0.871
0.878
0.524

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
0.9189
0.0238
0.0318
0.0190
0.0041
0.0024
building
0.0210
0.9670
0.0048
0.0053
0.0000
0.0018
low_veg
0.0616
0.0089
0.8007
0.1186
0.0002
0.0100
tree
0.0430
0.0039
0.0630
0.8876
0.0009
0.0015
car
0.1989
0.0097
0.0057
0.0171
0.7670
0.0016
clutter
0.2451
0.1536
0.1727
0.0159
0.0049
0.4078
Precision/Correctness
0.899
0.915
0.774
0.790
0.905
0.903
Recall/Completeness
0.919
0.967
0.801
0.888
0.767
0.408
F1
0.909
0.940
0.787
0.836
0.830
0.562

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_15_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9344
0.0200
0.0278
0.0143
0.0017
0.0018
building
0.0163
0.9753
0.0030
0.0039
0.0000
0.0016
low_veg
0.0403
0.0040
0.8351
0.1111
0.0000
0.0095
tree
0.0318
0.0022
0.0541
0.9101
0.0006
0.0013
car
0.1043
0.0108
0.0021
0.0132
0.8680
0.0016
clutter
0.2335
0.1215
0.1678
0.0130
0.0048
0.4593
Precision/Correctness
0.924
0.936
0.797
0.820
0.945
0.920
Recall/Completeness
0.934
0.975
0.835
0.910
0.868
0.459
F1
0.929
0.955
0.816
0.863
0.905
0.613

Overall accuracy 0.896

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
0.9414
0.0156
0.0175
0.0179
0.0027
0.0049
building
0.0331
0.9446
0.0160
0.0016
0.0001
0.0046
low_veg
0.0270
0.0075
0.9254
0.0389
0.0000
0.0011
tree
0.1023
0.0030
0.0673
0.8260
0.0007
0.0007
car
0.2173
0.0069
0.0004
0.0127
0.7627
0.0000
clutter
0.3293
0.0623
0.3603
0.0299
0.0046
0.2137
Precision/Correctness
0.908
0.960
0.894
0.852
0.884
0.571
Recall/Completeness
0.941
0.945
0.925
0.826
0.763
0.214
F1
0.924
0.952
0.910
0.839
0.819
0.311

Overall accuracy 0.908

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
0.9567
0.0126
0.0132
0.0124
0.0012
0.0040
building
0.0287
0.9527
0.0132
0.0010
0.0001
0.0043
low_veg
0.0176
0.0041
0.9457
0.0319
0.0000
0.0008
tree
0.0833
0.0018
0.0569
0.8569
0.0005
0.0006
car
0.1284
0.0086
0.0002
0.0091
0.8537
0.0000
clutter
0.3093
0.0629
0.3722
0.0274
0.0040
0.2241
Precision/Correctness
0.928
0.971
0.912
0.884
0.931
0.592
Recall/Completeness
0.957
0.953
0.946
0.857
0.854
0.224
F1
0.942
0.962
0.928
0.870
0.891
0.325

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_14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9514
0.0147
0.0125
0.0118
0.0087
0.0008
building
0.0323
0.9608
0.0039
0.0017
0.0008
0.0004
low_veg
0.0790
0.0135
0.8229
0.0774
0.0008
0.0064
tree
0.0485
0.0074
0.0657
0.8750
0.0032
0.0002
car
0.1827
0.0018
0.0030
0.0164
0.7930
0.0031
clutter
0.3202
0.2995
0.1740
0.0164
0.0157
0.1742
Precision/Correctness
0.904
0.949
0.885
0.822
0.849
0.722
Recall/Completeness
0.951
0.961
0.823
0.875
0.793
0.174
F1
0.927
0.955
0.853
0.848
0.820
0.281

Overall accuracy 0.904

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
0.9671
0.0106
0.0090
0.0086
0.0041
0.0006
building
0.0279
0.9681
0.0021
0.0011
0.0005
0.0003
low_veg
0.0583
0.0063
0.8599
0.0699
0.0003
0.0053
tree
0.0364
0.0059
0.0552
0.9001
0.0024
0.0001
car
0.0964
0.0007
0.0011
0.0119
0.8863
0.0037
clutter
0.2959
0.3287
0.1584
0.0119
0.0157
0.1895
Precision/Correctness
0.928
0.960
0.910
0.853
0.903
0.763
Recall/Completeness
0.967
0.968
0.860
0.900
0.886
0.189
F1
0.947
0.964
0.884
0.876
0.895
0.304

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
0.9003
0.0349
0.0229
0.0099
0.0107
0.0213
building
0.0237
0.9079
0.0587
0.0039
0.0000
0.0058
low_veg
0.0898
0.0071
0.8217
0.0599
0.0013
0.0203
tree
0.0553
0.0051
0.1339
0.7994
0.0033
0.0029
car
0.1622
0.0052
0.0078
0.0179
0.8045
0.0025
clutter
0.1444
0.0780
0.2021
0.0063
0.0042
0.5649
Precision/Correctness
0.892
0.914
0.688
0.896
0.831
0.724
Recall/Completeness
0.900
0.908
0.822
0.799
0.805
0.565
F1
0.896
0.911
0.749
0.845
0.818
0.634

Overall accuracy 0.852

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
0.9174
0.0331
0.0169
0.0064
0.0055
0.0207
building
0.0189
0.9157
0.0576
0.0029
0.0000
0.0050
low_veg
0.0714
0.0040
0.8539
0.0505
0.0005
0.0198
tree
0.0441
0.0036
0.1179
0.8292
0.0026
0.0027
car
0.0884
0.0051
0.0023
0.0146
0.8874
0.0023
clutter
0.1327
0.0786
0.1901
0.0045
0.0045
0.5896
Precision/Correctness
0.917
0.923
0.705
0.923
0.887
0.740
Recall/Completeness
0.917
0.916
0.854
0.829
0.887
0.590
F1
0.917
0.919
0.772
0.873
0.887
0.656

Overall accuracy 0.874

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
0.9314
0.0235
0.0071
0.0295
0.0030
0.0055
building
0.0370
0.9473
0.0036
0.0028
0.0000
0.0091
low_veg
0.0684
0.0083
0.7354
0.1824
0.0002
0.0053
tree
0.0250
0.0044
0.0915
0.8732
0.0002
0.0058
car
0.1836
0.0169
0.0007
0.0382
0.7605
0.0001
clutter
0.1555
0.0169
0.1426
0.0031
0.0112
0.6706
Precision/Correctness
0.932
0.936
0.753
0.663
0.803
0.913
Recall/Completeness
0.931
0.947
0.735
0.873
0.760
0.671
F1
0.932
0.941
0.744
0.754
0.781
0.773

Overall accuracy 0.885

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
0.9460
0.0203
0.0049
0.0222
0.0015
0.0052
building
0.0336
0.9530
0.0026
0.0019
0.0000
0.0089
low_veg
0.0532
0.0050
0.7644
0.1725
0.0000
0.0049
tree
0.0152
0.0031
0.0852
0.8905
0.0001
0.0060
car
0.1035
0.0199
0.0000
0.0299
0.8467
0.0000
clutter
0.1433
0.0157
0.1431
0.0025
0.0110
0.6843
Precision/Correctness
0.944
0.946
0.766
0.695
0.838
0.919
Recall/Completeness
0.946
0.953
0.764
0.890
0.847
0.684
F1
0.945
0.949
0.765
0.780
0.842
0.784

Overall accuracy 0.901

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image