Detailled semantic labeling (2D) result


Name B. Yu et al.
Affiliation Chinese Academy of Sciences, China
Abbreviation CAS_Y1
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
0.783
0.183
0.004
0.009
0.004
0.016
building
0.021
0.951
0.003
0.005
0.001
0.020
low_veg
0.276
0.118
0.503
0.060
0.001
0.043
tree
0.158
0.015
0.145
0.652
0.004
0.026
car
0.076
0.029
0.000
0.006
0.875
0.013
clutter
0.470
0.233
0.024
0.015
0.020
0.237
Precision/Correctness
0.686
0.703
0.792
0.865
0.844
0.318
Recall/Completeness
0.783
0.951
0.503
0.652
0.875
0.237
F1
0.731
0.808
0.615
0.743
0.859
0.271

Overall accuracy 0.719

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
0.790
0.185
0.003
0.007
0.001
0.014
building
0.017
0.959
0.002
0.004
0.001
0.017
low_veg
0.263
0.114
0.527
0.055
0.000
0.040
tree
0.142
0.013
0.137
0.679
0.004
0.024
car
0.016
0.029
0.000
0.004
0.938
0.013
clutter
0.486
0.220
0.016
0.012
0.020
0.245
Precision/Correctness
0.706
0.718
0.809
0.885
0.888
0.327
Recall/Completeness
0.790
0.959
0.527
0.679
0.938
0.245
F1
0.746
0.821
0.638
0.769
0.912
0.280

Overall accuracy 0.738
=============================================================

Tile top_potsdam_2_13_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.5974
0.3365
0.0053
0.0049
0.0030
0.0529
building
0.0193
0.9089
0.0119
0.0089
0.0016
0.0495
low_veg
0.2306
0.0944
0.5855
0.0321
0.0008
0.0566
tree
0.1481
0.0243
0.1943
0.5878
0.0024
0.0432
car
0.0869
0.0860
0.0003
0.0018
0.7745
0.0505
clutter
0.1518
0.4198
0.0397
0.0153
0.0374
0.3360
Precision/Correctness
0.568
0.571
0.782
0.914
0.781
0.115
Recall/Completeness
0.597
0.909
0.585
0.588
0.775
0.336
F1
0.582
0.701
0.670
0.715
0.778
0.171

Overall accuracy 0.646

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
0.5988
0.3460
0.0030
0.0030
0.0007
0.0486
building
0.0158
0.9187
0.0109
0.0082
0.0013
0.0452
low_veg
0.2234
0.0882
0.6091
0.0266
0.0004
0.0523
tree
0.1364
0.0208
0.1824
0.6189
0.0019
0.0397
car
0.0274
0.0859
0.0001
0.0008
0.8295
0.0563
clutter
0.1264
0.4314
0.0285
0.0108
0.0401
0.3628
Precision/Correctness
0.584
0.588
0.801
0.932
0.835
0.101
Recall/Completeness
0.599
0.919
0.609
0.619
0.829
0.363
F1
0.591
0.717
0.692
0.744
0.832
0.158

Overall accuracy 0.667

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
0.4892
0.4467
0.0069
0.0057
0.0027
0.0488
building
0.0149
0.9442
0.0039
0.0083
0.0005
0.0283
low_veg
0.3040
0.1058
0.5144
0.0264
0.0003
0.0491
tree
0.1110
0.0133
0.2476
0.5898
0.0011
0.0372
car
0.0368
0.1757
0.0000
0.0043
0.7502
0.0330
clutter
0.1656
0.3455
0.0499
0.0320
0.0660
0.3409
Precision/Correctness
0.311
0.353
0.704
0.944
0.791
0.028
Recall/Completeness
0.489
0.944
0.514
0.590
0.750
0.341
F1
0.380
0.513
0.594
0.726
0.770
0.051

Overall accuracy 0.568

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
0.4850
0.4621
0.0032
0.0036
0.0004
0.0457
building
0.0109
0.9565
0.0022
0.0070
0.0005
0.0229
low_veg
0.3040
0.1007
0.5248
0.0219
0.0002
0.0484
tree
0.1028
0.0113
0.2409
0.6087
0.0009
0.0355
car
0.0022
0.1780
0.0000
0.0031
0.7846
0.0321
clutter
0.1315
0.3718
0.0240
0.0308
0.0637
0.3782
Precision/Correctness
0.305
0.364
0.713
0.956
0.853
0.023
Recall/Completeness
0.485
0.957
0.525
0.609
0.785
0.378
F1
0.375
0.528
0.605
0.744
0.817
0.043

Overall accuracy 0.580

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
0.8230
0.1337
0.0043
0.0093
0.0041
0.0256
building
0.0156
0.9547
0.0017
0.0035
0.0002
0.0242
low_veg
0.3328
0.0927
0.4845
0.0527
0.0004
0.0369
tree
0.1406
0.0130
0.1331
0.6951
0.0035
0.0148
car
0.0848
0.0284
0.0004
0.0100
0.8681
0.0084
clutter
0.4242
0.2631
0.0697
0.0263
0.0219
0.1948
Precision/Correctness
0.624
0.784
0.767
0.909
0.857
0.159
Recall/Completeness
0.823
0.955
0.485
0.695
0.868
0.195
F1
0.710
0.861
0.594
0.788
0.863
0.175

Overall accuracy 0.734

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
0.8359
0.1322
0.0022
0.0060
0.0005
0.0232
building
0.0119
0.9643
0.0008
0.0024
0.0001
0.0206
low_veg
0.3252
0.0845
0.5110
0.0465
0.0002
0.0326
tree
0.1261
0.0114
0.1235
0.7239
0.0028
0.0123
car
0.0192
0.0298
0.0000
0.0069
0.9349
0.0092
clutter
0.4434
0.2582
0.0563
0.0240
0.0250
0.1931
Precision/Correctness
0.646
0.807
0.788
0.929
0.908
0.142
Recall/Completeness
0.836
0.964
0.511
0.724
0.935
0.193
F1
0.729
0.878
0.620
0.814
0.921
0.164

Overall accuracy 0.759

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
0.7149
0.2208
0.0085
0.0090
0.0040
0.0429
building
0.0139
0.9673
0.0017
0.0023
0.0002
0.0145
low_veg
0.3235
0.1372
0.4630
0.0522
0.0002
0.0238
tree
0.1755
0.0087
0.1745
0.5977
0.0021
0.0415
car
0.0561
0.1057
0.0000
0.0147
0.7630
0.0606
clutter
0.2321
0.3332
0.0549
0.0167
0.0123
0.3508
Precision/Correctness
0.511
0.686
0.720
0.899
0.831
0.261
Recall/Completeness
0.715
0.967
0.463
0.598
0.763
0.351
F1
0.596
0.803
0.564
0.718
0.795
0.299

Overall accuracy 0.662

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
0.7248
0.2229
0.0058
0.0060
0.0005
0.0399
building
0.0113
0.9744
0.0009
0.0016
0.0001
0.0117
low_veg
0.3162
0.1333
0.4833
0.0473
0.0001
0.0198
tree
0.1634
0.0070
0.1674
0.6204
0.0018
0.0400
car
0.0086
0.1034
0.0000
0.0086
0.8176
0.0618
clutter
0.2155
0.3269
0.0419
0.0127
0.0123
0.3906
Precision/Correctness
0.525
0.707
0.735
0.918
0.889
0.267
Recall/Completeness
0.725
0.974
0.483
0.620
0.818
0.391
F1
0.609
0.819
0.583
0.740
0.852
0.317

Overall accuracy 0.683

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
0.7600
0.2091
0.0028
0.0116
0.0028
0.0137
building
0.0323
0.9435
0.0014
0.0053
0.0003
0.0171
low_veg
0.3473
0.1667
0.3788
0.0603
0.0003
0.0467
tree
0.1636
0.0212
0.0695
0.7290
0.0037
0.0130
car
0.0839
0.0221
0.0006
0.0142
0.8604
0.0189
clutter
0.6248
0.2365
0.0234
0.0262
0.0084
0.0807
Precision/Correctness
0.646
0.746
0.823
0.853
0.902
0.226
Recall/Completeness
0.760
0.943
0.379
0.729
0.860
0.081
F1
0.698
0.833
0.519
0.786
0.881
0.119

Overall accuracy 0.718

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
0.7650
0.2126
0.0018
0.0086
0.0005
0.0116
building
0.0269
0.9530
0.0007
0.0043
0.0003
0.0149
low_veg
0.3354
0.1627
0.4013
0.0545
0.0002
0.0459
tree
0.1402
0.0186
0.0617
0.7654
0.0028
0.0112
car
0.0141
0.0205
0.0001
0.0095
0.9381
0.0177
clutter
0.6558
0.2203
0.0168
0.0219
0.0087
0.0765
Precision/Correctness
0.671
0.763
0.852
0.881
0.936
0.216
Recall/Completeness
0.765
0.953
0.401
0.765
0.938
0.076
F1
0.715
0.848
0.546
0.819
0.937
0.113

Overall accuracy 0.742

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
0.8566
0.0921
0.0122
0.0061
0.0047
0.0283
building
0.0156
0.9600
0.0015
0.0046
0.0009
0.0174
low_veg
0.3599
0.0795
0.4228
0.0623
0.0010
0.0746
tree
0.2165
0.0111
0.1406
0.5985
0.0052
0.0282
car
0.0914
0.0111
0.0006
0.0081
0.8706
0.0183
clutter
0.4709
0.2876
0.0118
0.0096
0.0051
0.2150
Precision/Correctness
0.554
0.728
0.758
0.863
0.866
0.388
Recall/Completeness
0.857
0.960
0.423
0.598
0.871
0.215
F1
0.673
0.828
0.543
0.707
0.868
0.277

Overall accuracy 0.665

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
0.8651
0.0908
0.0114
0.0043
0.0008
0.0276
building
0.0119
0.9683
0.0006
0.0036
0.0008
0.0147
low_veg
0.3434
0.0762
0.4462
0.0604
0.0005
0.0733
tree
0.1999
0.0090
0.1330
0.6279
0.0045
0.0257
car
0.0317
0.0111
0.0005
0.0053
0.9326
0.0188
clutter
0.4699
0.2959
0.0070
0.0078
0.0038
0.2158
Precision/Correctness
0.569
0.741
0.776
0.880
0.911
0.406
Recall/Completeness
0.865
0.968
0.446
0.628
0.933
0.216
F1
0.686
0.839
0.567
0.733
0.922
0.282

Overall accuracy 0.683

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
0.8409
0.1437
0.0014
0.0026
0.0037
0.0076
building
0.0224
0.9536
0.0007
0.0039
0.0013
0.0181
low_veg
0.4007
0.2218
0.3016
0.0337
0.0017
0.0404
tree
0.2347
0.0357
0.0990
0.5874
0.0113
0.0319
car
0.0894
0.0173
0.0001
0.0013
0.8878
0.0041
clutter
0.2486
0.4441
0.0286
0.0119
0.0193
0.2475
Precision/Correctness
0.719
0.741
0.777
0.904
0.843
0.273
Recall/Completeness
0.841
0.954
0.302
0.587
0.888
0.247
F1
0.775
0.834
0.434
0.712
0.865
0.260

Overall accuracy 0.739

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
0.8496
0.1415
0.0007
0.0019
0.0008
0.0055
building
0.0185
0.9607
0.0003
0.0033
0.0011
0.0161
low_veg
0.3870
0.2249
0.3187
0.0307
0.0012
0.0375
tree
0.2129
0.0324
0.0883
0.6268
0.0095
0.0301
car
0.0162
0.0173
0.0000
0.0005
0.9619
0.0041
clutter
0.2128
0.4763
0.0238
0.0099
0.0171
0.2600
Precision/Correctness
0.746
0.759
0.804
0.922
0.883
0.269
Recall/Completeness
0.850
0.961
0.319
0.627
0.962
0.260
F1
0.795
0.848
0.457
0.746
0.921
0.264

Overall accuracy 0.762

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
0.7941
0.1801
0.0019
0.0105
0.0034
0.0099
building
0.0136
0.9690
0.0004
0.0034
0.0000
0.0136
low_veg
0.2365
0.1269
0.5035
0.0869
0.0005
0.0457
tree
0.1508
0.0163
0.0926
0.7189
0.0039
0.0175
car
0.0941
0.0017
0.0000
0.0047
0.8924
0.0071
clutter
0.2594
0.3231
0.0348
0.0133
0.1127
0.2568
Precision/Correctness
0.771
0.741
0.878
0.804
0.810
0.311
Recall/Completeness
0.794
0.969
0.504
0.719
0.892
0.257
F1
0.782
0.840
0.640
0.759
0.849
0.281

Overall accuracy 0.764

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
0.8020
0.1806
0.0011
0.0076
0.0007
0.0080
building
0.0103
0.9748
0.0002
0.0028
0.0000
0.0119
low_veg
0.2284
0.1240
0.5241
0.0805
0.0004
0.0425
tree
0.1275
0.0138
0.0864
0.7530
0.0034
0.0158
car
0.0159
0.0012
0.0000
0.0032
0.9714
0.0082
clutter
0.2321
0.3118
0.0303
0.0109
0.1378
0.2771
Precision/Correctness
0.796
0.756
0.896
0.827
0.832
0.297
Recall/Completeness
0.802
0.975
0.524
0.753
0.971
0.277
F1
0.799
0.851
0.661
0.788
0.897
0.287

Overall accuracy 0.784

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
0.8858
0.0814
0.0048
0.0155
0.0029
0.0096
building
0.0062
0.9766
0.0014
0.0054
0.0004
0.0100
low_veg
0.2025
0.2148
0.4530
0.0902
0.0008
0.0386
tree
0.1630
0.0182
0.0898
0.6988
0.0070
0.0232
car
0.0875
0.0124
0.0007
0.0033
0.8886
0.0075
clutter
0.4463
0.2537
0.0266
0.0300
0.0207
0.2227
Precision/Correctness
0.690
0.734
0.829
0.822
0.821
0.488
Recall/Completeness
0.886
0.977
0.453
0.699
0.889
0.223
F1
0.776
0.838
0.586
0.755
0.853
0.306

Overall accuracy 0.736

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
0.8980
0.0778
0.0030
0.0128
0.0005
0.0079
building
0.0041
0.9815
0.0008
0.0049
0.0003
0.0084
low_veg
0.1838
0.2159
0.4738
0.0885
0.0002
0.0378
tree
0.1466
0.0166
0.0846
0.7243
0.0059
0.0221
car
0.0171
0.0133
0.0002
0.0018
0.9595
0.0081
clutter
0.4684
0.2218
0.0184
0.0295
0.0212
0.2406
Precision/Correctness
0.712
0.757
0.848
0.839
0.848
0.503
Recall/Completeness
0.898
0.981
0.474
0.724
0.959
0.241
F1
0.794
0.855
0.608
0.777
0.901
0.325

Overall accuracy 0.758

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
0.8666
0.1036
0.0041
0.0105
0.0027
0.0125
building
0.0318
0.9378
0.0006
0.0121
0.0013
0.0164
low_veg
0.2848
0.1172
0.4663
0.0909
0.0005
0.0402
tree
0.1241
0.0108
0.0785
0.7545
0.0048
0.0272
car
0.1036
0.0018
0.0003
0.0060
0.8804
0.0079
clutter
0.6095
0.2305
0.0109
0.0097
0.0148
0.1247
Precision/Correctness
0.761
0.754
0.851
0.806
0.883
0.325
Recall/Completeness
0.867
0.938
0.466
0.755
0.880
0.125
F1
0.811
0.836
0.602
0.780
0.882
0.180

Overall accuracy 0.763

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
0.8752
0.1019
0.0034
0.0078
0.0008
0.0108
building
0.0268
0.9463
0.0004
0.0107
0.0011
0.0147
low_veg
0.2593
0.1183
0.4963
0.0863
0.0002
0.0396
tree
0.1077
0.0088
0.0725
0.7810
0.0041
0.0260
car
0.0299
0.0010
0.0000
0.0048
0.9566
0.0076
clutter
0.6450
0.2082
0.0051
0.0078
0.0143
0.1195
Precision/Correctness
0.783
0.772
0.870
0.832
0.908
0.318
Recall/Completeness
0.875
0.946
0.496
0.781
0.957
0.119
F1
0.826
0.850
0.632
0.805
0.932
0.174

Overall accuracy 0.783

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
0.7392
0.2369
0.0056
0.0093
0.0026
0.0065
building
0.0164
0.9600
0.0097
0.0012
0.0004
0.0122
low_veg
0.0983
0.0684
0.7510
0.0462
0.0006
0.0355
tree
0.1997
0.0141
0.0963
0.6833
0.0028
0.0038
car
0.0476
0.1120
0.0000
0.0028
0.8298
0.0078
clutter
0.6232
0.1656
0.0353
0.0217
0.0092
0.1450
Precision/Correctness
0.812
0.674
0.914
0.843
0.854
0.172
Recall/Completeness
0.739
0.960
0.751
0.683
0.830
0.145
F1
0.774
0.792
0.825
0.755
0.842
0.157

Overall accuracy 0.777

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
0.7451
0.2397
0.0031
0.0063
0.0009
0.0049
building
0.0128
0.9679
0.0077
0.0007
0.0004
0.0104
low_veg
0.0853
0.0637
0.7729
0.0423
0.0005
0.0353
tree
0.1765
0.0120
0.0894
0.7165
0.0025
0.0031
car
0.0040
0.1060
0.0000
0.0013
0.8817
0.0070
clutter
0.6767
0.1499
0.0169
0.0201
0.0074
0.1291
Precision/Correctness
0.834
0.685
0.930
0.865
0.899
0.152
Recall/Completeness
0.745
0.968
0.773
0.717
0.882
0.129
F1
0.787
0.802
0.844
0.784
0.890
0.139

Overall accuracy 0.794

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
0.9415
0.0402
0.0019
0.0066
0.0062
0.0037
building
0.0457
0.9332
0.0018
0.0020
0.0011
0.0161
low_veg
0.2698
0.1211
0.4901
0.0949
0.0013
0.0227
tree
0.1761
0.0044
0.0490
0.7541
0.0103
0.0062
car
0.0531
0.0167
0.0000
0.0035
0.9213
0.0053
clutter
0.3588
0.3613
0.0294
0.0221
0.0483
0.1802
Precision/Correctness
0.818
0.865
0.919
0.795
0.852
0.284
Recall/Completeness
0.941
0.933
0.490
0.754
0.921
0.180
F1
0.875
0.898
0.639
0.774
0.885
0.220

Overall accuracy 0.833

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
0.9545
0.0366
0.0010
0.0047
0.0007
0.0025
building
0.0398
0.9427
0.0009
0.0015
0.0007
0.0143
low_veg
0.2485
0.1206
0.5199
0.0907
0.0006
0.0197
tree
0.1600
0.0030
0.0431
0.7804
0.0089
0.0046
car
0.0054
0.0170
0.0000
0.0019
0.9707
0.0050
clutter
0.3450
0.3973
0.0168
0.0183
0.0470
0.1756
Precision/Correctness
0.839
0.878
0.938
0.819
0.909
0.289
Recall/Completeness
0.955
0.943
0.520
0.780
0.971
0.176
F1
0.893
0.909
0.669
0.799
0.939
0.219

Overall accuracy 0.854

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
0.8709
0.1036
0.0076
0.0046
0.0072
0.0061
building
0.0142
0.9599
0.0033
0.0048
0.0007
0.0171
low_veg
0.2560
0.0628
0.6101
0.0399
0.0020
0.0292
tree
0.1430
0.0169
0.1982
0.6199
0.0089
0.0130
car
0.0562
0.0197
0.0011
0.0053
0.9164
0.0013
clutter
0.3504
0.0995
0.0330
0.0075
0.0208
0.4889
Precision/Correctness
0.792
0.794
0.713
0.911
0.828
0.698
Recall/Completeness
0.871
0.960
0.610
0.620
0.916
0.489
F1
0.830
0.869
0.658
0.738
0.870
0.575

Overall accuracy 0.793

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
0.8816
0.1041
0.0051
0.0031
0.0012
0.0049
building
0.0086
0.9686
0.0023
0.0039
0.0007
0.0159
low_veg
0.2325
0.0599
0.6430
0.0358
0.0005
0.0283
tree
0.1287
0.0145
0.1861
0.6513
0.0078
0.0116
car
0.0095
0.0204
0.0001
0.0036
0.9656
0.0009
clutter
0.3447
0.0983
0.0247
0.0067
0.0174
0.5082
Precision/Correctness
0.816
0.806
0.732
0.930
0.890
0.724
Recall/Completeness
0.882
0.969
0.643
0.651
0.966
0.508
F1
0.847
0.880
0.685
0.766
0.926
0.597

Overall accuracy 0.813

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
0.6206
0.3475
0.0016
0.0193
0.0025
0.0085
building
0.0149
0.9426
0.0018
0.0037
0.0003
0.0367
low_veg
0.2141
0.0603
0.4997
0.1941
0.0007
0.0311
tree
0.1075
0.0042
0.0647
0.8050
0.0016
0.0169
car
0.0436
0.0209
0.0000
0.0156
0.9099
0.0099
clutter
0.6570
0.0418
0.0058
0.0015
0.0095
0.2844
Precision/Correctness
0.768
0.536
0.882
0.674
0.840
0.601
Recall/Completeness
0.621
0.943
0.500
0.805
0.910
0.284
F1
0.686
0.683
0.638
0.733
0.873
0.386

Overall accuracy 0.668

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
0.6225
0.3542
0.0010
0.0141
0.0006
0.0076
building
0.0121
0.9483
0.0013
0.0030
0.0003
0.0351
low_veg
0.1911
0.0603
0.5282
0.1909
0.0002
0.0293
tree
0.0893
0.0030
0.0608
0.8312
0.0010
0.0146
car
0.0051
0.0202
0.0000
0.0107
0.9551
0.0090
clutter
0.6609
0.0382
0.0048
0.0011
0.0086
0.2864
Precision/Correctness
0.779
0.538
0.899
0.698
0.892
0.624
Recall/Completeness
0.622
0.948
0.528
0.831
0.955
0.286
F1
0.692
0.686
0.666
0.759
0.923
0.393

Overall accuracy 0.676

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image