Detailled semantic labeling (2D) result

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 0.913 0.021 0.029 0.021 0.008 0.008 building 0.016 0.967 0.010 0.004 0.000 0.003 low_veg 0.063 0.015 0.829 0.081 0.000 0.011 tree 0.034 0.006 0.091 0.864 0.002 0.003 car 0.150 0.008 0.005 0.017 0.812 0.008 clutter 0.293 0.131 0.144 0.027 0.005 0.401 Precision/Correctness 0.880 0.931 0.838 0.849 0.833 0.741 Recall/Completeness 0.913 0.967 0.829 0.864 0.812 0.401 F1 0.896 0.949 0.834 0.856 0.823 0.520 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 0.936 0.016 0.022 0.015 0.003 0.007 building 0.012 0.975 0.007 0.003 0.000 0.003 low_veg 0.049 0.009 0.861 0.071 0.000 0.010 tree 0.025 0.004 0.078 0.889 0.001 0.002 car 0.077 0.008 0.002 0.011 0.892 0.009 clutter 0.288 0.115 0.128 0.022 0.005 0.442 Precision/Correctness 0.905 0.950 0.866 0.878 0.899 0.771 Recall/Completeness 0.936 0.975 0.861 0.889 0.892 0.442 F1 0.920 0.962 0.864 0.883 0.895 0.562 Overall accuracy 0.901
============================================================= Tile top_potsdam_2_13_class.tif, reference set: full_reference ↓ predicted || reference → imp_surf building low_veg tree car clutter imp_surf 0.9011 0.0248 0.0305 0.0239 0.0056 0.0142 building 0.0377 0.9225 0.0252 0.0092 0.0001 0.0052 low_veg 0.0604 0.0243 0.8379 0.0699 0.0002 0.0073 tree 0.0279 0.0070 0.1028 0.8590 0.0010 0.0023 car 0.1293 0.0389 0.0018 0.0151 0.8047 0.0102 clutter 0.1762 0.2272 0.1529 0.0494 0.0020 0.3922 Precision/Correctness 0.858 0.895 0.867 0.870 0.844 0.508 Recall/Completeness 0.901 0.923 0.838 0.859 0.805 0.392 F1 0.879 0.909 0.852 0.864 0.824 0.443 Overall accuracy 0.865 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 0.9285 0.0191 0.0224 0.0162 0.0021 0.0118 building 0.0320 0.9352 0.0207 0.0073 0.0001 0.0047 low_veg 0.0498 0.0210 0.8690 0.0546 0.0001 0.0056 tree 0.0197 0.0054 0.0850 0.8876 0.0006 0.0017 car 0.0605 0.0447 0.0013 0.0087 0.8726 0.0121 clutter 0.1344 0.2308 0.1226 0.0405 0.0020 0.4697 Precision/Correctness 0.889 0.918 0.894 0.902 0.920 0.528 Recall/Completeness 0.929 0.935 0.869 0.888 0.873 0.470 F1 0.909 0.927 0.882 0.895 0.896 0.497 Overall accuracy 0.895 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 0.9085 0.0091 0.0543 0.0194 0.0055 0.0032 building 0.0052 0.9818 0.0056 0.0051 0.0002 0.0021 low_veg 0.0274 0.0056 0.8987 0.0643 0.0001 0.0039 tree 0.0146 0.0048 0.0990 0.8794 0.0006 0.0016 car 0.1074 0.0160 0.0022 0.0196 0.8535 0.0013 clutter 0.1042 0.2798 0.1986 0.0780 0.0016 0.3378 Precision/Correctness 0.887 0.912 0.893 0.912 0.821 0.300 Recall/Completeness 0.909 0.982 0.899 0.879 0.853 0.338 F1 0.897 0.946 0.896 0.895 0.837 0.318 Overall accuracy 0.897 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 0.9369 0.0050 0.0406 0.0126 0.0023 0.0027 building 0.0015 0.9922 0.0017 0.0028 0.0002 0.0016 low_veg 0.0189 0.0030 0.9216 0.0531 0.0000 0.0035 tree 0.0105 0.0038 0.0856 0.8982 0.0004 0.0014 car 0.0455 0.0153 0.0005 0.0131 0.9240 0.0016 clutter 0.0651 0.3125 0.1579 0.0783 0.0022 0.3840 Precision/Correctness 0.920 0.942 0.911 0.930 0.899 0.284 Recall/Completeness 0.937 0.992 0.922 0.898 0.924 0.384 F1 0.928 0.967 0.916 0.914 0.911 0.327 Overall accuracy 0.919 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 0.8954 0.0233 0.0364 0.0265 0.0094 0.0090 building 0.0130 0.9752 0.0047 0.0043 0.0000 0.0028 low_veg 0.0652 0.0268 0.8205 0.0781 0.0002 0.0092 tree 0.0238 0.0070 0.0747 0.8903 0.0017 0.0025 car 0.1341 0.0041 0.0029 0.0240 0.8286 0.0064 clutter 0.3023 0.1662 0.2523 0.0305 0.0034 0.2454 Precision/Correctness 0.871 0.932 0.849 0.884 0.830 0.503 Recall/Completeness 0.895 0.975 0.821 0.890 0.829 0.245 F1 0.883 0.953 0.835 0.887 0.829 0.330 Overall accuracy 0.880 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 0.9281 0.0159 0.0264 0.0178 0.0038 0.0080 building 0.0091 0.9835 0.0023 0.0029 0.0000 0.0022 low_veg 0.0514 0.0153 0.8625 0.0639 0.0001 0.0068 tree 0.0164 0.0053 0.0608 0.9145 0.0012 0.0019 car 0.0614 0.0041 0.0012 0.0154 0.9107 0.0072 clutter 0.3121 0.1657 0.2204 0.0192 0.0038 0.2787 Precision/Correctness 0.905 0.956 0.883 0.915 0.905 0.525 Recall/Completeness 0.928 0.983 0.862 0.914 0.911 0.279 F1 0.916 0.969 0.873 0.915 0.908 0.364 Overall accuracy 0.912 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 0.8839 0.0239 0.0487 0.0304 0.0076 0.0055 building 0.0072 0.9830 0.0047 0.0035 0.0000 0.0016 low_veg 0.0401 0.0151 0.8657 0.0720 0.0001 0.0070 tree 0.0195 0.0045 0.1223 0.8496 0.0010 0.0031 car 0.1262 0.0129 0.0030 0.0389 0.8118 0.0071 clutter 0.2386 0.1958 0.1917 0.0482 0.0022 0.3234 Precision/Correctness 0.876 0.930 0.828 0.885 0.831 0.695 Recall/Completeness 0.884 0.983 0.866 0.850 0.812 0.323 F1 0.880 0.956 0.847 0.867 0.821 0.441 Overall accuracy 0.874 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 0.9182 0.0164 0.0362 0.0217 0.0031 0.0045 building 0.0047 0.9900 0.0019 0.0022 0.0000 0.0012 low_veg 0.0275 0.0077 0.9007 0.0586 0.0000 0.0055 tree 0.0145 0.0032 0.1074 0.8716 0.0008 0.0024 car 0.0574 0.0137 0.0015 0.0274 0.8924 0.0075 clutter 0.2296 0.1874 0.1609 0.0404 0.0021 0.3795 Precision/Correctness 0.910 0.953 0.856 0.913 0.903 0.743 Recall/Completeness 0.918 0.990 0.901 0.872 0.892 0.380 F1 0.914 0.971 0.878 0.892 0.898 0.503 Overall accuracy 0.904 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 0.9193 0.0112 0.0276 0.0285 0.0093 0.0040 building 0.0136 0.9729 0.0063 0.0054 0.0001 0.0017 low_veg 0.0717 0.0152 0.7877 0.1166 0.0005 0.0083 tree 0.0298 0.0037 0.0473 0.9160 0.0016 0.0016 car 0.1382 0.0033 0.0044 0.0304 0.8224 0.0012 clutter 0.5265 0.1263 0.1500 0.0720 0.0028 0.1224 Precision/Correctness 0.829 0.957 0.810 0.781 0.842 0.721 Recall/Completeness 0.919 0.973 0.788 0.916 0.822 0.122 F1 0.872 0.965 0.799 0.843 0.832 0.209 Overall accuracy 0.861 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 0.9447 0.0062 0.0212 0.0207 0.0041 0.0031 building 0.0083 0.9833 0.0031 0.0039 0.0001 0.0013 low_veg 0.0548 0.0061 0.8281 0.1042 0.0001 0.0067 tree 0.0185 0.0021 0.0357 0.9417 0.0009 0.0011 car 0.0676 0.0026 0.0016 0.0209 0.9061 0.0011 clutter 0.5526 0.1094 0.1365 0.0629 0.0031 0.1356 Precision/Correctness 0.858 0.973 0.849 0.820 0.910 0.769 Recall/Completeness 0.945 0.983 0.828 0.942 0.906 0.136 F1 0.899 0.978 0.838 0.877 0.908 0.230 Overall accuracy 0.892 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 0.8714 0.0255 0.0648 0.0169 0.0103 0.0112 building 0.0118 0.9742 0.0033 0.0055 0.0001 0.0050 low_veg 0.0972 0.0216 0.7567 0.0887 0.0004 0.0354 tree 0.0306 0.0068 0.1113 0.8427 0.0019 0.0066 car 0.1499 0.0066 0.0041 0.0234 0.7917 0.0243 clutter 0.4665 0.0553 0.1516 0.0123 0.0008 0.3136 Precision/Correctness 0.735 0.915 0.766 0.852 0.850 0.692 Recall/Completeness 0.871 0.974 0.757 0.843 0.792 0.314 F1 0.797 0.943 0.761 0.847 0.820 0.432 Overall accuracy 0.803 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 0.8975 0.0198 0.0574 0.0108 0.0047 0.0097 building 0.0090 0.9810 0.0014 0.0043 0.0001 0.0042 low_veg 0.0830 0.0146 0.7863 0.0807 0.0001 0.0353 tree 0.0221 0.0050 0.0943 0.8716 0.0013 0.0057 car 0.0812 0.0067 0.0016 0.0163 0.8675 0.0266 clutter 0.4804 0.0507 0.1372 0.0090 0.0005 0.3221 Precision/Correctness 0.752 0.935 0.791 0.878 0.913 0.716 Recall/Completeness 0.898 0.981 0.786 0.872 0.868 0.322 F1 0.818 0.957 0.789 0.875 0.890 0.444 Overall accuracy 0.827 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 0.9264 0.0183 0.0350 0.0065 0.0076 0.0061 building 0.0066 0.9841 0.0044 0.0021 0.0001 0.0028 low_veg 0.0759 0.0260 0.8452 0.0432 0.0013 0.0084 tree 0.0685 0.0091 0.1196 0.7930 0.0056 0.0041 car 0.1537 0.0047 0.0071 0.0079 0.8112 0.0153 clutter 0.3689 0.2603 0.1437 0.0150 0.0010 0.2112 Precision/Correctness 0.894 0.941 0.808 0.906 0.844 0.538 Recall/Completeness 0.926 0.984 0.845 0.793 0.811 0.211 F1 0.910 0.962 0.826 0.846 0.827 0.303 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_15_class.tif, reference set: no_boundary ↓ predicted || reference → imp_surf building low_veg tree car clutter imp_surf 0.9492 0.0116 0.0279 0.0039 0.0029 0.0045 building 0.0034 0.9904 0.0024 0.0015 0.0000 0.0024 low_veg 0.0534 0.0147 0.8902 0.0347 0.0005 0.0065 tree 0.0502 0.0069 0.1002 0.8353 0.0040 0.0034 car 0.0767 0.0049 0.0023 0.0037 0.8948 0.0176 clutter 0.3801 0.2689 0.1151 0.0104 0.0007 0.2247 Precision/Correctness 0.925 0.959 0.843 0.933 0.915 0.567 Recall/Completeness 0.949 0.990 0.890 0.835 0.895 0.225 F1 0.937 0.975 0.866 0.882 0.905 0.322 Overall accuracy 0.920 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 0.9141 0.0187 0.0240 0.0278 0.0065 0.0089 building 0.0112 0.9781 0.0044 0.0039 0.0000 0.0025 low_veg 0.0959 0.0145 0.7784 0.1021 0.0002 0.0089 tree 0.0446 0.0054 0.0846 0.8615 0.0024 0.0015 car 0.1587 0.0005 0.0011 0.0066 0.8260 0.0071 clutter 0.2722 0.2772 0.1491 0.0222 0.0290 0.2503 Precision/Correctness 0.884 0.937 0.859 0.776 0.861 0.584 Recall/Completeness 0.914 0.978 0.778 0.862 0.826 0.250 F1 0.899 0.957 0.817 0.816 0.843 0.350 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_13_class.tif, reference set: no_boundary ↓ predicted || reference → imp_surf building low_veg tree car clutter imp_surf 0.9368 0.0120 0.0204 0.0214 0.0025 0.0069 building 0.0075 0.9843 0.0032 0.0029 0.0000 0.0021 low_veg 0.0894 0.0086 0.8036 0.0915 0.0001 0.0067 tree 0.0297 0.0035 0.0737 0.8902 0.0019 0.0011 car 0.0747 0.0006 0.0004 0.0036 0.9134 0.0072 clutter 0.2391 0.2721 0.1442 0.0201 0.0357 0.2887 Precision/Correctness 0.911 0.958 0.883 0.806 0.908 0.617 Recall/Completeness 0.937 0.984 0.804 0.890 0.913 0.289 F1 0.924 0.971 0.842 0.846 0.911 0.393 Overall accuracy 0.903 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 0.9231 0.0107 0.0188 0.0307 0.0079 0.0089 building 0.0156 0.9777 0.0034 0.0020 0.0001 0.0012 low_veg 0.0888 0.0115 0.7923 0.0964 0.0005 0.0105 tree 0.0448 0.0043 0.0633 0.8808 0.0034 0.0034 car 0.1699 0.0097 0.0124 0.0122 0.7798 0.0160 clutter 0.2332 0.2064 0.1048 0.0529 0.0028 0.4000 Precision/Correctness 0.830 0.919 0.868 0.825 0.834 0.836 Recall/Completeness 0.923 0.978 0.792 0.881 0.780 0.400 F1 0.874 0.948 0.828 0.852 0.806 0.541 Overall accuracy 0.860 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 0.9463 0.0069 0.0122 0.0234 0.0035 0.0077 building 0.0135 0.9824 0.0017 0.0013 0.0001 0.0011 low_veg 0.0709 0.0065 0.8242 0.0892 0.0001 0.0091 tree 0.0344 0.0033 0.0552 0.9017 0.0025 0.0028 car 0.0900 0.0099 0.0116 0.0073 0.8633 0.0181 clutter 0.2223 0.1787 0.0895 0.0504 0.0029 0.4562 Precision/Correctness 0.861 0.942 0.895 0.850 0.895 0.861 Recall/Completeness 0.946 0.982 0.824 0.902 0.863 0.456 F1 0.902 0.962 0.858 0.875 0.879 0.596 Overall accuracy 0.888 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 0.9134 0.0231 0.0329 0.0191 0.0067 0.0048 building 0.0116 0.9753 0.0048 0.0059 0.0000 0.0024 low_veg 0.0574 0.0079 0.7881 0.1218 0.0002 0.0246 tree 0.0402 0.0038 0.0626 0.8893 0.0018 0.0023 car 0.1652 0.0076 0.0058 0.0156 0.7989 0.0070 clutter 0.1595 0.1548 0.1213 0.0142 0.0025 0.5477 Precision/Correctness 0.921 0.917 0.788 0.787 0.872 0.856 Recall/Completeness 0.913 0.975 0.788 0.889 0.799 0.548 F1 0.917 0.945 0.788 0.835 0.834 0.668 Overall accuracy 0.879 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 0.9343 0.0181 0.0272 0.0142 0.0025 0.0037 building 0.0072 0.9836 0.0029 0.0043 0.0000 0.0019 low_veg 0.0377 0.0027 0.8218 0.1139 0.0001 0.0238 tree 0.0291 0.0021 0.0543 0.9113 0.0013 0.0020 car 0.0829 0.0086 0.0022 0.0123 0.8865 0.0075 clutter 0.1409 0.1222 0.1108 0.0110 0.0024 0.6127 Precision/Correctness 0.946 0.940 0.817 0.819 0.934 0.877 Recall/Completeness 0.934 0.984 0.822 0.911 0.887 0.613 F1 0.940 0.961 0.820 0.862 0.910 0.721 Overall accuracy 0.907 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 0.9295 0.0204 0.0209 0.0172 0.0040 0.0080 building 0.0302 0.9485 0.0157 0.0024 0.0000 0.0032 low_veg 0.0261 0.0087 0.9222 0.0411 0.0000 0.0018 tree 0.0856 0.0033 0.0670 0.8417 0.0014 0.0012 car 0.1762 0.0143 0.0007 0.0043 0.8024 0.0019 clutter 0.2961 0.0644 0.3354 0.0264 0.0032 0.2744 Precision/Correctness 0.917 0.952 0.892 0.852 0.850 0.563 Recall/Completeness 0.930 0.948 0.922 0.842 0.802 0.274 F1 0.923 0.950 0.907 0.847 0.826 0.369 Overall accuracy 0.907 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 0.9486 0.0163 0.0147 0.0116 0.0018 0.0070 building 0.0262 0.9570 0.0123 0.0018 0.0000 0.0027 low_veg 0.0159 0.0042 0.9446 0.0342 0.0000 0.0012 tree 0.0666 0.0019 0.0567 0.8729 0.0011 0.0009 car 0.1026 0.0183 0.0004 0.0026 0.8746 0.0016 clutter 0.2731 0.0641 0.3455 0.0236 0.0028 0.2909 Precision/Correctness 0.937 0.965 0.912 0.883 0.913 0.587 Recall/Completeness 0.949 0.957 0.945 0.873 0.875 0.291 F1 0.943 0.961 0.928 0.878 0.893 0.389 Overall accuracy 0.928 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 0.9466 0.0114 0.0143 0.0116 0.0130 0.0031 building 0.0315 0.9602 0.0040 0.0021 0.0009 0.0013 low_veg 0.0856 0.0155 0.8105 0.0779 0.0010 0.0095 tree 0.0442 0.0077 0.0652 0.8779 0.0049 0.0002 car 0.1465 0.0016 0.0021 0.0120 0.8325 0.0052 clutter 0.3556 0.2532 0.1279 0.0197 0.0121 0.2314 Precision/Correctness 0.903 0.955 0.886 0.822 0.807 0.623 Recall/Completeness 0.947 0.960 0.811 0.878 0.833 0.231 F1 0.924 0.958 0.847 0.849 0.820 0.337 Overall accuracy 0.903 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 0.9673 0.0062 0.0096 0.0082 0.0060 0.0026 building 0.0276 0.9677 0.0019 0.0014 0.0005 0.0009 low_veg 0.0634 0.0074 0.8512 0.0701 0.0003 0.0076 tree 0.0323 0.0065 0.0550 0.9025 0.0037 0.0001 car 0.0751 0.0005 0.0009 0.0077 0.9101 0.0057 clutter 0.3394 0.2759 0.1098 0.0159 0.0118 0.2472 Precision/Correctness 0.927 0.968 0.914 0.853 0.878 0.664 Recall/Completeness 0.967 0.968 0.851 0.902 0.910 0.247 F1 0.947 0.968 0.882 0.877 0.894 0.360 Overall accuracy 0.928 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 0.9160 0.0143 0.0255 0.0096 0.0134 0.0211 building 0.0131 0.9257 0.0521 0.0036 0.0000 0.0054 low_veg 0.0915 0.0097 0.8107 0.0610 0.0017 0.0254 tree 0.0503 0.0065 0.1237 0.8116 0.0046 0.0034 car 0.1494 0.0018 0.0099 0.0128 0.8202 0.0059 clutter 0.1559 0.0522 0.1581 0.0068 0.0017 0.6253 Precision/Correctness 0.900 0.951 0.704 0.898 0.803 0.732 Recall/Completeness 0.916 0.926 0.811 0.812 0.820 0.625 F1 0.908 0.938 0.753 0.853 0.811 0.674 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_15_class.tif, reference set: no_boundary ↓ predicted || reference → imp_surf building low_veg tree car clutter imp_surf 0.9392 0.0102 0.0179 0.0061 0.0069 0.0198 building 0.0083 0.9341 0.0504 0.0026 0.0000 0.0046 low_veg 0.0716 0.0049 0.8474 0.0511 0.0008 0.0243 tree 0.0386 0.0045 0.1082 0.8421 0.0036 0.0029 car 0.0851 0.0014 0.0032 0.0094 0.8951 0.0058 clutter 0.1470 0.0514 0.1451 0.0047 0.0015 0.6503 Precision/Correctness 0.925 0.964 0.727 0.925 0.869 0.754 Recall/Completeness 0.939 0.934 0.847 0.842 0.895 0.650 F1 0.932 0.949 0.783 0.882 0.882 0.698 Overall accuracy 0.892 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 0.8976 0.0483 0.0085 0.0323 0.0046 0.0086 building 0.0089 0.9704 0.0041 0.0036 0.0000 0.0130 low_veg 0.0625 0.0139 0.6976 0.2181 0.0005 0.0073 tree 0.0181 0.0049 0.0685 0.9028 0.0004 0.0052 car 0.1526 0.0213 0.0008 0.0336 0.7913 0.0004 clutter 0.1836 0.0220 0.0790 0.0027 0.0085 0.7041 Precision/Correctness 0.940 0.886 0.800 0.640 0.776 0.884 Recall/Completeness 0.898 0.970 0.698 0.903 0.791 0.704 F1 0.918 0.926 0.745 0.749 0.783 0.784 Overall accuracy 0.875 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 0.9145 0.0449 0.0060 0.0243 0.0022 0.0082 building 0.0061 0.9758 0.0029 0.0026 0.0000 0.0127 low_veg 0.0482 0.0097 0.7270 0.2089 0.0002 0.0061 tree 0.0091 0.0033 0.0621 0.9201 0.0002 0.0053 car 0.0837 0.0266 0.0002 0.0248 0.8644 0.0003 clutter 0.1732 0.0201 0.0781 0.0021 0.0081 0.7184 Precision/Correctness 0.951 0.897 0.820 0.670 0.832 0.892 Recall/Completeness 0.914 0.976 0.727 0.920 0.864 0.718 F1 0.932 0.935 0.771 0.775 0.848 0.796 Overall accuracy 0.892 Red/green image, indicating wrongly classified pixels Original, resized true ortho image for better comparison Classified image