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 0.016 0.039 0.016 0.003 0.013 building 0.015 0.969 0.007 0.003 0.000 0.006 low_veg 0.038 0.009 0.883 0.059 0.000 0.010 tree 0.037 0.006 0.133 0.819 0.002 0.003 car 0.076 0.006 0.005 0.015 0.885 0.014 clutter 0.287 0.102 0.170 0.024 0.010 0.407 Precision/Correctness 0.898 0.948 0.807 0.876 0.907 0.688 Recall/Completeness 0.913 0.969 0.883 0.819 0.885 0.407 F1 0.905 0.958 0.843 0.847 0.896 0.512 Overall accuracy 0.880	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 0.933 0.012 0.031 0.011 0.000 0.012 building 0.012 0.977 0.004 0.002 0.000 0.005 low_veg 0.028 0.005 0.909 0.050 0.000 0.009 tree 0.028 0.004 0.118 0.846 0.002 0.003 car 0.013 0.006 0.002 0.009 0.955 0.015 clutter 0.287 0.087 0.156 0.019 0.010 0.440 Precision/Correctness 0.920 0.964 0.832 0.903 0.948 0.714 Recall/Completeness 0.933 0.977 0.909 0.846 0.955 0.440 F1 0.926 0.970 0.869 0.874 0.952 0.544 Overall accuracy 0.903
============================================================= Tile top_potsdam_2_13_class.tif, reference set: full_reference ↓ predicted || reference → imp_surf building low_veg tree car clutter imp_surf 0.8957 0.0190 0.0477 0.0124 0.0024 0.0229 building 0.0105 0.9660 0.0107 0.0081 0.0001 0.0046 low_veg 0.0265 0.0090 0.9033 0.0400 0.0002 0.0211 tree 0.0369 0.0097 0.1389 0.8082 0.0012 0.0051 car 0.0885 0.0067 0.0049 0.0080 0.8403 0.0517 clutter 0.1415 0.1785 0.1382 0.0282 0.0132 0.5004 Precision/Correctness 0.908 0.931 0.849 0.916 0.892 0.396 Recall/Completeness 0.896 0.966 0.903 0.808 0.840 0.500 F1 0.902 0.948 0.875 0.859 0.866 0.442 Overall accuracy 0.882 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 0.9200 0.0135 0.0375 0.0078 0.0003 0.0209 building 0.0067 0.9763 0.0064 0.0069 0.0001 0.0036 low_veg 0.0190 0.0049 0.9257 0.0302 0.0001 0.0202 tree 0.0282 0.0079 0.1170 0.8418 0.0008 0.0043 car 0.0278 0.0074 0.0011 0.0033 0.9000 0.0603 clutter 0.1043 0.1773 0.0925 0.0214 0.0153 0.5892 Precision/Correctness 0.936 0.953 0.878 0.939 0.943 0.384 Recall/Completeness 0.920 0.976 0.926 0.842 0.900 0.589 F1 0.928 0.965 0.901 0.888 0.921 0.465 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 0.8935 0.0070 0.0827 0.0112 0.0018 0.0039 building 0.0051 0.9770 0.0097 0.0051 0.0001 0.0030 low_veg 0.0117 0.0029 0.9519 0.0319 0.0000 0.0015 tree 0.0178 0.0034 0.2008 0.7762 0.0008 0.0010 car 0.0653 0.0096 0.0063 0.0093 0.9068 0.0026 clutter 0.1076 0.2427 0.2022 0.0749 0.0076 0.3650 Precision/Correctness 0.918 0.938 0.821 0.947 0.901 0.420 Recall/Completeness 0.893 0.977 0.952 0.776 0.907 0.365 F1 0.906 0.957 0.882 0.853 0.904 0.391 Overall accuracy 0.880 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 0.9234 0.0034 0.0630 0.0068 0.0001 0.0034 building 0.0020 0.9883 0.0044 0.0031 0.0000 0.0022 low_veg 0.0068 0.0011 0.9654 0.0254 0.0000 0.0012 tree 0.0137 0.0026 0.1836 0.7985 0.0006 0.0009 car 0.0071 0.0126 0.0018 0.0054 0.9706 0.0025 clutter 0.0721 0.2680 0.1405 0.0880 0.0075 0.4239 Precision/Correctness 0.944 0.963 0.840 0.960 0.951 0.421 Recall/Completeness 0.923 0.988 0.965 0.799 0.971 0.424 F1 0.934 0.975 0.898 0.872 0.961 0.422 Overall accuracy 0.899 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 0.8902 0.0188 0.0516 0.0179 0.0037 0.0177 building 0.0098 0.9757 0.0068 0.0036 0.0000 0.0041 low_veg 0.0460 0.0171 0.8726 0.0546 0.0004 0.0092 tree 0.0314 0.0070 0.1091 0.8479 0.0022 0.0024 car 0.0707 0.0056 0.0063 0.0138 0.8961 0.0075 clutter 0.2950 0.1336 0.2621 0.0270 0.0042 0.2781 Precision/Correctness 0.887 0.947 0.814 0.912 0.905 0.443 Recall/Completeness 0.890 0.976 0.873 0.848 0.896 0.278 F1 0.889 0.961 0.842 0.879 0.901 0.342 Overall accuracy 0.883 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 0.9189 0.0121 0.0405 0.0115 0.0003 0.0166 building 0.0069 0.9844 0.0031 0.0025 0.0000 0.0032 low_veg 0.0352 0.0084 0.9043 0.0445 0.0003 0.0073 tree 0.0232 0.0055 0.0911 0.8768 0.0016 0.0018 car 0.0092 0.0059 0.0032 0.0078 0.9661 0.0077 clutter 0.3157 0.1261 0.2241 0.0193 0.0046 0.3102 Precision/Correctness 0.916 0.967 0.848 0.937 0.956 0.444 Recall/Completeness 0.919 0.984 0.904 0.877 0.966 0.310 F1 0.917 0.976 0.875 0.906 0.961 0.365 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 0.8725 0.0176 0.0708 0.0198 0.0034 0.0159 building 0.0116 0.9765 0.0060 0.0028 0.0002 0.0029 low_veg 0.0276 0.0090 0.9100 0.0456 0.0001 0.0077 tree 0.0251 0.0048 0.1886 0.7755 0.0012 0.0047 car 0.0659 0.0022 0.0039 0.0301 0.8728 0.0251 clutter 0.1997 0.1492 0.1808 0.0318 0.0064 0.4320 Precision/Correctness 0.887 0.947 0.778 0.916 0.891 0.632 Recall/Completeness 0.873 0.976 0.910 0.776 0.873 0.432 F1 0.880 0.962 0.839 0.840 0.882 0.513 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 0.9027 0.0116 0.0584 0.0133 0.0003 0.0137 building 0.0093 0.9843 0.0025 0.0017 0.0001 0.0020 low_veg 0.0176 0.0037 0.9378 0.0350 0.0000 0.0059 tree 0.0197 0.0037 0.1727 0.7989 0.0010 0.0040 car 0.0102 0.0026 0.0017 0.0206 0.9379 0.0271 clutter 0.1906 0.1428 0.1495 0.0246 0.0070 0.4855 Precision/Correctness 0.917 0.966 0.802 0.941 0.943 0.667 Recall/Completeness 0.903 0.984 0.938 0.799 0.938 0.485 F1 0.910 0.975 0.864 0.864 0.941 0.562 Overall accuracy 0.893 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 0.9258 0.0096 0.0330 0.0256 0.0022 0.0039 building 0.0112 0.9732 0.0069 0.0047 0.0000 0.0040 low_veg 0.0466 0.0095 0.8316 0.1076 0.0002 0.0045 tree 0.0300 0.0025 0.0504 0.9153 0.0012 0.0006 car 0.0735 0.0030 0.0078 0.0268 0.8823 0.0066 clutter 0.4728 0.1132 0.1825 0.0669 0.0122 0.1524 Precision/Correctness 0.854 0.963 0.794 0.796 0.920 0.753 Recall/Completeness 0.926 0.973 0.832 0.915 0.882 0.152 F1 0.888 0.968 0.812 0.851 0.901 0.253 Overall accuracy 0.873 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 0.9466 0.0056 0.0260 0.0187 0.0003 0.0028 building 0.0062 0.9845 0.0025 0.0033 0.0000 0.0035 low_veg 0.0347 0.0041 0.8613 0.0965 0.0000 0.0034 tree 0.0187 0.0012 0.0373 0.9417 0.0007 0.0003 car 0.0083 0.0024 0.0020 0.0164 0.9645 0.0064 clutter 0.4937 0.0982 0.1732 0.0576 0.0130 0.1644 Precision/Correctness 0.879 0.976 0.832 0.833 0.946 0.788 Recall/Completeness 0.947 0.985 0.861 0.942 0.964 0.164 F1 0.912 0.980 0.847 0.884 0.955 0.272 Overall accuracy 0.900 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 0.8791 0.0186 0.0648 0.0142 0.0039 0.0194 building 0.0079 0.9797 0.0039 0.0046 0.0000 0.0040 low_veg 0.0776 0.0146 0.8233 0.0651 0.0003 0.0191 tree 0.0335 0.0062 0.1365 0.8134 0.0025 0.0079 car 0.0727 0.0067 0.0051 0.0209 0.8759 0.0187 clutter 0.4248 0.0418 0.1531 0.0112 0.0023 0.3667 Precision/Correctness 0.763 0.936 0.765 0.879 0.918 0.751 Recall/Completeness 0.879 0.980 0.823 0.813 0.876 0.367 F1 0.817 0.957 0.793 0.845 0.897 0.493 Overall accuracy 0.823 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 0.8997 0.0137 0.0583 0.0094 0.0003 0.0187 building 0.0050 0.9867 0.0019 0.0033 0.0000 0.0031 low_veg 0.0654 0.0089 0.8515 0.0572 0.0000 0.0170 tree 0.0246 0.0047 0.1182 0.8435 0.0020 0.0070 car 0.0227 0.0070 0.0015 0.0146 0.9341 0.0201 clutter 0.4374 0.0382 0.1398 0.0082 0.0018 0.3746 Precision/Correctness 0.778 0.953 0.787 0.904 0.965 0.776 Recall/Completeness 0.900 0.987 0.851 0.843 0.934 0.375 F1 0.834 0.969 0.818 0.873 0.949 0.505 Overall accuracy 0.844 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 0.9091 0.0198 0.0550 0.0050 0.0028 0.0083 building 0.0042 0.9877 0.0044 0.0021 0.0000 0.0016 low_veg 0.0563 0.0191 0.8883 0.0294 0.0005 0.0064 tree 0.0752 0.0102 0.1422 0.7625 0.0059 0.0040 car 0.0845 0.0058 0.0066 0.0076 0.8772 0.0183 clutter 0.2861 0.2129 0.1769 0.0135 0.0059 0.3046 Precision/Correctness 0.911 0.946 0.769 0.927 0.914 0.623 Recall/Completeness 0.909 0.988 0.888 0.763 0.877 0.305 F1 0.910 0.966 0.825 0.837 0.895 0.409 Overall accuracy 0.893 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 0.9266 0.0140 0.0497 0.0029 0.0004 0.0065 building 0.0017 0.9934 0.0023 0.0014 0.0000 0.0011 low_veg 0.0366 0.0098 0.9273 0.0225 0.0001 0.0038 tree 0.0585 0.0081 0.1205 0.8050 0.0045 0.0033 car 0.0123 0.0061 0.0012 0.0033 0.9547 0.0223 clutter 0.2883 0.2134 0.1658 0.0099 0.0051 0.3175 Precision/Correctness 0.938 0.963 0.796 0.951 0.956 0.652 Recall/Completeness 0.927 0.993 0.927 0.805 0.955 0.317 F1 0.932 0.978 0.857 0.872 0.955 0.427 Overall accuracy 0.919 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 0.9272 0.0131 0.0284 0.0213 0.0022 0.0079 building 0.0086 0.9794 0.0054 0.0035 0.0000 0.0031 low_veg 0.0499 0.0102 0.8594 0.0716 0.0001 0.0088 tree 0.0463 0.0045 0.1122 0.8328 0.0021 0.0021 car 0.0929 0.0022 0.0014 0.0122 0.8814 0.0099 clutter 0.2381 0.2307 0.1613 0.0163 0.0436 0.3100 Precision/Correctness 0.918 0.951 0.845 0.820 0.903 0.636 Recall/Completeness 0.927 0.979 0.859 0.833 0.881 0.310 F1 0.922 0.965 0.852 0.826 0.892 0.417 Overall accuracy 0.896 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 0.9471 0.0078 0.0233 0.0156 0.0002 0.0060 building 0.0047 0.9865 0.0038 0.0027 0.0000 0.0023 low_veg 0.0436 0.0054 0.8818 0.0623 0.0000 0.0069 tree 0.0312 0.0029 0.0980 0.8646 0.0018 0.0017 car 0.0158 0.0022 0.0002 0.0087 0.9620 0.0110 clutter 0.2061 0.2202 0.1566 0.0133 0.0545 0.3493 Precision/Correctness 0.943 0.968 0.872 0.851 0.926 0.663 Recall/Completeness 0.947 0.986 0.882 0.865 0.962 0.349 F1 0.945 0.977 0.877 0.858 0.944 0.458 Overall accuracy 0.921 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 0.9273 0.0074 0.0274 0.0254 0.0018 0.0107 building 0.0163 0.9736 0.0049 0.0018 0.0000 0.0033 low_veg 0.0473 0.0077 0.8427 0.0915 0.0003 0.0105 tree 0.0488 0.0059 0.0901 0.8473 0.0033 0.0046 car 0.0926 0.0111 0.0100 0.0164 0.8555 0.0145 clutter 0.1974 0.1775 0.1911 0.0435 0.0090 0.3815 Precision/Correctness 0.865 0.931 0.818 0.833 0.903 0.803 Recall/Completeness 0.927 0.974 0.843 0.847 0.855 0.382 F1 0.895 0.952 0.830 0.840 0.879 0.517 Overall accuracy 0.865 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 0.9480 0.0046 0.0189 0.0194 0.0001 0.0090 building 0.0141 0.9793 0.0026 0.0013 0.0000 0.0028 low_veg 0.0334 0.0047 0.8669 0.0862 0.0000 0.0087 tree 0.0382 0.0046 0.0808 0.8699 0.0025 0.0040 car 0.0150 0.0130 0.0059 0.0097 0.9412 0.0152 clutter 0.1854 0.1515 0.1831 0.0414 0.0091 0.4296 Precision/Correctness 0.893 0.950 0.844 0.856 0.932 0.832 Recall/Completeness 0.948 0.979 0.867 0.870 0.941 0.430 F1 0.920 0.965 0.856 0.863 0.937 0.567 Overall accuracy 0.890 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 0.9067 0.0097 0.0509 0.0180 0.0017 0.0129 building 0.0109 0.9694 0.0083 0.0051 0.0001 0.0063 low_veg 0.0368 0.0056 0.8263 0.1138 0.0001 0.0172 tree 0.0359 0.0034 0.0840 0.8729 0.0023 0.0015 car 0.1047 0.0054 0.0060 0.0135 0.8587 0.0117 clutter 0.2436 0.1169 0.1399 0.0149 0.0094 0.4754 Precision/Correctness 0.919 0.947 0.737 0.795 0.932 0.775 Recall/Completeness 0.907 0.969 0.826 0.873 0.859 0.475 F1 0.913 0.958 0.779 0.832 0.894 0.589 Overall accuracy 0.875 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 0.9235 0.0057 0.0452 0.0136 0.0004 0.0117 building 0.0068 0.9776 0.0063 0.0038 0.0000 0.0054 low_veg 0.0216 0.0018 0.8522 0.1079 0.0000 0.0166 tree 0.0253 0.0018 0.0731 0.8967 0.0017 0.0013 car 0.0213 0.0058 0.0014 0.0111 0.9476 0.0128 clutter 0.2379 0.0864 0.1286 0.0120 0.0097 0.5254 Precision/Correctness 0.940 0.967 0.761 0.824 0.953 0.796 Recall/Completeness 0.923 0.978 0.852 0.897 0.948 0.525 F1 0.932 0.972 0.804 0.859 0.950 0.633 Overall accuracy 0.899 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 0.9198 0.0250 0.0235 0.0150 0.0017 0.0149 building 0.0283 0.9489 0.0172 0.0016 0.0002 0.0038 low_veg 0.0179 0.0052 0.9446 0.0305 0.0000 0.0018 tree 0.0745 0.0031 0.0909 0.8285 0.0009 0.0021 car 0.0570 0.0151 0.0009 0.0025 0.9061 0.0184 clutter 0.3656 0.0582 0.3145 0.0162 0.0013 0.2443 Precision/Correctness 0.925 0.949 0.881 0.879 0.929 0.402 Recall/Completeness 0.920 0.949 0.945 0.828 0.906 0.244 F1 0.922 0.949 0.912 0.853 0.918 0.304 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 0.9375 0.0217 0.0175 0.0096 0.0003 0.0134 building 0.0246 0.9590 0.0122 0.0009 0.0002 0.0031 low_veg 0.0106 0.0027 0.9604 0.0253 0.0000 0.0010 tree 0.0560 0.0020 0.0788 0.8610 0.0005 0.0017 car 0.0056 0.0180 0.0003 0.0007 0.9587 0.0166 clutter 0.3742 0.0563 0.3187 0.0143 0.0006 0.2358 Precision/Correctness 0.942 0.959 0.902 0.908 0.978 0.401 Recall/Completeness 0.938 0.959 0.960 0.861 0.959 0.236 F1 0.940 0.959 0.930 0.884 0.968 0.297 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 0.9522 0.0054 0.0235 0.0103 0.0046 0.0040 building 0.0457 0.9349 0.0073 0.0015 0.0005 0.0101 low_veg 0.0473 0.0075 0.8749 0.0633 0.0004 0.0067 tree 0.0432 0.0078 0.0860 0.8576 0.0047 0.0007 car 0.0587 0.0025 0.0024 0.0109 0.9200 0.0054 clutter 0.2618 0.1331 0.2885 0.0292 0.0120 0.2753 Precision/Correctness 0.918 0.975 0.837 0.842 0.914 0.522 Recall/Completeness 0.952 0.935 0.875 0.858 0.920 0.275 F1 0.935 0.955 0.855 0.850 0.917 0.360 Overall accuracy 0.909 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 0.9692 0.0021 0.0178 0.0074 0.0004 0.0032 building 0.0410 0.9436 0.0043 0.0010 0.0003 0.0097 low_veg 0.0323 0.0031 0.9026 0.0568 0.0000 0.0053 tree 0.0321 0.0064 0.0731 0.8844 0.0034 0.0006 car 0.0067 0.0019 0.0004 0.0059 0.9793 0.0058 clutter 0.2436 0.1462 0.2858 0.0263 0.0114 0.2868 Precision/Correctness 0.936 0.984 0.864 0.871 0.968 0.530 Recall/Completeness 0.969 0.944 0.903 0.884 0.979 0.287 F1 0.952 0.963 0.883 0.877 0.973 0.372 Overall accuracy 0.929 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 0.9270 0.0056 0.0337 0.0097 0.0050 0.0190 building 0.0113 0.9765 0.0046 0.0021 0.0001 0.0054 low_veg 0.0413 0.0043 0.8822 0.0484 0.0010 0.0228 tree 0.0413 0.0056 0.1504 0.7957 0.0039 0.0032 car 0.0607 0.0017 0.0072 0.0104 0.9070 0.0129 clutter 0.1392 0.0125 0.1813 0.0138 0.0134 0.6398 Precision/Correctness 0.932 0.981 0.731 0.908 0.889 0.751 Recall/Completeness 0.927 0.976 0.882 0.796 0.907 0.640 F1 0.930 0.979 0.799 0.848 0.898 0.691 Overall accuracy 0.894 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 0.9473 0.0026 0.0257 0.0064 0.0004 0.0176 building 0.0067 0.9860 0.0014 0.0013 0.0001 0.0045 low_veg 0.0261 0.0015 0.9111 0.0393 0.0001 0.0217 tree 0.0312 0.0038 0.1342 0.8253 0.0030 0.0025 car 0.0084 0.0015 0.0017 0.0057 0.9685 0.0142 clutter 0.1316 0.0110 0.1706 0.0117 0.0134 0.6617 Precision/Correctness 0.952 0.990 0.756 0.934 0.943 0.773 Recall/Completeness 0.947 0.986 0.911 0.825 0.969 0.662 F1 0.950 0.988 0.826 0.876 0.955 0.713 Overall accuracy 0.916 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 0.9092 0.0355 0.0115 0.0230 0.0019 0.0189 building 0.0198 0.9461 0.0026 0.0022 0.0002 0.0290 low_veg 0.0574 0.0159 0.7821 0.1346 0.0002 0.0098 tree 0.0271 0.0076 0.1001 0.8637 0.0006 0.0010 car 0.0537 0.0119 0.0018 0.0269 0.9003 0.0054 clutter 0.2738 0.0213 0.0908 0.0018 0.0086 0.6036 Precision/Correctness 0.923 0.906 0.780 0.721 0.869 0.762 Recall/Completeness 0.909 0.946 0.782 0.864 0.900 0.604 F1 0.916 0.926 0.781 0.786 0.884 0.674 Overall accuracy 0.873 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 0.9241 0.0321 0.0089 0.0163 0.0004 0.0182 building 0.0173 0.9512 0.0019 0.0014 0.0002 0.0280 low_veg 0.0444 0.0111 0.8098 0.1259 0.0000 0.0088 tree 0.0170 0.0058 0.0921 0.8840 0.0004 0.0008 car 0.0054 0.0142 0.0006 0.0178 0.9575 0.0045 clutter 0.2683 0.0193 0.0896 0.0013 0.0080 0.6135 Precision/Correctness 0.931 0.918 0.797 0.756 0.909 0.774 Recall/Completeness 0.924 0.951 0.810 0.884 0.957 0.614 F1 0.928 0.934 0.803 0.815 0.932 0.685 Overall accuracy 0.888 Red/green image, indicating wrongly classified pixels Original, resized true ortho image for better comparison Classified image