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 0.895 0.030 0.040 0.020 0.004 0.010 building 0.049 0.915 0.014 0.005 0.001 0.017 low_veg 0.050 0.016 0.847 0.078 0.000 0.009 tree 0.044 0.007 0.155 0.788 0.002 0.004 car 0.103 0.012 0.005 0.019 0.844 0.017 clutter 0.340 0.143 0.164 0.029 0.015 0.310 Precision/Correctness 0.856 0.913 0.781 0.840 0.870 0.584 Recall/Completeness 0.895 0.915 0.847 0.788 0.844 0.310 F1 0.876 0.914 0.812 0.813 0.857 0.405 Overall accuracy 0.844	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 0.916 0.026 0.033 0.015 0.001 0.009 building 0.045 0.924 0.011 0.004 0.001 0.016 low_veg 0.038 0.012 0.872 0.070 0.000 0.008 tree 0.034 0.006 0.141 0.814 0.002 0.003 car 0.031 0.013 0.002 0.013 0.922 0.019 clutter 0.345 0.129 0.150 0.024 0.014 0.337 Precision/Correctness 0.878 0.930 0.805 0.867 0.919 0.602 Recall/Completeness 0.916 0.924 0.872 0.814 0.922 0.337 F1 0.897 0.927 0.837 0.840 0.921 0.432 Overall accuracy 0.867
============================================================= Tile top_potsdam_2_13_class.tif, reference set: full_reference ↓ predicted || reference → imp_surf building low_veg tree car clutter imp_surf 0.8635 0.0230 0.0602 0.0103 0.0024 0.0405 building 0.0561 0.8698 0.0438 0.0049 0.0003 0.0251 low_veg 0.0376 0.0125 0.8760 0.0443 0.0002 0.0294 tree 0.0463 0.0107 0.1917 0.7421 0.0016 0.0076 car 0.1368 0.0131 0.0055 0.0109 0.7812 0.0525 clutter 0.1499 0.2074 0.1659 0.0279 0.0391 0.4098 Precision/Correctness 0.852 0.909 0.790 0.908 0.832 0.234 Recall/Completeness 0.864 0.870 0.876 0.742 0.781 0.410 F1 0.858 0.889 0.831 0.817 0.806 0.298 Overall accuracy 0.830 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 0.8884 0.0179 0.0485 0.0062 0.0003 0.0386 building 0.0520 0.8807 0.0401 0.0039 0.0002 0.0230 low_veg 0.0303 0.0090 0.8969 0.0349 0.0001 0.0287 tree 0.0375 0.0093 0.1697 0.7754 0.0012 0.0068 car 0.0610 0.0142 0.0020 0.0056 0.8573 0.0599 clutter 0.1158 0.2087 0.1291 0.0211 0.0482 0.4771 Precision/Correctness 0.878 0.932 0.816 0.932 0.877 0.216 Recall/Completeness 0.888 0.881 0.897 0.775 0.857 0.477 F1 0.883 0.905 0.855 0.847 0.867 0.297 Overall accuracy 0.857 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 0.8224 0.0204 0.1260 0.0085 0.0019 0.0209 building 0.0348 0.9032 0.0234 0.0065 0.0001 0.0321 low_veg 0.0236 0.0035 0.9407 0.0255 0.0001 0.0067 tree 0.0262 0.0047 0.2570 0.7064 0.0009 0.0048 car 0.1103 0.0170 0.0054 0.0099 0.8342 0.0232 clutter 0.1271 0.2794 0.2079 0.0474 0.0290 0.3092 Precision/Correctness 0.849 0.894 0.775 0.953 0.871 0.100 Recall/Completeness 0.822 0.903 0.941 0.706 0.834 0.309 F1 0.835 0.899 0.850 0.811 0.852 0.151 Overall accuracy 0.834 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 0.8574 0.0189 0.0978 0.0051 0.0003 0.0205 building 0.0308 0.9176 0.0161 0.0047 0.0000 0.0307 low_veg 0.0201 0.0019 0.9516 0.0201 0.0000 0.0063 tree 0.0223 0.0042 0.2403 0.7278 0.0006 0.0047 car 0.0362 0.0189 0.0011 0.0056 0.9147 0.0235 clutter 0.0998 0.3245 0.1684 0.0535 0.0301 0.3238 Precision/Correctness 0.870 0.914 0.793 0.965 0.923 0.079 Recall/Completeness 0.857 0.918 0.952 0.728 0.915 0.324 F1 0.864 0.916 0.865 0.830 0.919 0.127 Overall accuracy 0.853 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 0.8855 0.0276 0.0506 0.0226 0.0040 0.0097 building 0.0320 0.9466 0.0082 0.0040 0.0002 0.0090 low_veg 0.0563 0.0229 0.8475 0.0651 0.0002 0.0081 tree 0.0303 0.0098 0.1234 0.8317 0.0024 0.0024 car 0.1198 0.0074 0.0039 0.0216 0.8419 0.0055 clutter 0.2896 0.2096 0.2733 0.0243 0.0115 0.1916 Precision/Correctness 0.858 0.923 0.797 0.895 0.883 0.389 Recall/Completeness 0.886 0.947 0.847 0.832 0.842 0.192 F1 0.872 0.934 0.822 0.862 0.862 0.257 Overall accuracy 0.862 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 0.9157 0.0208 0.0386 0.0156 0.0005 0.0087 building 0.0287 0.9550 0.0052 0.0029 0.0002 0.0080 low_veg 0.0455 0.0133 0.8798 0.0544 0.0000 0.0069 tree 0.0228 0.0082 0.1068 0.8583 0.0018 0.0021 car 0.0402 0.0076 0.0019 0.0141 0.9305 0.0056 clutter 0.3072 0.2151 0.2357 0.0163 0.0123 0.2135 Precision/Correctness 0.886 0.946 0.830 0.921 0.938 0.380 Recall/Completeness 0.916 0.955 0.880 0.858 0.931 0.213 F1 0.901 0.950 0.854 0.889 0.934 0.273 Overall accuracy 0.892 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 0.8486 0.0324 0.0813 0.0196 0.0047 0.0135 building 0.0365 0.9446 0.0099 0.0030 0.0002 0.0058 low_veg 0.0298 0.0134 0.9102 0.0410 0.0001 0.0056 tree 0.0326 0.0052 0.2703 0.6832 0.0017 0.0069 car 0.0994 0.0064 0.0046 0.0285 0.8498 0.0115 clutter 0.1969 0.2134 0.2251 0.0168 0.0059 0.3419 Precision/Correctness 0.849 0.919 0.720 0.913 0.859 0.578 Recall/Completeness 0.849 0.945 0.910 0.683 0.850 0.342 F1 0.849 0.932 0.804 0.782 0.854 0.430 Overall accuracy 0.828 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 0.8798 0.0264 0.0665 0.0140 0.0012 0.0121 building 0.0341 0.9526 0.0062 0.0020 0.0002 0.0048 low_veg 0.0205 0.0081 0.9344 0.0326 0.0000 0.0044 tree 0.0273 0.0043 0.2552 0.7055 0.0015 0.0062 car 0.0316 0.0070 0.0020 0.0202 0.9279 0.0113 clutter 0.1873 0.2093 0.1915 0.0120 0.0065 0.3933 Precision/Correctness 0.877 0.939 0.741 0.937 0.913 0.607 Recall/Completeness 0.880 0.953 0.934 0.706 0.928 0.393 F1 0.878 0.946 0.826 0.805 0.920 0.477 Overall accuracy 0.853 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 0.9181 0.0165 0.0350 0.0258 0.0027 0.0020 building 0.0361 0.9408 0.0113 0.0071 0.0001 0.0047 low_veg 0.0519 0.0131 0.8171 0.1152 0.0002 0.0025 tree 0.0464 0.0045 0.0550 0.8920 0.0017 0.0004 car 0.1118 0.0016 0.0113 0.0314 0.8338 0.0102 clutter 0.5765 0.1253 0.1497 0.0617 0.0136 0.0733 Precision/Correctness 0.806 0.951 0.789 0.783 0.901 0.638 Recall/Completeness 0.918 0.941 0.817 0.892 0.834 0.073 F1 0.858 0.946 0.803 0.834 0.866 0.131 Overall accuracy 0.849 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 0.9390 0.0128 0.0277 0.0187 0.0004 0.0014 building 0.0305 0.9530 0.0066 0.0056 0.0000 0.0042 low_veg 0.0393 0.0079 0.8467 0.1044 0.0000 0.0017 tree 0.0344 0.0030 0.0425 0.9188 0.0011 0.0002 car 0.0324 0.0015 0.0039 0.0218 0.9298 0.0106 clutter 0.6085 0.1114 0.1336 0.0529 0.0142 0.0794 Precision/Correctness 0.831 0.965 0.828 0.820 0.936 0.672 Recall/Completeness 0.939 0.953 0.847 0.919 0.930 0.079 F1 0.882 0.959 0.837 0.866 0.933 0.142 Overall accuracy 0.876 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 0.8681 0.0246 0.0711 0.0146 0.0050 0.0167 building 0.0239 0.9517 0.0068 0.0063 0.0007 0.0105 low_veg 0.0787 0.0189 0.8028 0.0777 0.0006 0.0212 tree 0.0417 0.0076 0.1593 0.7824 0.0026 0.0063 car 0.0917 0.0149 0.0058 0.0288 0.8432 0.0155 clutter 0.4631 0.0511 0.2020 0.0108 0.0029 0.2701 Precision/Correctness 0.738 0.917 0.726 0.857 0.892 0.676 Recall/Completeness 0.868 0.952 0.803 0.782 0.843 0.270 F1 0.797 0.934 0.763 0.818 0.867 0.386 Overall accuracy 0.794 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 0.8908 0.0200 0.0633 0.0096 0.0007 0.0156 building 0.0204 0.9608 0.0037 0.0051 0.0006 0.0094 low_veg 0.0668 0.0141 0.8283 0.0705 0.0002 0.0200 tree 0.0328 0.0063 0.1412 0.8120 0.0020 0.0056 car 0.0335 0.0157 0.0026 0.0212 0.9111 0.0159 clutter 0.4781 0.0473 0.1888 0.0084 0.0023 0.2751 Precision/Correctness 0.751 0.934 0.748 0.882 0.947 0.702 Recall/Completeness 0.891 0.961 0.828 0.812 0.911 0.275 F1 0.815 0.947 0.786 0.846 0.929 0.395 Overall accuracy 0.815 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 0.8945 0.0192 0.0676 0.0073 0.0039 0.0075 building 0.1129 0.8377 0.0136 0.0052 0.0004 0.0302 low_veg 0.0566 0.0378 0.8437 0.0517 0.0009 0.0093 tree 0.0862 0.0098 0.1511 0.7410 0.0069 0.0051 car 0.1034 0.0160 0.0096 0.0092 0.8292 0.0326 clutter 0.3389 0.2159 0.1601 0.0237 0.0128 0.2486 Precision/Correctness 0.819 0.926 0.729 0.876 0.874 0.325 Recall/Completeness 0.894 0.838 0.844 0.741 0.829 0.249 F1 0.855 0.880 0.782 0.803 0.851 0.282 Overall accuracy 0.828 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 0.9129 0.0140 0.0619 0.0048 0.0008 0.0057 building 0.1117 0.8436 0.0114 0.0043 0.0003 0.0287 low_veg 0.0360 0.0315 0.8810 0.0444 0.0002 0.0068 tree 0.0703 0.0084 0.1305 0.7807 0.0054 0.0048 car 0.0281 0.0181 0.0032 0.0050 0.9082 0.0374 clutter 0.3578 0.2176 0.1399 0.0177 0.0109 0.2562 Precision/Correctness 0.839 0.944 0.754 0.904 0.928 0.309 Recall/Completeness 0.913 0.844 0.881 0.781 0.908 0.256 F1 0.874 0.891 0.813 0.838 0.918 0.280 Overall accuracy 0.852 Red/green image, indicating wrongly classified pixels Original, resized true ortho image for better comparison Classified image
============================================================= Tile top_potsdam_5_13_class.tif, reference set: full_reference ↓ predicted || reference → imp_surf building low_veg tree car clutter imp_surf 0.9074 0.0325 0.0240 0.0279 0.0032 0.0049 building 0.0218 0.9620 0.0051 0.0043 0.0004 0.0063 low_veg 0.0710 0.0229 0.7882 0.1146 0.0001 0.0031 tree 0.0497 0.0080 0.0951 0.8437 0.0024 0.0011 car 0.1220 0.0082 0.0006 0.0098 0.8534 0.0061 clutter 0.2734 0.3321 0.1313 0.0365 0.0705 0.1563 Precision/Correctness 0.889 0.910 0.856 0.755 0.853 0.537 Recall/Completeness 0.907 0.962 0.788 0.844 0.853 0.156 F1 0.898 0.935 0.820 0.797 0.853 0.242 Overall accuracy 0.866 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 0.9272 0.0266 0.0199 0.0216 0.0005 0.0043 building 0.0184 0.9681 0.0036 0.0035 0.0003 0.0062 low_veg 0.0642 0.0170 0.8119 0.1050 0.0000 0.0019 tree 0.0351 0.0062 0.0832 0.8727 0.0019 0.0009 car 0.0373 0.0088 0.0001 0.0067 0.9403 0.0069 clutter 0.2438 0.3377 0.1206 0.0359 0.0840 0.1780 Precision/Correctness 0.914 0.930 0.882 0.783 0.886 0.535 Recall/Completeness 0.927 0.968 0.812 0.873 0.940 0.178 F1 0.921 0.949 0.845 0.825 0.912 0.267 Overall accuracy 0.892 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 0.9125 0.0164 0.0251 0.0323 0.0034 0.0103 building 0.0308 0.9378 0.0061 0.0034 0.0002 0.0217 low_veg 0.0783 0.0201 0.7595 0.1358 0.0004 0.0059 tree 0.0494 0.0063 0.0675 0.8703 0.0031 0.0034 car 0.1103 0.0168 0.0046 0.0258 0.8281 0.0144 clutter 0.2218 0.2193 0.1710 0.0596 0.0152 0.3132 Precision/Correctness 0.827 0.900 0.827 0.782 0.857 0.693 Recall/Completeness 0.913 0.938 0.760 0.870 0.828 0.313 F1 0.868 0.918 0.792 0.824 0.842 0.431 Overall accuracy 0.833 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 0.9336 0.0132 0.0185 0.0250 0.0007 0.0091 building 0.0286 0.9431 0.0039 0.0028 0.0001 0.0216 low_veg 0.0626 0.0164 0.7861 0.1299 0.0001 0.0049 tree 0.0388 0.0054 0.0602 0.8903 0.0024 0.0029 car 0.0372 0.0176 0.0026 0.0167 0.9111 0.0147 clutter 0.2108 0.1968 0.1670 0.0545 0.0153 0.3554 Precision/Correctness 0.855 0.921 0.851 0.807 0.897 0.709 Recall/Completeness 0.934 0.943 0.786 0.890 0.911 0.355 F1 0.893 0.932 0.817 0.847 0.904 0.474 Overall accuracy 0.858 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 0.8812 0.0475 0.0385 0.0244 0.0031 0.0052 building 0.0707 0.8853 0.0133 0.0118 0.0015 0.0174 low_veg 0.0591 0.0183 0.7475 0.1679 0.0002 0.0071 tree 0.0378 0.0060 0.0696 0.8840 0.0017 0.0009 car 0.1223 0.0112 0.0048 0.0242 0.8224 0.0151 clutter 0.4085 0.1560 0.1357 0.0399 0.0041 0.2557 Precision/Correctness 0.850 0.867 0.746 0.719 0.910 0.694 Recall/Completeness 0.881 0.885 0.747 0.884 0.822 0.256 F1 0.865 0.876 0.747 0.793 0.864 0.374 Overall accuracy 0.818 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 0.8979 0.0448 0.0335 0.0188 0.0009 0.0040 building 0.0666 0.8942 0.0114 0.0100 0.0012 0.0166 low_veg 0.0404 0.0147 0.7768 0.1618 0.0001 0.0062 tree 0.0277 0.0043 0.0618 0.9041 0.0014 0.0007 car 0.0366 0.0121 0.0014 0.0208 0.9135 0.0157 clutter 0.4253 0.1259 0.1294 0.0383 0.0038 0.2773 Precision/Correctness 0.869 0.888 0.769 0.748 0.946 0.718 Recall/Completeness 0.898 0.894 0.777 0.904 0.913 0.277 F1 0.883 0.891 0.773 0.819 0.930 0.400 Overall accuracy 0.841 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 0.9236 0.0340 0.0180 0.0162 0.0021 0.0060 building 0.0463 0.9278 0.0143 0.0015 0.0002 0.0098 low_veg 0.0276 0.0082 0.9140 0.0496 0.0000 0.0006 tree 0.0972 0.0040 0.0883 0.8087 0.0013 0.0006 car 0.0900 0.0151 0.0005 0.0047 0.8505 0.0391 clutter 0.5456 0.0716 0.2229 0.0265 0.0152 0.1181 Precision/Correctness 0.893 0.930 0.895 0.837 0.887 0.313 Recall/Completeness 0.924 0.928 0.914 0.809 0.850 0.118 F1 0.908 0.929 0.904 0.822 0.868 0.172 Overall accuracy 0.891 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 0.9389 0.0314 0.0132 0.0111 0.0005 0.0049 building 0.0411 0.9386 0.0098 0.0008 0.0002 0.0094 low_veg 0.0174 0.0053 0.9342 0.0429 0.0000 0.0003 tree 0.0775 0.0032 0.0771 0.8408 0.0010 0.0004 car 0.0226 0.0192 0.0003 0.0015 0.9155 0.0409 clutter 0.5708 0.0619 0.2139 0.0220 0.0135 0.1179 Precision/Correctness 0.913 0.940 0.916 0.867 0.940 0.313 Recall/Completeness 0.939 0.939 0.934 0.841 0.916 0.118 F1 0.926 0.939 0.925 0.854 0.928 0.171 Overall accuracy 0.911 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 0.9460 0.0142 0.0200 0.0127 0.0056 0.0015 building 0.0720 0.8854 0.0066 0.0060 0.0018 0.0282 low_veg 0.0672 0.0142 0.8338 0.0818 0.0006 0.0024 tree 0.0584 0.0081 0.0921 0.8352 0.0054 0.0008 car 0.0875 0.0120 0.0030 0.0130 0.8697 0.0148 clutter 0.3510 0.3155 0.1825 0.0260 0.0215 0.1036 Precision/Correctness 0.884 0.943 0.849 0.795 0.877 0.194 Recall/Completeness 0.946 0.885 0.834 0.835 0.870 0.104 F1 0.914 0.913 0.841 0.815 0.874 0.135 Overall accuracy 0.878 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 0.9629 0.0108 0.0150 0.0096 0.0007 0.0010 building 0.0667 0.8943 0.0037 0.0056 0.0015 0.0282 low_veg 0.0491 0.0088 0.8658 0.0747 0.0001 0.0015 tree 0.0474 0.0069 0.0813 0.8595 0.0043 0.0007 car 0.0191 0.0122 0.0010 0.0080 0.9430 0.0167 clutter 0.3426 0.3495 0.1687 0.0205 0.0196 0.0990 Precision/Correctness 0.904 0.953 0.876 0.824 0.937 0.172 Recall/Completeness 0.963 0.894 0.866 0.859 0.943 0.099 F1 0.932 0.923 0.871 0.841 0.940 0.126 Overall accuracy 0.898 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 0.8740 0.0529 0.0332 0.0195 0.0071 0.0134 building 0.0218 0.9274 0.0420 0.0045 0.0001 0.0042 low_veg 0.0482 0.0295 0.7938 0.1189 0.0013 0.0083 tree 0.0351 0.0100 0.1317 0.8171 0.0040 0.0020 car 0.0722 0.0149 0.0066 0.0195 0.8664 0.0204 clutter 0.1609 0.0961 0.2138 0.0195 0.0141 0.4955 Precision/Correctness 0.918 0.870 0.679 0.820 0.858 0.784 Recall/Completeness 0.874 0.927 0.794 0.817 0.866 0.496 F1 0.895 0.898 0.732 0.819 0.862 0.607 Overall accuracy 0.843 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 0.8952 0.0502 0.0264 0.0146 0.0014 0.0122 building 0.0177 0.9358 0.0394 0.0033 0.0001 0.0036 low_veg 0.0303 0.0276 0.8241 0.1111 0.0002 0.0068 tree 0.0269 0.0080 0.1171 0.8431 0.0032 0.0017 car 0.0175 0.0160 0.0020 0.0135 0.9300 0.0210 clutter 0.1549 0.0944 0.2053 0.0159 0.0124 0.5171 Precision/Correctness 0.938 0.883 0.698 0.850 0.927 0.810 Recall/Completeness 0.895 0.936 0.824 0.843 0.930 0.517 F1 0.916 0.908 0.756 0.847 0.929 0.631 Overall accuracy 0.865 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 0.9091 0.0379 0.0096 0.0303 0.0029 0.0102 building 0.0595 0.8912 0.0032 0.0035 0.0024 0.0403 low_veg 0.0693 0.0077 0.7241 0.1947 0.0004 0.0038 tree 0.0290 0.0021 0.0937 0.8738 0.0010 0.0003 car 0.0811 0.0216 0.0010 0.0279 0.8667 0.0017 clutter 0.2730 0.0349 0.0592 0.0024 0.0125 0.6179 Precision/Correctness 0.903 0.896 0.801 0.653 0.787 0.793 Recall/Completeness 0.909 0.891 0.724 0.874 0.867 0.618 F1 0.906 0.894 0.761 0.748 0.825 0.695 Overall accuracy 0.857 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 0.9232 0.0353 0.0073 0.0231 0.0011 0.0100 building 0.0560 0.8970 0.0021 0.0023 0.0021 0.0405 low_veg 0.0561 0.0035 0.7514 0.1857 0.0001 0.0032 tree 0.0188 0.0008 0.0877 0.8918 0.0006 0.0003 car 0.0153 0.0265 0.0002 0.0185 0.9383 0.0012 clutter 0.2648 0.0331 0.0579 0.0020 0.0113 0.6309 Precision/Correctness 0.913 0.906 0.820 0.683 0.826 0.799 Recall/Completeness 0.923 0.897 0.751 0.892 0.938 0.631 F1 0.918 0.901 0.784 0.774 0.879 0.705 Overall accuracy 0.872 Red/green image, indicating wrongly classified pixels Original, resized true ortho image for better comparison Classified image