Supplementary Materials for: "Morphological reconstruction based on U-Net segmentation of optical images for the seabed semi-buried polymetallic nodules"

Supplementary Materials for: “Morphological reconstruction based on U-Net segmentation of optical images for the seabed semiburied polymetallic nodules” Xiang Meng 1,2, Kehong Yang 2,*, Mingwei Wang 2,*, Qian Yu 1, and Zinyin Wu 2 1School of Geography and Ocean Science, Nanjing University, Nanjing, 210093, China 2State Key Laboratory of Submarine Geoscience & Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou, 310012, China Abstract This supplementary document provides additional details and results supporting the main manuscript “Morphological reconstruction based on U-net segmentation of optical images for the seabed semiburied polymetallic nodules” . It includes extended tables not included in the main publication. 1. Additional Tables Table S1. U-Net versus two fitting methods on each metric (boundaries region). Metric Method Maximum Minimum Mean Median Std Dev AR U-Net 95.66% 47.08% 81.62% 85.43% 0.1297 EAM 222.39% 79.69% 111.88% 105.06% 0.2585 CIM 106.39% 66.06% 95.52% 97.81% 0.0766 APC U-Net 52.92% 3.34% 18.38% 14.57% 0.1297 EAM 122.39% 0.97% 16.76% 8.35% 0.2293 CIM 33.94% 0.05% 5.94% 3.36% 0.0657 IoU U-Net 93.83% 47.08% 81.11% 85.02% 0.1291 EAM 95.29% 43.18% 83.33% 87.07% 0.1165 CIM 98.29% 64.56% 91.14% 94.03% 0.0767 CD U-Net 2.9101 0.1351 0.6699 0.3915 0.6162 EAM 5.0425 0.0325 0.5149 0.2146 0.8792 CIM 1.9361 0.0089 0.2112 0.0476 0.3574 ∗ Corresponding author Email addresses: ya[email protected] (Kehong Yang), wan[email protected] (Mingwei Wang) 2 Table S2. Binary shapes (BS) versus two fitting methods on each metric. Metric Method Maximum Minimum Mean Median Std Dev AR BS 92.43% 20.7% 64.1% 64.5% 0.1586 EAM 162.67% 24.27% 110.29% 110.57% 0.1722 CIM 101.12% 14.44% 90.67% 92.94% 0.1183 APC BS 87.40% 7.57% 36.06% 35.5% 0.1633 EAM 75.73% 0.04% 14.34% 10.98% 0.1398 CIM 85.56% 0.06% 9.44% 7.06% 0.1175 IoU BS 91.47% 12.35% 63.01% 63.23% 0.1613 EAM 94.89% 23.79% 82.61% 84.45% 0.0996 CIM 98.85% 14.15% 88.9% 90.83% 0.1143 CD BS 9.1221 0.0919 1.3684 0.8611 1.5048 EAM 6.3361 0.0403 0.6286 0.3079 0.8369 CIM 8.7895 0.0065 0.4299 0.0863 1.0847 Table S3. EA-CIM on each metric (binary images). Metirc Maximum Minimum Mean Median Std Dev AR 106.62% 24.27% 93.93% 95.47% 0.1006 APC 75.73% 0.06% 6.87% 4.68% 0.0952 IoU 98.85% 23.79% 90.21% 92.1% 0.0927 CD 6.3361 0.0065 0.2964 0.082 0.7501