Semantic Segmentation on Trans10K v2
95.01AccuracyTrans4Trans-M
Evaluation Results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Trans4Trans-MGFLOPs=34.382021.08 | 95.01 | 75.14 | 96.08 | 5,581 | 71.46 | 6,925 | 65.16 | 63.96 | 83.84 | 88.21 | 80.29 | 76.33 | 8,309 | 68.09 | |
| Trans4Trans-SGFLOPs=19.922021.08 | 94.57 | 74.15 | 95.6 | 5,705 | 71.18 | 7,021 | 63.95 | 61.25 | 81.67 | 87.34 | 78.52 | 77.13 | 8,100 | 64.88 | |
| Trans2SegGFLOPs=49.032021.08 | 94.14 | 72.15 | 95.35 | 5,343 | 67.82 | 6,420 | 59.64 | 60.56 | 88.52 | 86.67 | 75.99 | 73.98 | 8,243 | 57.17 | |
| Trans4Trans-TGFLOPs=10.452021.08 | 93.23 | 68.63 | 94.44 | 4,839 | 61.89 | 6,186 | 61.14 | 54.83 | 73.6 | 83.03 | 75.2 | 74.69 | 7,526 | 59.19 | |
| DeepLabv3+GFLOPs=37.982021.08 | 92.75 | 68.87 | 93.82 | 5,129 | 64.65 | 6,571 | 55.26 | 57.19 | 77.06 | 81.89 | 72.64 | 70.81 | 7,744 | 58.63 | |
| DANetGFLOPs=1982021.08 | 92.7 | 68.81 | 93.69 | 4,769 | 66.05 | 7,018 | 53.01 | 56.15 | 77.73 | 82.89 | 72.24 | 72.18 | 7,787 | 56.06 | |
| TransLabGFLOPs=61.312021.08 | 92.67 | 69 | 93.9 | 5,436 | 64.48 | 6,514 | 54.58 | 57.72 | 79.85 | 81.61 | 72.82 | 69.63 | 7,750 | 56.43 | |
| PSPNetGFLOPs=187.032021.08 | 92.47 | 68.23 | 93.62 | 5,033 | 64.24 | 7,019 | 51.51 | 55.27 | 79.27 | 81.93 | 71.95 | 68.91 | 7,713 | 54.43 | |
| OCNetGFLOPs=43.312021.08 | 92.03 | 66.31 | 93.12 | 4,147 | 63.54 | 6,005 | 54.1 | 51.01 | 79.57 | 81.95 | 69.4 | 68.44 | 7,841 | 54.65 | |
| FCNGFLOPs=42.232021.08 | 91.65 | 62.75 | 93.62 | 3,884 | 56.05 | 5,876 | 46.91 | 50.74 | 82.56 | 78.71 | 68.78 | 57.87 | 7,366 | 46.54 | |
| DenseASPPGFLOPs=36.22021.08 | 90.86 | 63.01 | 91.39 | 4,241 | 60.93 | 6,475 | 48.97 | 51.4 | 65.72 | 75.64 | 67.93 | 67.03 | 7,026 | 49.64 | |
| DUNetGFLOPs=123.692021.08 | 90.67 | 59.01 | 93.07 | 3,420 | 50.95 | 5,496 | 43.19 | 45.05 | 79.8 | 76.07 | 65.29 | 54.33 | 6,857 | 42.64 | |
| HarDNetGFLOPs=4.422021.08 | 90.19 | 56.19 | 92.87 | 3,462 | 47.5 | 4,240 | 49.78 | 49.19 | 62.33 | 72.93 | 68.32 | 58.14 | 6,533 | 30.9 | |
| HRNet_w18GFLOPs=4.22021.08 | 89.58 | 54.25 | 92.47 | 2,766 | 45.08 | 4,053 | 45.66 | 45 | 68.05 | 73.24 | 64.86 | 52.85 | 6,252 | 33.02 | |
| BiSeNetGFLOPs=19.912021.08 | 89.13 | 58.4 | 90.12 | 3,954 | 53.71 | 5,090 | 46.95 | 44.68 | 64.32 | 72.86 | 63.57 | 61.38 | 6,788 | 44.85 | |
| DeepLabv3+MBv2GFLOPs=2.622021.08 | 88.39 | 54.16 | 89.95 | 3,179 | 48.29 | 4,618 | 41.39 | 43.42 | 61.97 | 69.48 | 61.65 | 54.89 | 6,347 | 37.36 | |
| FastSCNNGFLOPs=1.012021.08 | 88.05 | 51.93 | 90.64 | 3,276 | 41.12 | 4,728 | 47.47 | 44.64 | 48.99 | 67.88 | 63.8 | 55.08 | 5,886 | 24.65 | |
| RefineNetGFLOPs=44.562021.08 | 87.99 | 58.18 | 90.63 | 3,062 | 53.17 | 5,595 | 42.72 | 46.59 | 70.85 | 76.01 | 62.91 | 57.05 | 7,034 | 41.32 | |
| ContextNetGFLOPs=0.872021.08 | 86.75 | 46.69 | 89.86 | 2,322 | 34.88 | 3,234 | 44.24 | 42.25 | 50.36 | 65.23 | 60 | 43.88 | 5,381 | 20.17 | |
| LEDNetGFLOPs=6.232021.08 | 86.07 | 46.4 | 88.59 | 2,813 | 36.72 | 3,245 | 43.77 | 38.55 | 41.51 | 64.19 | 60.05 | 42.4 | 5,312 | 27.29 | |
| DFANetGFLOPs=1.022021.08 | 85.15 | 42.54 | 88.49 | 2,665 | 27.84 | 2,894 | 46.27 | 39.47 | 33.06 | 58.87 | 59.45 | 43.22 | 4,487 | 13.37 | |
| U-NetGFLOPs=124.552021.08 | 81.9 | 29.23 | 86.34 | 876 | 15.18 | 1,902 | 27.13 | 24.73 | 17.26 | 53.4 | 47.36 | 11.97 | 3,779 | 1.77 | |
| ICNetGFLOPs=10.642021.08 | 78.23 | 23.39 | 83.29 | 296 | 4.91 | 933 | 19.24 | 15.35 | 24.11 | 44.54 | 41.49 | 7.58 | 2,747 | 3.8 | |
| DABNetGFLOPs=5.182021.08 | 77.43 | 15.27 | 81.19 | 0 | 0.09 | 0 | 4.1 | 10.49 | 0 | 36.18 | 42.83 | 0 | 830 | 0 | |
| ESPNetv2GFLOPs=0.832021.08 | 73.03 | 12.27 | 78.98 | 0 | 0 | 0 | 0 | 6.17 | 0 | 30.65 | 37.03 | 0 | 0 | 0 | |
| ENetGFLOPs=2.092021.08 | 71.67 | 8.5 | 79.74 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 22.25 | 0 | 0 | 0 | |
| FPENetGFLOPs=0.762021.08 | 70.31 | 10.14 | 74.97 | 1 | 0 | 2 | 2.11 | 2.83 | 0 | 16.84 | 24.81 | 0 | 4 | 0 |