Semantic Segmentation on SUN RGB-D (val)
74.99Wall mIoUDeepLab v3
Evaluation results
| Method | Links | ||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DeepLab v3Input image dim=768 x 384, Left-right flips=false, Multiscale testing=false2018.08 | 74.99 | 85.44 | 39.64 | 54.14 | 64.23 | 45.84 | 47.35 | 38.3 | 45.5 | 30.7 | 45.3 | 15.21 | 24.92 | 18.18 | 6.37 | 48.84 | 28.2 | 32.87 | 31.23 | 0 | 16.98 | 60.09 | 31.74 | 37.86 | 33.8 | 17.84 | 20.51 | 0 | 19.58 | 44.1 | 31.36 | 18.15 | 68.76 | 52.01 | 35.36 | 45.97 | 10.37 | 35.74 | |
| AdapNet++Input image dim=768 x 384, Left-right flips=false, Multiscale testing=false2018.08 | 73.81 | 84.79 | 47.45 | 64.31 | 65.76 | 52.15 | 49.97 | 41.66 | 46.99 | 32.83 | 45.19 | 32.71 | 34.72 | 21.16 | 5.03 | 50.32 | 37.45 | 32.35 | 25.89 | 0.12 | 21.68 | 60.65 | 32.41 | 46.96 | 18.81 | 20.79 | 27.95 | 1.32 | 22.7 | 51.08 | 33.77 | 20.59 | 71.28 | 55.31 | 36.74 | 38.6 | 15.34 | 38.4 | |
| AdapNetInput image dim=768 x 384, Left-right flips=false, Multiscale testing=false2018.08 | 72.82 | 86.61 | 41.9 | 56.63 | 64.99 | 46.14 | 46.74 | 36.14 | 44.43 | 28.4 | 46.53 | 25.17 | 31.07 | 13.41 | 3.98 | 47.9 | 28.58 | 32.9 | 19.25 | 0 | 17.3 | 52.6 | 30.63 | 20.32 | 16.56 | 18.21 | 13.69 | 0 | 15.72 | 43.02 | 17.48 | 9.44 | 61.11 | 49.43 | 31.32 | 21.56 | 9.38 | 32.47 | |
| ParseNetInput image dim=768 x 384, Left-right flips=false, Multiscale testing=false2018.08 | 70.83 | 81.85 | 40.25 | 55.64 | 61.12 | 45.06 | 45.86 | 30.41 | 41.73 | 34.45 | 44.11 | 28.63 | 32.59 | 12.78 | 8.58 | 46.51 | 34.05 | 28.07 | 24.82 | 0.02 | 19.1 | 52.28 | 34.36 | 27.05 | 31.49 | 21.65 | 21.49 | 6.46 | 20.77 | 44.48 | 34.59 | 15.37 | 66.96 | 47.15 | 31.49 | 28.14 | 13.19 | 34.68 | |
| SegNetInput image dim=768 x 384, Left-right flips=false, Multiscale testing=false2018.08 | 70.18 | 82.01 | 37.62 | 45.77 | 57.22 | 35.86 | 40.6 | 30.32 | 39.37 | 29.67 | 40.67 | 22.48 | 32.13 | 14.79 | 5.15 | 39.69 | 27.46 | 26.05 | 15.64 | 0 | 16.01 | 53.89 | 26.21 | 12.95 | 24.23 | 14.57 | 10.88 | 0 | 13.5 | 38.02 | 3.83 | 0.76 | 60.98 | 45.97 | 22.82 | 31.94 | 8.97 | 29.14 | |
| FCN-8sInput image dim=768 x 384, Left-right flips=false, Multiscale testing=false2018.08 | 68.57 | 79.94 | 37.27 | 52.85 | 58.42 | 42.79 | 43.69 | 27.58 | 43.49 | 28.46 | 42.21 | 27.08 | 28.34 | 11.38 | 6.01 | 46.96 | 28.8 | 27.73 | 18.97 | 0 | 20.34 | 51.98 | 30.98 | 9.06 | 34.23 | 15.95 | 12.79 | 0 | 18.24 | 41.38 | 8.92 | 0.97 | 59.64 | 44.21 | 22.83 | 26.2 | 8.81 | 30.46 | |
| FastNetInput image dim=768 x 384, Left-right flips=false, Multiscale testing=false2018.08 | 68.39 | 79.73 | 34.98 | 44.93 | 55.26 | 36.68 | 38.57 | 26.88 | 38.28 | 26.02 | 39.72 | 20.55 | 22.47 | 11.71 | 5.67 | 41.98 | 25.08 | 22.76 | 11.76 | 0.1 | 13.37 | 49.14 | 29.38 | 17.49 | 19.8 | 17.95 | 13.55 | 2.93 | 15.33 | 38.12 | 15.82 | 7.93 | 54.08 | 39.59 | 22.2 | 23.71 | 9.8 | 28.15 | |
| DeepLab v2Input image dim=768 x 384, Left-right flips=false, Multiscale testing=false2018.08 | 63.68 | 74.07 | 31.15 | 47 | 55.46 | 40.81 | 40.6 | 29.1 | 37.95 | 18.67 | 35.67 | 23.2 | 25.08 | 13.52 | 2.55 | 42.91 | 24.56 | 29.81 | 27.16 | 0 | 15.85 | 43.47 | 25.39 | 35.55 | 23.21 | 16.53 | 22 | 0.34 | 13.56 | 22.32 | 46.18 | 11.45 | 55.94 | 43.33 | 29.52 | 32.73 | 7.63 | 30.06 | |
| DeconvNetInput image dim=768 x 384, Left-right flips=false, Multiscale testing=false2018.08 | 61.53 | 75.52 | 26.97 | 38.52 | 48.46 | 30.94 | 37.12 | 21.99 | 27.48 | 0 | 0 | 0 | 0 | 9.4 | 0 | 29.59 | 0 | 0 | 8.32 | 0 | 0 | 46.4 | 0 | 0 | 0 | 0 | 0 | 0 | 4.43 | 32.06 | 0 | 0 | 47.35 | 32.85 | 0 | 0 | 0 | 15.64 |