Skip to main content
. 2024 Apr 27;24(9):2795. doi: 10.3390/s24092795

Table 2.

Low-resolution comparison experiments on the AIM dataset.

Model MAD (1 × 103) MSE (1 × 103) Grad Conn dtSSD Fgr (1 × 103)
DeepLabV3 29.64 23.78 20.17 7.71 432 -
FBA 23.45 17.66 9.05 6.05 229 6.32
BGMv2 44.61 39.08 5.54 11.60 269 3.31
MODNet 21.66 14.27 5.37 5.23 176 9.51
RVM 14.84 8.93 4.35 3.83 101 5.01
RVM-Large 13.48 4.58 3.95 3.38 98 4.79
Our Model 11.86 4.28 3.73 3.25 94 4.55
Ours + CE 11.32 4.19 3.68 3.09 90 4.33