L+M
| Dataset | Fitting error | Sensitivity adjustment |
| Muli-layer Perceptron | Muli-layer Perceptron | |
| Average training | 4.83 ± 0.27 [dB] | N/A |
| Average testing | 5.83 ± 0.21 [dB] | N/A |
| modelfest | 3.15 [dB] | 1.000 |
| hdrvdp_csf | 5.34 [dB] | 1.000 |
| rovamo1993 | 5.22 [dB] | 1.000 |
| laird2006 | 1.99 [dB] | 1.000 |
| snowden1995 | 3.93 [dB] | 1.000 |
| robson1966 | 0.70 [dB] | 1.000 |
| virsu1979 | 3.53 [dB] | 1.000 |
| virsu1982 | 0.75 [dB] | 1.000 |
| wright1983 | 0.36 [dB] | 1.000 |
| colorfest | 4.51 [dB] | 1.000 |
| hdr_csf | 5.02 [dB] | 1.000 |
| kim2013 | 7.70 [dB] | 1.000 |
| five_centres | 6.12 [dB] | 1.000 |
| lucassen2018 | 4.58 [dB] | 1.000 |
| hdr_csf_disc | 7.11 [dB] | 1.000 |
| kong2018 | 4.03 [dB] | 1.000 |
| vanderHorst1969_b | 3.59 [dB] | 1.000 |
| hansen2009 | 2.39 [dB] | 1.000 |
To keep the plots legible, only up to 3 models are plotted.
