Table 2.
Quantitative evaluation of the sparse reconstructions using Fu’s, Arad’s and our MSFA (see Figure 7) on synthetic datasets. Note that the bold font indicates the minimal error.
| MSFA | Dataset | ||||||
|---|---|---|---|---|---|---|---|
| CAVE | Harvard | ICVL | |||||
| RMSE | PSNR | RMSE | PSNR | RMSE | PSNR | ||
| Noisy (SNR = ∞) |
Ours | 0.258 | 20.21 | 0.212 | 21.97 | 0.071 | 28.59 |
| Fu’s | 0.322 | 19.89 | 0.219 | 21.63 | 0.086 | 28.57 | |
| Arad | 0.386 | 19.75 | 0.232 | 21.14 | 0.094 | 28.51 | |
| Noisy (SNR ≈ 30db) |
Ours | 0.281 | 17.86 | 0.246 | 17.97 | 0.095 | 24.26 |
| Fu’s | 0.328 | 16.36 | 0.257 | 16.77 | 0.109 | 22.33 | |
| Arad | 0.440 | 15.32 | 0.288 | 16.03 | 0.156 | 20.98 | |