Research Article
HMFT: Hyperspectral and Multispectral Image Fusion Super-Resolution Method Based on Efficient Transformer and Spatial-Spectral Attention Mechanism
Table 4
QIs of the results by different methods on specified test computer images from the Harvard dataset.
| Computer (512 × 512 × 31) | Method | PSNR | SAM | ERGAS | RMSE | CC | SSIM |
| FUSE | 36.72 | 2.71 | 8.8726 | 4.4762 | 0.9875 | 0.9705 | GLP-HS | 49.56 | 1.65 | 2.48 | 0.9116 | 0.9991 | 0.9932 | CSTF | 48.21 | 14.27 | 2.99 | 1.0664 | 0.9986 | 0.9835 | MHFnet | 47.18 | 4.48 | 7.58 | 0.0047 | 0.9628 | 0.9840 | HSRnet | 43.36 | 2.90 | 3.05 | 0.0074 | 0.9916 | 0.9770 | Ours | 52.08 | 2.78 | 2.17 | 0.0026 | 0.9996 | 0.9948 | Best value | +∞ | 0 | 0 | 0 | 1 | 1 |
|
|
The best values are underlined.
|