Research Article
Feature-Based Retinal Image Registration Using D-Saddle Feature
Table 4
Correlation between registration accuracy and image quality metrics in FIRE dataset.
| | | Registration accuracy | GDB-ICP | Harris-PIIFD | H-M | Saddle | D-Saddle |
| Intensity difference (MSE) | Spearman’s rho (rs) | 0.205 | 0.395 | 0.326 | 0.341 | 0.269 | value (p) | 0.064 | <0.001 | <0.001 | <0.001 | 0.002 | Structure similarity (SSIM) | Spearman’s rho (rs) | −0.030 | −0.094 | −0.013 | −0.071 | −0.015 | value (p) | 0.788 | 0.288 | 0.878 | 0.417 | 0.863 | Intensity uniformity (UN) | Spearman’s rho (rs) | 0.193 | −0.093 | 0.037 | 0.016 | 0.001 | value (p) | 0.083 | 0.297 | 0.673 | 0.853 | 0.994 |
|
|
Correlation is significant at the 0.05 level. Correlation is significant at the 0.01 level.
|