Super-resolution for OCT Medical Images

Evaluation Metrics

• MSR (Mean-to-Standard Deviation Ratio)

$$ \text{MSR} = \left( \frac{\mu_f}{\mu_b} \right) $$

• CNR (Contrast-to-Noise Ratio)

$$ \text{CNR} = \frac{|\mu_f - \mu_b|}{\sqrt{\frac{\sigma_f^2 + \sigma_b^2}{2}}} $$

• $\mu_f$ is the mean intensity of the foreground.
• $\mu_b$ is the mean intensity of the background.
• $\sigma_f$ is the standard deviation of the foreground.
• $\sigma_b$ is the standard deviation of the background.

Source : Reduction of speckle noise from optical coherence tomography images using multi-frame weighted nuclear norm minimization method