Table 1 Comparison between different distance metrics used to calculate the PANN change score for different natural disaster classes.

From: Training-free AI for earth observation change detection using physics aware neuromorphic networks

Distance metric

Fires

Floods

Hurricanes

Landslides

Euclidean

\(90.084 \pm 0.007\)

\(42.79 \pm 0.04\)

\(43.1 \pm 0.07\)

\(27.38 \pm 0.09\)

Cosine

\(86.71 \pm 0.01\)

\(39.36 \pm 0.03\)

\(22.50 \pm 0.01\)

\(35.2 \pm 0.2\)

Correlation

\(\mathbf {93.66 \pm 0.01}\)

\(\mathbf {59.26 \pm 0.02}\)

\(\mathbf {76.51 \pm 0.05}\)

\(\mathbf {73.5 \pm 0.2}\)

  1. All values are calculated as mean AUPRC (%) ± SEM over 5 runs. Best scores are highlighted.