Table 1 The learned \(\{\beta _k\}\) for depth images with different levels of noise
From: Optimal graph edge weights driven nlms with multi-layer residual compensation
| Â | \(\beta _0\) | \(\beta _1\) | \(\beta _2\) | \(\beta _3\) | \(\beta _4\) |
|---|---|---|---|---|---|
\(\sigma <5\) | 1.0223 | 0.0002 | 0.0001 | 0.0001 | 0.0021 |
\(\sigma =10\) | 0.9992 | 0.0000 | − 0.0005 | − 0.0001 | 0.0002 |
\(\sigma =20\) | 0.9976 | − 0.0003 | − 0.0004 | − 0.0004 | − 0.0002 |
\(\sigma =30\) | 0.9766 | − 0.0006 | − 0.0007 | − 0.0009 | − 0.0001 |
\(\sigma =40\) | 0.9885 | − 0.0012 | − 0.0013 | − 0.0019 | 0.0011 |
\(\sigma =50\) | 0.9862 | − 0.0021 | − 0.0025 | − 0.0032 | − 0.0016 |