add keywords for all nodes
This commit is contained in:
@@ -19,13 +19,13 @@ Restore an image via regularised deconvolution. Assumes the image was blurred by
|
||||
| Name | Type | Default | Description |
|
||||
|------|------|---------|-------------|
|
||||
| method | dropdown | wiener | Deconvolution method: wiener or richardson_lucy |
|
||||
| sigma | FLOAT | 2.0 | Gaussian PSF sigma in pixels (0.1–50.0) |
|
||||
| regularisation | FLOAT | 0.01 | Regularisation parameter for Wiener filter (1e-6–1.0) |
|
||||
| iterations | INT | 10 | Number of iterations (Richardson-Lucy only, 1–200) |
|
||||
| sigma | FLOAT | 2.0 | Gaussian PSF sigma in pixels (0.1-50.0) |
|
||||
| regularisation | FLOAT | 0.01 | Regularisation parameter for Wiener filter (1e-6-1.0) |
|
||||
| iterations | INT | 10 | Number of iterations (Richardson-Lucy only, 1-200) |
|
||||
|
||||
## Notes
|
||||
|
||||
- **Wiener**: Fast, single-pass frequency-domain filter. The regularisation parameter controls the noise/sharpness tradeoff — smaller values sharpen more but amplify noise.
|
||||
- **Richardson-Lucy**: Iterative method that preserves positivity. More iterations = sharper result but risk of ringing artifacts.
|
||||
- The PSF sigma should match the actual blur in the image. If unknown, start with sigma=1–3 and adjust.
|
||||
- The PSF sigma should match the actual blur in the image. If unknown, start with sigma=1-3 and adjust.
|
||||
- For tip-shape deconvolution (non-Gaussian PSF), use Tip Deconvolution instead.
|
||||
|
||||
Reference in New Issue
Block a user