30 lines
1.3 KiB
Markdown
30 lines
1.3 KiB
Markdown
# Wavelet Denoise
|
||
|
||
Denoise a DATA_FIELD using wavelet coefficient thresholding. BayesShrink adapts the threshold per sub-band; VisuShrink uses a global threshold. Equivalent to applying DWT from Gwyddion dwt.c with coefficient thresholding.
|
||
|
||
## Inputs
|
||
|
||
| Name | Type | Required | Description |
|
||
|------|------|----------|-------------|
|
||
| field | DATA_FIELD | Yes | Input field to denoise |
|
||
|
||
## Outputs
|
||
|
||
| Name | Type | Description |
|
||
|------|------|-------------|
|
||
| denoised | DATA_FIELD | Denoised field |
|
||
|
||
## Controls
|
||
|
||
| Name | Type | Default | Description |
|
||
|------|------|---------|-------------|
|
||
| wavelet | dropdown | db4 | Wavelet family: db1 (Haar), db2, db4, db8, sym4, coif1, or bior1.3 |
|
||
| method | dropdown | BayesShrink | Threshold estimation method: BayesShrink (per sub-band adaptive) or VisuShrink (global universal) |
|
||
| sigma | FLOAT | 0.0 | Noise level estimate in data units; 0 = automatic estimation (0–1.0) |
|
||
| mode | dropdown | soft | Thresholding mode: soft (smooth shrinkage) or hard (zero below threshold) |
|
||
|
||
## Limitations
|
||
|
||
- The field size should ideally be a power of two for best wavelet decomposition; other sizes are handled by padding.
|
||
- sigma = 0 triggers automatic noise estimation from the finest-scale wavelet coefficients; this may be inaccurate on strongly structured surfaces.
|