Sensitive Nonlinear Iterative Peak-clipping (SNIP)
Table of contents
Introduction
Sensitive Nonlinear Iterative Peak-clipping (SNIP) estimates a slowly varying spectral background by repeatedly clipping peak-like structures from a transformed spectrum.
How to use
- Upload data, then open Processing Page.
- Enable Baseline removal.
- In Select baseline removal function, choose SNIP.
- Set SNIP Iterations.
- Apply the processing workflow to subtract the estimated baseline.
Behavior
SNIP estimates a slowly varying background by iteratively clipping peaks from a transformed spectrum. Increasing SNIP Iterations allows the baseline estimate to remove wider peak structures, usually producing a smoother and lower background.
Method
SpectraGuru applies a log-log-square-root transform before iterative clipping:
\[v = \log(\log(\sqrt{\max(y,0)+1}+1)+1)\]The clipping update replaces each point with the smaller of the current value and the average of points at a symmetric offset:
\[v_i \leftarrow \min\left(v_i,\frac{v_{i-k}+v_{i+k}}{2}\right)\]| Parameter | Tunable or fixed | Implementation |
|---|---|---|
| SNIP Iterations | Tunable | Default 50, UI range 10-200 |
| Transform | Fixed | Log-log-square-root transform with inverse transform after clipping |
| Baseline action | Fixed | Estimated baseline is subtracted from each selected spectrum |
References
- Morhac, M., Kliman, J., Matousek, V., Veselsky, M., & Turzo, I. (1997). Background elimination methods for multidimensional coincidence gamma-ray spectra. Nuclear Instruments and Methods in Physics Research Section A, 401(1), 113-132. https://doi.org/10.1016/S0168-9002(97)01023-1