BSE Field Correction
BSE Field Correction corrects two common imaging artifacts in backscattered electron images:
- Vignetting — radial brightness falloff where the center of each field appears brighter than the edges, caused by detector geometry and optics
- Photometric drift — per-field brightness and contrast shifts caused by temperature fluctuations, beam instability, or detector gain changes over the course of a long measurement
{{FIG: Before/after comparison showing a multi-field mosaic with visible vignetting (brightness rings) on the left and the corrected uniform image on the right | BSE field correction before/after}}
When to Use BSE Field Correction
Use BSE Field Correction when:
- You can see brightness rings or center-to-edge falloff within individual fields
- Adjacent fields have noticeably different brightness despite measuring the same material
- Long measurements show progressive brightness drift across the dataset
- Quantitative BSE analysis requires consistent grey levels across all fields
Artifact | Best Tool |
|---|---|
Uniform brightness offset between fields | BSE Adjustor |
Radial brightness falloff (vignetting) + drift | BSE Field Correction |
Both vignetting and uniform offset | BSE Field Correction |
Accessing BSE Field Correction
From the Process ribbon Processing tab:
- Click the BSE Adjust dropdown in the Tools panel
- Select BSE Field Correction
{{FIG: Process ribbon Tools panel showing BSE Adjust dropdown with BSE Field Correction option | BSE Adjust dropdown menu}}
Workflow
First-Time Correction
- Click BSE Adjust → BSE Field Correction
- AMICS prompts: "Perform BSE correction?" → Click Yes
- A progress bar appears while correction parameters are calculated
- The corrected BSE images are displayed immediately
{{FIG: Progress dialog during BSE field correction parameter calculation | BSE correction progress}}
Re-Applying Existing Correction
If correction parameters have already been calculated for this measurement:
- Click BSE Adjust → BSE Field Correction
- AMICS prompts: "Correction already available. Recalculate?"
- Yes — recalculate from scratch (use if you've modified the dataset)
- No — apply the existing parameters immediately (faster)
- Cancel — abort
Toggling Correction On/Off
Once calculated, the correction can be toggled:
- Use the Apply BSE Shift checkbox to switch between corrected and original BSE images
- This lets you compare the before and after images instantly
Saving
- Correction is applied in-memory only until you explicitly save
- Original BSE images in the FLD files are never modified — the correction is non-destructive
- When you save, correction parameters are written as a JSON file in the MRE folder
- Parameters include: vignetting polynomial coefficients, per-tile gain and offset values
How the Correction Works
Step 1: Vignetting Correction (Flat Field)
AMICS samples approximately 200 random fields and computes a radial brightness profile — measuring how intensity varies from the center to the edges. A 4th-degree polynomial is fit to this profile, and a flat-field correction image is generated.
The flat-field image is the same for all fields in the measurement, since vignetting is caused by the optical system geometry.
Step 2: Photometric Correction (Per-Field Gain/Offset)
For each pair of overlapping fields, AMICS compares the pixel intensities in the overlapping region using QQ-regression (quantile-quantile regression), which is robust to outliers. This produces a gain and offset value describing how one field's brightness relates to its neighbor's.
These pairwise relationships are propagated across the entire measurement grid using a least-squares optimization, producing a consistent global gain and offset for each field.
Step 3: Apply Correction
Each pixel is corrected as:
corrected = pixel × flat_field × gain + offsetWith these rules:
- Background pixels (grey level 0-25) are left unchanged
- Zero pixels (empty/masked) remain zero
- All other pixels receive the full vignetting + photometric correction
- Results are clamped to a minimum of 1 (non-zero pixels never become zero)
Performance
BSE Field Correction is optimized for large datasets:
- Flat-field calculation uses random tile sampling (not all tiles)
- QQ-regression is parallelized across tile pairs
- Per-pixel correction is trivially parallelizable
- Memory cost: one additional BSE image copy per loaded field (same as BSE Adjustor)
Correction Parameters
Correction values are automatically clamped to prevent extreme adjustments:
Parameter | Range | Purpose |
|---|---|---|
Gain | 0.5 — 2.0 | Per-tile brightness scaling |
Offset | -50 — +50 | Per-tile brightness shift |
Background range | 0 — 25 | Grey levels treated as background (uncorrected) |
Limitations
- Requires overlapping fields — the photometric correction uses overlap regions to compute gain/offset between neighbors. Non-overlapping measurements cannot be corrected for photometric drift (vignetting correction still works).
- Extremely sparse datasets (few fields) may produce less accurate photometric corrections due to limited overlap data.
- The correction is calculated from the BSE images as they exist — if the BSE images have been modified by other tools, recalculate the correction.
Reverting Correction
To remove BSE Field Correction:
- Click BSE Adjust → BSE Adjust Revert to undo the last correction
- Click BSE Adjust → Clear ALL BSE Offsets to remove all BSE corrections
These buttons apply to whichever correction mode (BSE Adjustor or BSE Field Correction) is currently active.
See Also
- [BSE Adjustor] — Simple offset-based BSE brightness adjustment
- [MIST Stitching] — Advanced field stitching (complements BSE Field Correction)
- [Image Stability] — Monitoring BSE stability during acquisition
- [Image and EDS Standards] — Auto-adjusting BSE brightness/contrast during measurement