Quick start: Upload → Adjust → Detect → Correct → Export
1. Load a field
Select an immunohistochemistry (IHC) stained histopathology image (JPG/PNG). Works for any nuclear IHC marker
such as ER, PR, Ki-67, p53. On a phone, Camera photographs the field directly through the eyepiece.
Tip for phone camera photos: keep the slide field flat and evenly lit, and avoid glare. After
loading, run Auto-enhance to optimise the image before detection.
2. Automated detection
Detect nuclei auto-classifies strong positive (brown/black DAB), weak/blush positive (pale brown) and negative
(blue hematoxylin) nuclei. Weak positives appear as orange dots for observer review.
3. Manual correction
Use Strong +, Weak / blush and Negative to add missed nuclei. Detected nuclei are filled dots, manual marks are
hollow rings. The IHC index reports both strong-only and all-positive percentages.
4. Erase
Select Erase, set the brush size, then press and drag to remove incorrect detected or manual marks.
5. ROI (freehand pen trace)
Choose Trace ROI and draw around the region of interest with the mouse or a finger. Release to close the
polygon. Detect inside ROI restricts detection to the traced region. Esc cancels mid-trace.
6. Image adjustments for eyepiece photographs
- Auto-enhance — one-click optimisation of brightness, contrast, gamma, white balance, background
whitening, saturation and sharpening from the image histogram. Recommended first step for all phone camera
images.
- Manual — opens the sliders for fine tuning: brightness for overall light/dark correction, contrast to
make nuclei stand out, exposure (gamma) to lift dark areas, saturation to separate brown DAB from blue
hematoxylin, sharpen for crisper nuclear boundaries, whiten background to clean up uneven phone lighting, and
temperature to correct warm or cool white-balance casts.
- Revert — returns to the original loaded image.
- Apply — locks adjustments so detection analyses the processed image.
Critical: image adjustments change pixel values. If detection has
already run, existing dots were computed on the original image and may no longer match. The safe workflow is
adjust → apply → detect → correct → export. Never adjust between detect and export if count accuracy matters.
7. Sensitivity settings
- Minimum nucleus size — smallest area in pixels a blob must have to count as a nucleus. Increase if
debris is picked up; decrease if small nuclei are missed.
- DAB sensitivity — how strictly brown/black positive staining is filtered. Increase to catch lighter
brown nuclei; decrease if background tissue is falsely counted. Phone photographs often need a lower value
(~25–35) because phone cameras boost saturation.
- Hematoxylin sensitivity — how strictly blue negative nuclei are filtered. Increase to catch pale blue
nuclei; decrease if collagen or mucin is counted. For phone images, try ~30–40.
8. Weak / blush staining
- Orange dots are weak/blush DAB staining detected but classified as low intensity. They are not
included in the main IHC index by default.
- IHC index (strong) counts only strong brown/black nuclei as positive — strict scoring, e.g. Ki-67.
- All Pos % counts strong and weak together — for ER/PR, where any nuclear staining is positive per
ASCO/CAP guidelines.
- Reclassify manually — tap an orange dot in Strong + mode to upgrade it, or erase it if it is artefact.
The choice is recorded in the report.
- Weak staining is hardest to detect. If many nuclei carry no dot at all, lower DAB sensitivity to ~20. If too
many blue nuclei turn orange, raise it to ~50.
9. Export
Save an annotated PNG, a CSV of coordinates, a session JSON, or print the full report. The report includes both
strong-only and all-positive IHC indices.
Input and shortcuts
Touch: one finger marks or erases, two fingers pinch to zoom and drag to pan. Pan / inspect mode lets one
finger pan without marking.
Mouse: wheel zooms at the cursor, middle-button or Pan mode drags the field.
Keys: 1 strong positive, 2 negative, 3 erase, 4 weak/blush, P pan, R ROI, F fit, Ctrl+Z undo,
Ctrl+Y redo, Esc cancel.
Adjustment risk
| Adjustment | Risk | Recommendation |
|---|
| Whiten background | High | Apply before detection only |
| Brightness | High | Apply before detection only |
| Contrast | High | Apply before detection only |
| Gamma | High | Apply before detection only |
| Sharpen | Medium | Safe if subtle; re-run if aggressive |
| Saturation | Low | Generally safe after detection |
| Temperature | Low | Safe after detection |
Troubleshooting
- Too many false positives — raise the minimum nucleus size, lower DAB sensitivity, whiten the
background before detection.
- Too many missed nuclei — lower the minimum nucleus size, raise DAB or hematoxylin sensitivity, raise
brightness or gamma before detection.
- Clusters counted as one nucleus — sharpen before detection, then split the cluster by erasing and
marking manually.
- Counts changed after an image adjustment — expected. Follow adjust → apply → detect → correct →
export.
- Very large photograph — images above roughly 12 megapixels are scaled down to stay within the
device's canvas limit; the notice states the working size.
Golden rule: adjust → apply → detect → correct → export
