StrataQuest Home StrataQuest Glossary Removal operations
Image Processing & Correction

Removal operations

Six engines that select from a detection result by criterion — keeping the events the analysis actually wants

Definition
Removal operations take a coded imageLoading... (the labeled output of detection) and produce a refined coded image with some labels removed. The criteria are: Remove labels (larger / smaller) — drop events above or below a size threshold, measured by area or by another size proxy. Remove labels (stronger / weaker) — drop events above or below an intensity threshold (mean, integrated, or another intensity statistic). Remove objects (with seeds) and Remove objects (without seeds) — drop events based on their relationship to a separate seed mask, retaining either those that contain seeds or those that don't. All six produce a coded image of the same shape as the input, with the targeted labels gone — and the surviving labels usually renumbered to fill the resulting gaps.
Supporting
Detection - Nuclei Segmentation Classic
Video · Supporting
Detection - Automatic Background Threshold params
Video · Supporting
Pre processing - Import - Coded Image
Video · Supporting
Filter by what you can measure
Size and intensity are the standard criteria
Seeded removal connects detections to other detections
Keep events that contain seeds, or that don't
Where they sit in the workflow
Post-Processing slot, after measurement
Compose multiple criteria with chains
Cascade removals to express AND

Choosing the criterion, by what should survive

The choice of removal engine is determined by the property that distinguishes events the analysis cares about from events it doesn't:

Surviving events should be at least a certain size: Remove labels (smaller).

Surviving events should be no larger than a certain size: Remove labels (larger).

Surviving events should be at least a certain brightness: Remove labels (weaker).

Surviving events should be no brighter than a certain level: Remove labels (stronger).

Surviving events should overlap a known feature mask: Remove objects (without seeds), with the feature mask as the seeds. ("Drop events that don't contain a seed" = "keep events that do.")

Surviving events should NOT overlap a known feature mask: Remove objects (with seeds).

Compound criteria become chains. "Cells larger than 50 µm² and brighter than threshold X and in tumor regions" is three engines in sequence — each criterion expressed as one removal step.

Share This Term
Term Connections