| Operation | Inputs | Description |
|---|---|---|
| Fuse events | Coded image A, Coded image B | Merges all events from both coded images into a single output. Labels are renumbered for uniqueness. Priority rules resolve spatial overlaps. |
| Fuse events (with mask) | Coded image A, Coded image B, Binary mask | Fuses events conditionally: from input A, only events within the mask are included; from input B, only events outside the mask (or vice versa, depending on configuration). Enables spatial mixing of detection strategies. |
BOM Group
Fuse Events
Merging events from multiple coded images
View
Definition
Operations that merge events from multiple coded images into a single unified coded image — combining detection results from different methods, layers, or processing passes, with optional mask-based spatial constraints for selective fusion.
Two Variants
Basic fusion and mask-constrained fusion
Label Renumbering
Output labels are unique and contiguous
Overlap Resolution
Priority rules handle spatial conflicts
Multi-Strategy Detection
Combine the best results from different approaches
Operations Reference
Simplified
Fuse events combines two sets of detected objects into one. Fuse events (with mask) lets you control which regions each input contributes to — using a mask to divide the image into zones handled by different detection methods.
When to Use Which
Fuse Events (Basic)
Use when combining detection results that occupy different spatial areas with minimal overlap — e.g., nuclei detected in Layer 1 and total tissue area detected in Layer 2. Since the objects don't overlap spatially, simple fusion merges them cleanly.
Fuse Events (With Mask)
Use when different regions of the same image require different detection strategies. Common scenarios:
- Tumor vs. stroma: Use deep learning nuclei detection in morphologically complex tumor regions, classical thresholding in clean stromal areas. The Classifier-generated tumor mask determines which method's results to keep in each region.
- Dense vs. sparse tissue: Use aggressive watershed separation in dense cellular areas, gentle detection in sparse regions. A density-based mask controls the blend.
- Multi-layer combination: Different layers may detect different cell populations. Fuse with mask combines them into a single coded image for unified measurement.
Simplified
Use basic Fuse for non-overlapping detection results. Use Fuse with mask when different regions need different detection methods — the mask tells the operation where to take results from each input.