Find & count nuclei
Boundary U-Net segmentation followed by deterministic instance separation.
Measure morphology
Quantify nuclear geometry and inspect population distributions.
Measure fluorescence
Summarize fluorescence within each detected nuclear instance.
Interactive object review
Inspect the returned overlay with quantitative nuclei rendered as measurement-aware points.
Spatial ROI review
Draw a temporary rectangle over the analysis view and summarize the nuclei inside it.
Evidence report
Package visual evidence, machine-readable measurements, provenance and specialist review.
Colocalization
Multi-channel overlap statistics require a validated acquisition model and channel-aware validation before public release.
Time-series tracking
Tracking requires temporal metrics, identity handling and explicit failure analysis before release.
Learned object classification
A public classifier requires a released model and independent validation set. User images are not silently used to train it.
Review workflow
See what the algorithm actually measured.
The Lab now separates prediction, object measurements and interpretation. A scientist can inspect the overlay, view the per-nucleus table, explore measurement-aware spatial patterns and then download the complete evidence bundle.
1 · Predict
Boundary U-Net produces semantic nuclear information.
2 · Separate
Deterministic processing converts predictions into individual instances.
3 · Measure
Geometry and intensity are computed from the returned instances.
4 · Review
Viewer, table, ROI and evidence-constrained report expose the result.
Run the complete workflow.
Upload a supported TIFF or Nikon ND2 in the BioNuclei Analyzer, inspect the returned evidence, download the report and delete the transient server-side job.
Open BioNuclei Analyzer