Segment microscopy images into reviewable cell measurements

Why use this

1. A microscopy researcher may need cell measurements from images whose channels, pixel sizes and dimensionality vary. A segmentation overlay can appear convincing while merging neighboring cells or missing dim structures.

2. This job could prepare a reproducible segmentation plan and the review artifacts needed to inspect later masks. Keeping acquisition metadata and expert quality checks visible helps prevent measurements from being treated as reliable merely because software produced a table.

Why an agent / simpler alternative

Useful judgment

Reconcile image metadata, identify ambiguous channel/mode choices and organize mask-review criteria for the intended research question.

Simpler alternative

Use a validated lab segmentation workflow when acquisition conditions and expected morphology already match it.

Unsuitable when

The task would support diagnosis, process protected images without approval or report biological conclusions without expert validation.

Proposed inputs

1. Name

   Image manifest

   Owner

   Authorized researcher

   Minimum scope

   File formats, channels, dimensions and scale metadata

   Quality check

   Check pixel-size provenance and sample identifiers

2. Name

   Segmentation brief

   Owner

   Domain lead

   Minimum scope

   Desired objects, approved model/mode and review criteria

   Quality check

   Specify CPU/GPU and volumetric choices explicitly

Access and setup — proposed or unknown

Proposed minimum

1. Review synthetic image manifests and proposed overlay checks; no image processing, model downloads or GPU use.

Compatibility unknowns

1. Current model support, image readers, hardware resources and runtime compatibility are unknown.

External actions requiring approval

1. Sensitive-image access, uploads, model installation, compute spending and output overwrites requires authorization.

Expected deliverable and review checks

Proposed artifact

Segmentation-plan candidate with metadata exceptions, parameter record and mask/measurement review checklist.

Critical checks

1. Require correct scale and channel mapping; inspect merges/splits and missed objects; separate planned review from measured performance.

Missing output evidence

This overlay adds no new output inspection, execution or validation; consult the referenced canonical source outcome limits.

Failure signals and proposed responses

1. Signal

   Pixel size is missing or inconsistent

   Why it matters

   Physical measurements lack a reliable basis

   Proposed response

   Hold scaled outputs until acquisition metadata is verified.

2. Signal

   A channel is selected by position rather than identity

   Why it matters

   The model may process the wrong signal

   Proposed response

   Confirm channel meaning before any run.

Effort and cost — unknown

Setup work

1. Define acquisition context and scientific acceptance criteria.

Ongoing review

1. Researchers inspect actual overlays and measurements after any authorized analysis.

Unknowns

1. Segmentation quality, model suitability, memory demand and compute costs are unknown.

First check — proposed, unexecuted

Status

proposed_unexecuted

Synthetic fixture

Invent two image manifests with a swapped channel order and absent pixel size; prepare a preflight table only.

Acceptance checks

1. Flag both issues; segment nothing; report no cell measurements or validated accuracy.

Stop conditions

1. Stop if scale, channel identity or protected-data authority is unresolved.

Limits

1. A manifest review cannot establish segmentation accuracy, meaningful morphology measurements or biological interpretation.