Pipeline stage

Quality Control

Systematically flag images, fields, and runs that do not meet quality thresholds — out-of-focus fields, illumination drift, registration failure, phototoxicity — before they corrupt segmentation, features, or biological readout.

The problem — A QC failure that reaches segmentation produces masks for cells that were never in focus. One that reaches profiling inflates or deflates the morphological signal of an entire plate edge. One that reaches a slide-level classifier embeds blurry tiles at full weight. The failure is invisible from the output alone — the result looks well-formed because the pipeline completed without error. QC exists to catch these failures before they propagate.

What it is / how it works — Fovea treats quality as four nested levels, as described in Image Quality Is Not One Number: image-space physics (focus, SNR, illumination uniformity), run/sample consistency (field-to-field drift, batch structure, registration stability), readout fidelity (segmentation accuracy, percent-replicating), and the operational decision (reportable / needs-review / reject). The QC stage primarily instruments the first two levels — per-FOV focus scoring, illumination drift detection, saturation flags, drift-over-time curves, bleaching profiles — because these are computable directly from images without ground truth. Metrics Reloaded (Maier-Hein et al.) grounds the readout-level checks: metric selection must reflect domain interest, not imaging convenience. The QC stage outputs a structured QC report that feeds the final QC-aware report.

Where it breaks — The general failure is instrumenting only the image-space level and reporting as if the readout level were clear. Sharp fields can sit inside a batch-confounded run; a per-FOV focus score cannot detect a well that saturated at 10 minutes. QC must also be modality-aware: for live-cell imaging, photobleaching and drift accumulate over time; for spatial omics, cycle registration error is the dominant failure; for digital pathology, tissue folds and pen marks are invisible to a sharpness metric. A QC stage that is not parameterized by modality will pass failures specific to that modality.

A green image-space QC is not a green readout. A field can be sharp, correctly illuminated, and still produce a morphological profile that is non-replicating for reasons invisible to per-image metrics.

References

Appears in these notes

  • The QC-Aware ReportA QC-aware report doesn't just present a result — it surfaces the evidence that the result is trustworthy, structured by the four nested levels of quality and ending in an explicit decision. Its job is to make a number un-trustable on sight when it shouldn't be trusted.
  • Choosing a Bioimage Analysis Pipeline — Off-the-Shelf vs CustomMost microscopy questions are answered by composing existing, validated tools — and a custom pipeline is justified only when off-the-shelf options fail on a specific, identifiable axis. This note gives the decision criteria and the signals that you have genuinely crossed into custom territory.
  • Foundation Models Are Not PipelinesA foundation model is a powerful feature extractor, not a deployable measurement pipeline. It still needs ingestion, QC, preprocessing, confounder correction, validation, and provenance around it — and a frozen encoder makes the surrounding system more important, not less.
  • Image Quality Is Not One NumberFovea treats microscopy quality as four nested levels — image-space, run/sample, readout, and decision — because a metric at one level says nothing about the levels above it. Full-reference scalars like SSIM and PSNR measure only the first, yet teams report as if they measured the third.
  • Verification — A Pipeline You Can Re-RunVerification is the engineering discipline of making a microscopy pipeline deterministic, version-pinned, and re-executable from a provenance record. Without it, every result is a one-off, and "we changed nothing" is unprovable.
  • The Challenge of Rare Events and MutantsWhen the phenotype you care about appears in 10 cells out of 300,000, accuracy is meaningless, validation has almost no positives to learn from, and false discovery dominates. Rare-event detection is a class-imbalance and ground-truth-scarcity problem — best answered by adaptive acquisition that goes looking for the positives.
  • From Research Script to Production PipelineA notebook that worked once is not a pipeline, and a community model that scored well on its paper's data is not a validated component. Turning either into something deployable is the engineering work of pinning, wrapping, gating, and validating — most of which the original artifact deliberately skipped.
  • Uncertainty, Drift, and Failure Modes in Bioimage AnalysisModels fail silently when the data drifts away from what they were trained on — and an overconfident, miscalibrated model gives no warning. Trustworthy deployment needs calibrated uncertainty, explicit distribution-shift detection, and a way to flag failure when there is no ground truth to check against.
  • Verification and Validation Are Two Different QuestionsVerification asks whether the pipeline was built right — deterministic, reproducible, correct to spec. Validation asks whether it is the right pipeline — outputs that are biologically true and fit for purpose. Microscopy makes both hard, and conflating them is how silently wrong results get shipped.
  • Measure Where It MattersAdaptive, uncertainty-driven acquisition treats the microscope as part of the model — spending photons and time only where the image is uncertain or the biology is happening, instead of scanning everything uniformly.

← Back to the constellation