preprocess-imaging

Preprocess-Imaging Skill

Purpose

This skill designs and audits the data-preparation stage of a medical-imaging model — the stage before a training repo is built — and proves it is leakage-safe by construction. Data leakage enters one step earlier than the split table can see: a normaliser fit on the whole dataset, a data-fitted transform run before the split exists, or a patient whose slices land in more than one partition. Each silently inflates every downstream metric (Kapoor & Narayanan, Patterns 2023; Varoquaux & Cheplygina, npj Digit Med 2022; CLAIM 2024 data items).

It is the missing first link in the lane: preprocess-imaging (prepare + audit) → /model-scaffold (build) → /model-validation (validate the split) → /model-evaluation + /analyze-stats (metrics) → /write-paper + /check-reporting (publish). It integrates MONAI / TorchIO transforms (referenced in the emitted plan); it does not reimplement them, and it never executes preprocessing on real patient data.

When to use

• You have a data manifest (one row per image/slice with a patient/subject ID) and want a leakage-safe preprocessing plan + a machine-checkable manifest before scaffolding a model.
• You want to audit an existing preprocessing pipeline for data-stage leakage.

When NOT to use

• Auditing the train/val/test split table itself → /model-validation (split-leakage gate).
• Building the training repo / model code → /model-scaffold (it consumes this manifest).
• Choosing the architecture → /architecture-zoo.
• Held-out metrics / calibration → /model-evaluation then /analyze-stats.
• Reimplementing MONAI / TorchIO transforms → out of scope (this skill wires and audits them).

Workflow

Phase 1 — Inventory the data and the intended steps

Collect: modality (CT / MR / X-ray / US / path), the data manifest (one row per image/slice with a patient_id), the intended resample spacing, the intensity transform (fixed HU window vs a fitted z-score / min-max / histogram match), and the augmentation plan. See references/preprocessing_guide.md for modality-aware guidance (what normalisation is standard per modality, which augmentations preserve vs break physiology).