Session profile
- Audience: learners working with EM data acquisition or QA.
- Duration: 75 minutes lecture + 15 minutes QA exercise.
- Output: acquisition risk register and QA gate plan.
Slide-by-slide lecture plan
- Slide 1 (2 min): Title and positioning
- Slide 2 (4 min): Acquisition quality as reconstruction ceiling
- Slide 3 (6 min): End-to-end prep chain
- Fixation, staining, sectioning/block-face, imaging.
- Slide 4 (6 min): Imaging parameter tradeoffs
- Dwell time, beam current, overlap, and throughput.
- Slide 5 (6 min): Artifact taxonomy I
- Physical artifacts: fold, tear, chatter, compression.
- Slide 6 (6 min): Artifact taxonomy II
- Contrast/charging/drift and alignment implications.
- Slide 7 (6 min): Stitching and stack assembly QA
- Slide 8 (6 min): Pilot segmentation as acquisition validation
- Slide 9 (7 min): Worked artifact triage example
- Slide 10 (5 min): Metadata standards and provenance requirements
- Slide 11 (5 min): Failure escalation protocol
- Slide 12 (5 min): Throughput vs fidelity governance decisions
- Slide 13 (5 min): Mini-lab
- Build a risk register for one tissue/imaging configuration.
- Slide 14 (6 min): Bridge to infrastructure and large-scale processing.
- Primary shortlist:
course/units/figures/03-em-prep-and-imaging-selected-v1.md.
- Include at least one before/after artifact correction panel.
Speaker notes (expert-level)
- Tie every artifact class to a concrete downstream segmentation failure type.
- Emphasize regional QA rather than global aggregate quality claims.
Assessment and artifacts
- Deliverable: acquisition QA checklist with quantitative triggers.
- Rubric dimensions: artifact coverage, detection metrics, mitigation realism.
Connections
Slide source file
- Marp draft source:
course/decks/marp/03-em-prep-and-imaging.marp.md
- Batch render helper:
./scripts/render_marp.sh
