Deck Spec: 08 Segmentation and Proofreading

Session

Slide production blocks

  1. Title and QC thesis
    • Key message: proofreading is the scientific quality layer for segmentation outputs.
    • Figure: FIG-SRC-MODULE14_LESSON2-S03-01
  2. Error taxonomy
    • Key message: merges, splits, boundary ambiguity, identity confusion.
    • Figure: FIG-SRC-MODULE14_LESSON2-S08-01
  3. Ultrastructure cues for correction
    • Key message: neuronal context cues improve correction precision.
    • Figure: FIG-RIV-ULTRA-S06-01
  4. Synapse-aware correction decisions
    • Key message: synaptic context guards against topology-breaking edits.
    • Figure: FIG-RIV-ULTRA-S09-01
  5. Organelle and boundary ambiguity
    • Key message: organelle cues help disambiguate uncertain boundaries.
    • Figure: FIG-RIV-ULTRA-S11-01
  6. Comparative ambiguity panel
    • Key message: similar local textures can map to different correction actions.
    • Figure: FIG-RIV-ULTRA-S17-01
  7. Boundary failure case
    • Key message: unresolved boundary ambiguity should remain explicitly tagged.
    • Figure: FIG-RIV-ULTRA-S23-01
  8. Identity disambiguation (axon/dendrite)
    • Key message: identity checks reduce downstream motif distortion.
    • Figure: FIG-RIV-AXDEN-S13-01
  9. Edge-case process morphology
    • Key message: difficult morphology requires adjudication, not solo correction.
    • Figure: FIG-RIV-AXDEN-S18-01
  10. High-complexity correction panel
    • Key message: correction priority should track downstream impact.
    • Figure: FIG-RIV-AXDEN-S22-01
  11. Metrics and release gates
    • Key message: pair VI, edge P/R, ERL, and synapse-centric F1 with audit logs.
    • Figure: FIG-SRC-MODULE14_LESSON2-S13-01
  12. Activity and bridge
    • Key message: write one correction log with metric deltas; bridge to motif analysis.
    • Figure: FIG-SRC-MODULE14_LESSON2-S10-01

Assessment