{
  "status": "PASS_WITH_DOCUMENTED_DISPLAY_NOTES",
  "method": "view_image of renderer-generated technical samples; no video/action correctness assessment",
  "samples": [
    "01_camera24_fight/data/B1_Directed_sample_0000.jpg",
    "02_camera22_crowd/data/B2_SeedFollow_sample_0140.jpg",
    "SELF_C09/data/B2_SeedFollow_sample_0157.jpg"
  ],
  "checked": ["readable Korean text", "headers and score/detail boxes not clipped", "scene and detail regions separated", "identity labels remain legible", "current-condition ROI border and source/reference frame fields visible"],
  "notes": [
    "bbox #1-3 marks are inherited raw first-three-object marks, not pair detail indices; README clarifies this",
    "B2 header distance/approach describe seed conditions; actual follow-up has different documented distance/direction conditions",
    "All-frame automated text boundaries/overlap checks and full decoding are separate from these three samples"
  ],
  "semantic_fight_judgment": "NOT_PERFORMED"
}