In a 10-procedure pilot, deep learning identified abdominal wall layers in optical trocar-entry video with promising accuracy—but this is not yet evidence of safer access.

• Across 951 annotated frames, mean Dice was 0.883; scores were 0.923 for rectus muscle, 0.903 for rectus sheath and 0.824 for peritoneum.
• Tissue classes were rarely confused; most errors occurred at tissue-background boundaries.

The work supports technical feasibility, not clinical readiness: the small, single-center dataset limits generalizability, and prospective testing is needed before AI guidance can inform entry decisions.