An ophthalmic video foundation model for surgical recognition and navigation with wet-lab porcine eye validation - PubMed
8 hours ago
- #AI in healthcare
- #ophthalmic surgery
- #surgical navigation
- Development of an ophthalmic video foundation model (OVFM) for surgical recognition and navigation.
- OVFM trained on 1.1 million clips across 144 surgical types using a self-supervised video transformer.
- Demonstrated superior performance in seven downstream tasks and real-time deployment via knowledge distillation.
- Enhanced surgical performance and reduced skill gaps in cataract surgeries on wet-lab porcine eyes.
- Potential for broad real-time, intraoperative applications in various surgical fields.