This study introduces Surgical-DINO, a model for depth estimation in endoscopic surgery. This model is a low-rank adaptation of the DINOv2 for depth estimation. The authors integrated LoRA layers into DINO to adapt it with surgery-specific domain knowledge. The model was validated on a MICCAI challenge dataset of SCARED, which was collected from da Vinci Xi endoscope surgery. The results show that Surgical-DINO significantly outperforms all state-of-the-art models in endoscopic depth estimation tasks.
Publication date: 11 Jan 2024
Project Page: https://arxiv.org/abs/2401.06013
Paper: https://arxiv.org/pdf/2401.06013
