Work

2025Vision

AIMS: Surgical Phase Detection

Research intern, AIMS (AI in Medicine and Surgery)

From March to November 2025 I was a computer-vision intern with AIMS (AI in Medicine and Surgery) at Leeds. I trained DINOv2 models for endoscopic submucosal dissection workflow, later published as the ISBI 2026 paper “Self-Supervised Vision Transformer for Surgical Phase Recognition in Endoscopic Submucosal Dissection” (DOI 10.1109/isbi61048.2026.11515812). That paper reports 89.5% accuracy on the patient set and 90.0% on porcine.

Endoscopic frames beside DINOv2 patch-norm and centre-similarity maps through an ESD sequence.
Fig. 1DINOv2 feature diagnostics across an endoscopic submucosal dissection sequence. Selected frames from one ESD sequence.
Stack
DINOv2, V-JEPA, Medical AI, Self-Supervised Learning
Links
Paper, Code

Also