New dataset advances novel view synthesis for gastroendoscopy

Researchers have introduced the GastroNVS dataset, the first real-world collection of gastroscopic images and associated data specifically designed for novel view synthesis (NVS) in medical endoscopy. This dataset aims to advance NVS techniques, such as those based on Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS), for applications like expanding endoscopic field of view and creating digital twins for training and archiving. The paper also evaluates existing 3DGS methods on this new dataset, highlighting current challenges and future research directions. AI

IMPACT This dataset could enable new AI-powered visualization and training tools for medical procedures.

RANK_REASON The item is an academic paper introducing a new dataset and evaluating methods. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• 3D Gaussian splatting
• arXiv
• Gastroendoscopy View Synthesis
• GastroNVS
• Nerf
• Neural radiance field

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →