SG2VID conditions a latent video diffusion model on scene graphs with temporal features to generate controllable surgical videos across cataract and cholecystectomy datasets.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SG2VID: Scene Graphs Enable Fine-Grained Control for Video Synthesis
SG2VID conditions a latent video diffusion model on scene graphs with temporal features to generate controllable surgical videos across cataract and cholecystectomy datasets.