AI Paper Atlas

Research map / Video / Video generation

Mitigating Surgical Data Imbalance with Dual-Prediction Video Diffusion Model

Danush Kumar Venkatesh, Adam Schmidt, Muhammad Abdullah Jamal, Omid Mohareri

ICML 2026 · Video › Video generation · approach: Efficient video diffusion models · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add SurgiFlowVid synthetic videos (Seg conditioning) for under-represented classesMViTv2 action recognition · SAR-RARP50mean Jaccard: 0.46 -> 0.52Improved
add dual-prediction module (RGB plus optical flow) to SurV-Gen (ablation)SurgiFlowVid · SAR-RARP50 action recognitionmean Jaccard: 0.46 -> 0.49Improved
add sparse visual encoder with RGB or mask conditioning (ablation)SurgiFlowVid · SAR-RARP50 action recognitionmean Jaccard: 0.49 -> 0.52 (Seg.)Improved
model each under-represented class individually with RGB conditioning (ablation)SurgiFlowVid · SAR-RARP50 action recognitionmean Jaccard: 0.46 -> 0.53; A7 0.32 -> 0.52Improved
model each class individually with segmentation conditioning (ablation)SurgiFlowVid · SAR-RARP50 action recognitionmean Jaccard: 0.52 -> 0.48Mixed
duplicate the training set rather than add synthetic samples (ablation)downstream action recognizer · SAR-RARP50mean Jaccard: 0.46 -> 0.45No clear effect
shuffle frames of synthetic videos (ablation)downstream action recognizer · SAR-RARP50mean Jaccard: 0.46 -> 0.43Hurt
add noise or use only sparse frames as synthetic data (ablation)downstream action recognizer · SAR-RARP50mean Jaccard: 0.46 -> 0.42 (sparse), 0.43 (noisy)Hurt
add SurgiFlowVid RGB-conditioned samples to real GraSP dataMViTv2 step recognition · GraSPmean mAP: 0.47 -> 0.52Improved
add Endora or SparseCtrl (RGB) synthetic samplesMViTv2 step recognition · GraSPmean mAP: 0.47 -> 0.45 (Endora), 0.46 (SparseCtrl)Hurt
add SurgiFlowVid (Seg) synthetic samples for tool presence detectiontool detector · SAR-RARP50mean Dice: 0.69 -> 0.79Improved
add SurgFlowVid (Seg) synthetic samples for tool detectiontool detector · GraSPmean Dice: 0.68 -> 0.72Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy