Research map / Video / Video-language understanding
SurgPub-Video: A Comprehensive Surgical Video Framework for Enhanced Surgical Intelligence in Vision-Language Model
Yaoqian Li, Xikai Yang, Dunyuan Xu, Yang YU, Litao Zhao, Xiaowei Hu, Jinpeng Li, Pheng-Ann Heng
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 changed | Where | Measured | Effect |
|---|---|---|---|
| train on SurgPub-Video video VQA data (SurgLLaVA-Video 3B) instead of general VLMs | TinyLLaVA-Video · SurgPub-Video benchmark | overall accuracy (%): 66.67 (GPT-4o) -> 82.84 | Improved |
| SurgLLaVA-Video 3B vs surgical SurgVLM 72B on action recognition | SurgLLaVA-Video · SAR-RARP | accuracy (%): 42.90 -> 65.65 | Improved |
| SurgLLaVA-Video vs SurgVLM on CVS assessment | SurgLLaVA-Video · Endoscapes CVS | balanced accuracy (%): 51.40 -> 58.88 | Improved |
| SurgLLaVA-Video vs SurgVLM on CholecT50 triplet | SurgLLaVA-Video · CholecT50 | accuracy/mAP (%): triplet acc 12.52 -> 39.30; target acc 57.07 -> 47.12; target mAP 21.01 -> 19.93 | Mixed |
| increase input frames from 8 to 16 (ablation) | SurgLLaVA-Video · SurgPub-Video benchmark | overall accuracy (%): peak 80.82 at 16 frames | Improved |
| increase input frames beyond 16 to 32 (ablation) | SurgLLaVA-Video · SurgPub-Video benchmark | overall accuracy (%) | Hurt |
| increase training data from 25% to 100% (ablation) | SurgLLaVA-Video · SurgPub-Video benchmark | overall accuracy (%): up to 80.82 at 100% | Improved |
| increase training data beyond 75% for surgery planning (ablation) | SurgLLaVA-Video · SurgPub-Video benchmark planning | accuracy (%): peak 82.40 at 75% | No clear effect |
Closest papers by meaning
- Procedure-Aware Surgical Video-language Pretraining with Hierarchical Knowledge Augmentation (NeurIPS 2024 · similarity 0.90)
- SurgCoT: Advancing Spatiotemporal Reasoning in Surgical Videos through a Chain-of-Thought Benchmark (CVPR 2026 · similarity 0.89)
- Bridging Vision and Language for Robust Context-Aware Surgical Point Tracking: The VL-SurgPT Dataset and Benchmark (AAAI 2026 · similarity 0.88)
- OmniMedVQA: A New Large-Scale Comprehensive Evaluation Benchmark for Medical LVLM (CVPR 2024 · similarity 0.87)
- MedHorizon: Towards Long-context Medical Video Understanding in the Wild (NeurIPS 2026 · similarity 0.86)
- MM-OR: A Large Multimodal Operating Room Dataset for Semantic Understanding of High-Intensity Surgical Environments (CVPR 2025 · similarity 0.85)
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.
