GynSurg: A Comprehensive Gynecology Laparoscopic Surgery Dataset

2025-06-12

AI Summary

Existing gynecological laparoscopic surgery datasets suffer from limited scale, task singularity, and coarse-grained annotations, hindering end-to-end surgical workflow modeling and interpretable analysis. To address these limitations, we introduce Gyn-Lap—the first large-scale, multi-task, fine-grained multimodal dataset for gynecological laparoscopy—supporting four core tasks: surgical action recognition, pixel-level semantic segmentation, structured intraoperative report generation, and novel procedure discovery. Our methodology features holistic, multi-center, hierarchical joint annotation (frame-level actions + pixel-level segmentation + structured textual reports) and establishes standardized acquisition and inter-annotator consistency verification protocols. The publicly released dataset comprises thousands of high-quality surgical videos with comprehensive annotations. Benchmark experiments demonstrate substantial improvements in state-of-the-art model generalization: +5.2% accuracy in action recognition and +4.8% mIoU in semantic segmentation.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Security and Privacy: Data transparency and provenance
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Recent advances in deep learning have transformed computer-assisted intervention and surgical video analysis, driving improvements not only in surgical training, intraoperative decision support, and patient outcomes, but also in postoperative documentation and surgical discovery. Central to these developments is the availability of large, high-quality annotated datasets. In gynecologic laparoscopy, surgical scene understanding and action recognition are fundamental for building intelligent systems that assist surgeons during operations and provide deeper analysis after surgery. However, existing datasets are often limited by small scale, narrow task focus, or insufficiently detailed annotations, limiting their utility for comprehensive, end-to-end workflow analysis. To address these limitations, we introduce GynSurg, the largest and most diverse multi-task dataset for gynecologic laparoscopic surgery to date. GynSurg provides rich annotations across multiple tasks, supporting applications in action recognition, semantic segmentation, surgical documentation, and discovery of novel procedural insights. We demonstrate the dataset quality and versatility by benchmarking state-of-the-art models under a standardized training protocol. To accelerate progress in the field, we publicly release the GynSurg dataset and its annotations

Problem

• Lack of large annotated datasets for gynecologic laparoscopy
• Limited utility of existing datasets for comprehensive workflow analysis
• Need for diverse multi-task surgical data for AI applications

Innovation

• Largest diverse gynecologic laparoscopic surgery dataset
• Rich multi-task annotations for surgical analysis
• Public release to accelerate research progress