Learning from Single Timestamps: Complexity Estimation in Laparoscopic Cholecystectomy

2025-11-06

AI Summary

This study addresses the challenge of automatic surgical complexity assessment in unedited laparoscopic cholecystectomy (LC) videos. To overcome the lack of frame-level annotations, the authors propose a single-timestamp weakly supervised learning method grounded in the Parkland Grading Scale (PGS) and introduce STC-Net—a unified framework jointly modeling temporal localization and inflammation severity classification. STC-Net incorporates a hard-soft hybrid localization loss and a background-aware classification mechanism to enable end-to-end training. Notably, it is the first method to perform both complexity grading and temporal localization across full-length videos using only a single timestamp label per video. Evaluated on 1,859 real-world LC videos, the model achieves 62.11% accuracy and 61.42% F1-score—improving over non-localizing baselines by over 10%. These results demonstrate significant gains in postoperative evaluation and surgical training efficiency, validating the clinical feasibility and scalability of weakly supervised surgical video analysis.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Video Understanding & Activity Analysis
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Security and Privacy: Large-scale security measurements
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Purpose: Accurate assessment of surgical complexity is essential in Laparoscopic Cholecystectomy (LC), where severe inflammation is associated with longer operative times and increased risk of postoperative complications. The Parkland Grading Scale (PGS) provides a clinically validated framework for stratifying inflammation severity; however, its automation in surgical videos remains largely unexplored, particularly in realistic scenarios where complete videos must be analyzed without prior manual curation. Methods: In this work, we introduce STC-Net, a novel framework for SingleTimestamp-based Complexity estimation in LC via the PGS, designed to operate under weak temporal supervision. Unlike prior methods limited to static images or manually trimmed clips, STC-Net operates directly on full videos. It jointly performs temporal localization and grading through a localization, window proposal, and grading module. We introduce a novel loss formulation combining hard and soft localization objectives and background-aware grading supervision. Results: Evaluated on a private dataset of 1,859 LC videos, STC-Net achieves an accuracy of 62.11% and an F1-score of 61.42%, outperforming non-localized baselines by over 10% in both metrics and highlighting the effectiveness of weak supervision for surgical complexity assessment. Conclusion: STC-Net demonstrates a scalable and effective approach for automated PGS-based surgical complexity estimation from full LC videos, making it promising for post-operative analysis and surgical training.

Problem

• Automating surgical complexity assessment in laparoscopic cholecystectomy videos
• Estimating Parkland Grading Scale from untrimmed surgical videos
• Developing weakly supervised temporal localization for complexity grading

Innovation

• STC-Net framework for single-timestamp complexity estimation
• Joint temporal localization and grading in full videos
• Hard-soft loss formulation with background-aware supervision