Exploring Pre-training Across Domains for Few-Shot Surgical Skill Assessment

2025-09-11

AI Summary

Automated surgical skill assessment (SSA) is hindered by the scarcity and high cost of expert annotations. To address the data bottleneck in few-shot learning (FSL) settings, this paper proposes a cross-domain self-supervised pretraining framework. It systematically investigates the efficacy of jointly pretraining on small-scale domain-related data and surgery-specific data—marking the first such exploration—and elucidates how domain discrepancy impacts transfer performance. Evaluated on an OSATS-annotated robotic surgery dataset under 1-, 2-, and 5-shot settings, our method achieves accuracies of 60.16%, 66.03%, and 73.65%, respectively. Incorporating domain-related data yields average improvements of +1.22% in accuracy and +2.28% in F1-score. The approach substantially enhances model generalization under extremely limited annotation budgets, establishing a novel, scalable, and cost-effective paradigm for SSA.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Search and Optimization: Learning to Search
• Application Domains: Humanities & Computational Social Science

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Automated surgical skill assessment (SSA) is a central task in surgical computer vision. Developing robust SSA models is challenging due to the scarcity of skill annotations, which are time-consuming to produce and require expert consensus. Few-shot learning (FSL) offers a scalable alternative enabling model development with minimal supervision, though its success critically depends on effective pre-training. While widely studied for several surgical downstream tasks, pre-training has remained largely unexplored in SSA. In this work, we formulate SSA as a few-shot task and investigate how self-supervised pre-training strategies affect downstream few-shot SSA performance. We annotate a publicly available robotic surgery dataset with Objective Structured Assessment of Technical Skill (OSATS) scores, and evaluate various pre-training sources across three few-shot settings. We quantify domain similarity and analyze how domain gap and the inclusion of procedure-specific data into pre-training influence transferability. Our results show that small but domain-relevant datasets can outperform large scale, less aligned ones, achieving accuracies of 60.16%, 66.03%, and 73.65% in the 1-, 2-, and 5-shot settings, respectively. Moreover, incorporating procedure-specific data into pre-training with a domain-relevant external dataset significantly boosts downstream performance, with an average gain of +1.22% in accuracy and +2.28% in F1-score; however, applying the same strategy with less similar but large-scale sources can instead lead to performance degradation. Code and models are available at https://github.com/anastadimi/ssa-fsl.

Problem

• Investigating pre-training impact on few-shot surgical skill assessment
• Addressing scarcity of expert-annotated surgical skill data
• Evaluating domain relevance versus dataset size for transfer learning

Innovation

• Self-supervised pre-training for surgical skill
• Domain-relevant datasets outperform large-scale ones
• Procedure-specific data boosts few-shot performance