An AI Framework for Microanastomosis Motion Assessment

TL;DR

The paper addresses the need for objective, scalable assessment of microanastomosis proficiency, replacing subjective expert judgments with an end-to-end AI framework. The approach integrates four modules—instrument detection with YOLOv11, instrument tracking with DeepSORT, tip localization via shape descriptors, and a supervised skill classifier—enhanced by a hybrid detection–tracking strategy to improve temporal consistency. Empirical results show instrument detection precision near $97\%$ and $mAP$ values on the order of $98\%$ (IoU ~ $50\%$ to $95\%$), with near real-time operation (~$29.7$ fps), and an overall skill-classification accuracy of around $87\%$ across three tiers. This framework offers a scalable, objective tool for neurosurgical training, enabling standardized evaluation and data-driven curriculum development; future work will broaden the dataset and incorporate more expert raters to improve ground-truth reliability.

Abstract

Proficiency in microanastomosis is a fundamental competency across multiple microsurgical disciplines. These procedures demand exceptional precision and refined technical skills, making effective, standardized assessment methods essential. Traditionally, the evaluation of microsurgical techniques has relied heavily on the subjective judgment of expert raters. They are inherently constrained by limitations such as inter-rater variability, lack of standardized evaluation criteria, susceptibility to cognitive bias, and the time-intensive nature of manual review. These shortcomings underscore the urgent need for an objective, reliable, and automated system capable of assessing microsurgical performance with consistency and scalability. To bridge this gap, we propose a novel AI framework for the automated assessment of microanastomosis instrument handling skills. The system integrates four core components: (1) an instrument detection module based on the You Only Look Once (YOLO) architecture; (2) an instrument tracking module developed from Deep Simple Online and Realtime Tracking (DeepSORT); (3) an instrument tip localization module employing shape descriptors; and (4) a supervised classification module trained on expert-labeled data to evaluate instrument handling proficiency. Experimental results demonstrate the effectiveness of the framework, achieving an instrument detection precision of 97%, with a mean Average Precision (mAP) of 96%, measured by Intersection over Union (IoU) thresholds ranging from 50% to 95% (mAP50-95).

Figures (6)

• Figure 1: An overview of the AI microanastomosis motion assessment system architecture.
• Figure 2: Data collection and video recording environment setup.
• Figure 3: Class distribution in the object detection dataset.
• Figure 4: YOLOv11 training results.
• Figure 5: YOLOv11 testing images