Microsurgical Instrument Segmentation for Robot-Assisted Surgery

Abstract-- Accurate segmentation of thin structures is critical for microsurgical scene understanding but remains challenging due to resolution loss, low contrast, and class imbalance. We propose Microsurgery Instrument Segmentation for Robotic Assistance(MISRA), a segmentation framework that augments RGB input with luminance channels, integrates skip attention to preserve elongated features, and employs an Iterative Feedback Module(IFM) for continuity restoration across multiple passes. Microsurgery(MS) is a surgical technique that manipulates blood vessels as small as 1-2 mm in diameter and plays a critical role in lymphedema treatment and soft tissue reconstruction [1], [2], [3]. While MS enables highly precise procedures with improved patient outcomes, it also demands exceptional dexterity and accuracy, as even minor errors can lead to complications [4]. Thus, robot-assisted surgery has emerged to improve stability and accuracy in microsurgery [5]. However, effective robotic assistance relies on accurate real-time segmentation of critical structures such as microvessels, needles, and wires [6]. Therefore, it is crucial to develop advanced segmentation methods tailored to the unique challenges of MS environments for enhanced robotic assistance in MS.