Saved in:
Bibliographic Details
Main Authors: Zhang, Chengyi, Ye, Zi, Wang, Ziyang
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.23068
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913154425946112
author Zhang, Chengyi
Ye, Zi
Wang, Ziyang
author_facet Zhang, Chengyi
Ye, Zi
Wang, Ziyang
contents Reliable visual understanding in robot-assisted and minimally invasive surgery (RMIS/MIS) demands more than accurate masks: in clinical practice, clinicians pose language-like questions about procedural context, visibility, artefacts, and the presence of anatomical structures and surgical instruments, often under degraded views caused by occlusion, smoke, bleeding, and specular highlights. We present \textbf{RoboSurg-VQA}, a segmentation-aware visual question answering (VQA) benchmark built by repurposing public surgical segmentation datasets under a shared schema. Each frame is paired with a fixed set of clinically motivated questions spanning procedure context, anatomy (including region), imaging modality/view, surgical artefacts, image quality, and basic visibility and spatial attributes, with closed answer sets to enable consistent evaluation. To scale annotation, we generate candidate answers via constrained prompting with automatic validity and consistency checks, followed by human auditing to improve plausibility and label consistency. We report benchmark statistics, sanity baselines, and common evaluation challenges under challenging surgical conditions. The code will be available on https://github.com/ziyangwang007/Robosurg-VQA.
format Preprint
id arxiv_https___arxiv_org_abs_2605_23068
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle RoboSurg-VQA: A Multimodal Benchmark for Surgical Segmentation-Aware Visual Question Answering
Zhang, Chengyi
Ye, Zi
Wang, Ziyang
Computer Vision and Pattern Recognition
Reliable visual understanding in robot-assisted and minimally invasive surgery (RMIS/MIS) demands more than accurate masks: in clinical practice, clinicians pose language-like questions about procedural context, visibility, artefacts, and the presence of anatomical structures and surgical instruments, often under degraded views caused by occlusion, smoke, bleeding, and specular highlights. We present \textbf{RoboSurg-VQA}, a segmentation-aware visual question answering (VQA) benchmark built by repurposing public surgical segmentation datasets under a shared schema. Each frame is paired with a fixed set of clinically motivated questions spanning procedure context, anatomy (including region), imaging modality/view, surgical artefacts, image quality, and basic visibility and spatial attributes, with closed answer sets to enable consistent evaluation. To scale annotation, we generate candidate answers via constrained prompting with automatic validity and consistency checks, followed by human auditing to improve plausibility and label consistency. We report benchmark statistics, sanity baselines, and common evaluation challenges under challenging surgical conditions. The code will be available on https://github.com/ziyangwang007/Robosurg-VQA.
title RoboSurg-VQA: A Multimodal Benchmark for Surgical Segmentation-Aware Visual Question Answering
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.23068