ORQA
收藏资源简介:
ORQA是一个全面的手术室理解基准数据集,它整合了四个公开的手术室数据集,包括MVOR、4D-OR、EgoSurgery和MM-OR,包含了各种手术室场景的图像、点云、音频和文本数据。数据集包含了23种不同类型的问答对,旨在帮助计算系统更全面地理解手术室环境,提高手术数据科学的研究水平。数据集的创建过程包括数据收集、标注、预处理和问答对的生成。ORQA数据集可用于手术室领域的多任务学习和推理,有助于解决手术室环境下的复杂问题,提高手术操作的精度和安全性。
ORQA is a comprehensive benchmark dataset for operating room (OR) understanding. It integrates four publicly available operating room datasets, including MVOR, 4D-OR, EgoSurgery, and MM-OR, and contains images, point clouds, audio, and text data across various operating room scenarios. The dataset encompasses 23 distinct types of question-answer pairs, aiming to help computational systems comprehensively understand operating room environments and advance research in surgical data science. The creation process of the ORQA dataset includes data collection, annotation, preprocessing, and generation of question-answer pairs. The ORQA dataset can be used for multi-task learning and reasoning in the operating room domain, assisting in solving complex problems in operating room settings and improving the accuracy and safety of surgical operations.

- 1ORQA: A Benchmark and Foundation Model for Holistic Operating Room Modeling慕尼黑工业大学计算机辅助医学程序系, 德国 · 2025年



