DESK (Dexterous Surgical Skill) dataset

DESK数据集是由普渡大学工业工程学院开发的一个外科手术技能数据集，专注于通过模拟环境加速真实机器人的学习过程。该数据集包含从三个不同的机器人平台（Taurus II机器人、模拟Taurus II机器人和YuMi机器人）收集的外科手术技能数据，主要用于外科手势分类任务。数据集的创建旨在解决不同机器人平台间知识转移的问题，特别是在军事和灾难救援场景中部署的可现场使用的手术机器人。数据集的收集过程包括随机化元素，如物体的位置和板的方向，以增加数据集的多样性和实用性。DESK数据集的应用领域主要集中在通过模拟数据提高真实机器人数据的学习效率，解决外科手术技能训练中的关键问题。

The DESK dataset is a surgical skill dataset developed by the School of Industrial Engineering at Purdue University, focusing on accelerating the learning pipeline of real-world robots via simulation environments. It collects surgical skill data from three distinct robotic platforms: the Taurus II robot, the simulated Taurus II robot, and the YuMi robot, and is primarily used for surgical gesture classification tasks. The dataset was created to address the knowledge transfer challenge across different robotic platforms, especially for field-deployable surgical robots used in military and disaster relief scenarios. The data collection process incorporates randomized factors such as object positions and board orientation to improve the dataset's diversity and practical utility. The main application scenarios of the DESK dataset center on enhancing the learning efficiency of real robot data through simulated data, and solving key problems in surgical skill training.